See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/4258437

Single Key Omni directional Pointing and command System (SKOPS) -a smart on
screen navigational tool for physically disabled persons

Conference Paper · July 2007

DOI: 10.1109/HEALTH.2007.381629 · Source: IEEE Xplore

CITATIONS READS

3 976

2 authors:

Suman Deb Sujay Deb

National Institute of Technology, Agartala Indraprastha Institute of Information Technology
86 PUBLICATIONS   241 CITATIONS    86 PUBLICATIONS   1,679 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Minimally Buffered Network on Chip Router for High Throughput Applications View project

Geometric Optimization for Map Based Path Planning Reactive Navigation with Sensor Fusion View project

All content following this page was uploaded by Suman Deb on 24 March 2015.

The user has requested enhancement of the downloaded file.

 ICMIT 2005
(Indian Conference on Medical Informatics & Telemedicine)

February 27 – March 1, 2005

Indian Institute of Technology Kharagpur

India
 MESSAGE

It gives me great pleasure that a national level conference on medical informatics and
telemedicine has been organized at IIT, Kharagpur during 27th February to 1st March, 2005
under the auspices of Department of Computer Science and Engineering, School of
Information Technology, and School of Medical Science and Technology.

IIT, Kharagpur has always encouraged inter-disciplinary research on new and emergent
technologies and took the pioneering role in initiating a postgraduate program in Medical
Science and Technology, first of its kind in the country. I am happy that delegates from
technical and medical fraternities from all over the country will be attending the
conference.

I am confident that the deliberations and exchange of ideas during the conference will
be fruitful and contribute towards enhanced use of information technology for better
healthcare services in India.

I extend my best wishes to all participants, invited delegates, and the organizers and hope
that ICMIT 2005 will be a grand success.

February 7, 2005 S. K. Dube

PHONE : +91-3222-255386 / 282002 (Off.), +91-3222-277201 / 282003 (Res.)

FAX : +91-3222-255239 / 282000 # E-mail : director@iitkgp.ernet.in, skdube@iitkgp.ernet.in
 INDIAN CONFERENCE ON MEDICAL INFORMATICS AND
TELEMEDICINE
(ICMIT 2005)
February 27 - March 1, 2005
Web : http://www.icmit.iitkgp.ernet.in
Indian Institute of Technology, Kharagpur-721302, India

From the Program Chair

The aim of the Indian Conference on Medical Informatics and Telemedicine (ICMIT 2005) is to
provide a common forum for researchers and medical professionals to exchange ideas and
deliberate on current research issues in this emerging field. This conference is the first of its kind
organized in India and we are overwhelmed by the enthusiastic response obtained from all over
the country.

We have received a large number of original contributions from academic institutions, R & D
laboratories, institutes of medical science and hospitals as well as from industry. All the papers have
been peer reviewed and I am grateful to the reviewers for taking out time from their busy schedules
to review the papers. Technical papers accepted for presentation, which have been received in
full camera-ready formats, have been included in this volume. Apart from these, there are several
invited lectures whose abstracts are available in a separate booklet.

The conference is going to be held over two and a half days – from 27th February afternoon to 1st
March, 2005. There will be eight technical sessions, four special invited talks, and a session for
industrial presentations.

I am extremely happy that most of the active research groups in India working on medical
informatics and telemedicine are participating in the conference. Several eminent personalities in
the field will also be here to share their views and ideas and I am sure the discussions and
interactions will enrich us all.

I take this opportunity to thank the Invited Speakers, Session Chairs and Committee members,
without whose support we could not have completed this task. I am sure the sessions will have live
discussions and will help in promoting research in this area throughout the country.

February 7, 2005. A. K. Majumdar

Professor
Computer Science and
Engineering Department
Indian Institute of Technology
Kharagpur 721302
 INDIAN CONFERENCE ON MEDICAL INFORMATICS AND
TELEMEDICINE
(ICMIT 2005)
February 27 - March 1, 2005
Web : http://www.icmit.iitkgp.ernet.in
Indian Institute of Technology, Kharagpur-721302, India

Message from the Organizing Chair

On behalf of the organizing committee let me extend my heartiest welcome to all the participants
and delegates of the Indian Conference on Medical Informatics and Telemedicine (ICMIT-2005) to
be held at the Indian Institute of Technology, Kharagpur during 27th February and 1st March of 2005.
The conference is jointly organized by the Department of Computer Science and Engineering, the
School of Medical Science and Technology, and the School of Information Technology of this
Institute. This also reflects the interdisciplinary character of the ICMIT-2005, an appropriate forum for
exchanging ideas among researchers, academicians and professionals in areas of Computer
Science and Information Technology and Medical Science and Technology. I hope the
conference will stimulate new collaborative research and development activities in the area of
Medical Informatics and Telemedicine to the benefit of our society. Lastly, I take this opportunity to
thank our sponsors of this conference without whose support it would have been very difficult for us
to organize such an event.

Wishing you all a pleasant stay and brain storming sessions during the conference.

February 7, 2005. Jayanta Mukhopadhyay

Professor
Department of Computer Science and
Engineering
Indian Institute of Technology
Kharagpur, 721302.
 COMMITTEE MEMBERS

Conference Chairs:
Program Chair : Prof. A.K.Majumdar, Department of Computer Science & Engineering, IIT Kharagpur
Organizing Chair : Prof. Jayanta Mukhopadhyay, Department of Computer Science & Engineering,
IIT Kharagpur
Finance Chair : Prof. Shamik Sural, School of Information Technology, IIT Kharagpur
Convenor : Prof. S.K.Ghosh, School of Information Technology, IIT Kharagpur

Committees:
Advisory Committee:
Dr. G.D. Gautama, Department of Information Technology, Government of West Bengal
Dr. Sangita Reddy, Apollo Hospitals Enterprise Ltd, Hyderabad
Prof. Ashoke Jhunjhunwala, Department of Electrical Engineering, IIT, Madras
Shri S. K. Roy, Commissioner & Secretary, Health & Family Welfare, Transport, I&C (IT), Government of Tripura
Prof. Anupam Basu, , Department of Computer Science & Engineering, IIT Kharagpur
Prof. S.C. DeSarkar, School of Information Technology, IIT Kharagpur
Prof. R. C. Arora, School of Medical Science & Technology, IIT Kharagpur
Prof. S.K. Guha, School of Medical Science & Technology, IIT Kharagpur

Program Committee:
Prof. Sneh Anand, Centre For Bio Medical Engineering, IIT Delhi
Prof. S.K. Mishra, Department of Endocrine Surgery, SGPGIMS, Lucknow
Dr. N.K.Singh, TeleVital India Pvt Ltd, Bangalore
Dr. L.S. Satyamurthy, Antrix Corporation Ltd, Bangalore
Dr. Sanjay Bedi, Dept. of Pathology, Shri Guru Ram Das Institute of Medical Sciences, Amritsar
Dr. B.S.Bedi, Department of Information Technology, Government of India, New Delhi
Dr. S. Sanyal, Central Hospital, SE Railway H.Q, Kolkata
Dr. S.K. Dey Biswas, Indian Council of Medical Research, New Delhi
Prof. A.G. Ramakrishnan, Department of Electrical Engineering, Indian Institute of Science, Bangalore
Prof. Soumyo Mukherjee, School of BioSciences & Bioengineering IIT, Bombay
Dr. S.K.Sharma, EKO X-Ray & Imaging Institute
Prof A.K Ray, Department of Electronics & Electrical Communication Engineering, IIT Kharagpur

Organizing Committee:
Shri Ravikant, Department of Information Technology, Government of West Bengal
Shri Suresh Kumar, Department of Health and Family Welfare, Government of West Bengal
Dr. J.N. Maiti, WEBEL Kolkata
Dr. B. Majumdar, IIT Kharagpur
Dr. A. K. Sangal, Space Application Centre (SAC), ISRO, Ahmedabad
Dr. O.P.Sharma, Ind. Medical Ltd, Chandigarh
Dr. Koel Chaudhury, School of Medical Science & Technology, IIT Kharagpur
Dr. P. K. Biswas, Department of Electronics & Electrical Communication Engineering, IIT Kharagpur
Dr. A. Jeyram, RRSSC, ISRO, Kharagpur
Dr. S. B. Gogia, IAMI, New Delhi
Dr. Chandrayee Das, Inspiration, Kolkata
Shri. R. Bhattacharya, FW & PM Government of Tripura
 CONTENTS

Title Authors Page No.

Telemedicine in India

Impact of Information and Communication Technology in Kartar Singh 1

Health Care, Medical Education and Research
Application of Telemedicine for Tropical Diseases Janaki N. Maiti ,Atanu Jana 5
Telemedicine Initiative of Asia Heart Foundation Vinod Kumar Gupta, Devi Prasad 9
Shetty, Alok Roy

Telemedicine Applications

National Informatics Centre’s Tele – CME for Peripheral Shefali S. Dash, Naina Pandita. 11
Doctors
Health Management Information System: The Current Indian Sita Naik, A.K. Sarkar, Rakesh 14
Perspective Aggarwal, Rakesh Pandey
Information Technology in Medicine – Measuring and A. Nandakumar 18
Mapping Cancer Patterns
Telepathology-Indian Scenario Manoj Jain, Ramesh K. Gupta, 20
Rakesh Pandey, Saroj K Mishra
Teleradiotherapy Network in Uttar Pradesh To Provide A N. R. Datta, Saroj K Mishra 22
Comprehensive Radiation Therapy Care for Cancer Patients
Evidence of Benefit from Telemedicine Sanjeev Garg, Vinod Gupta, 27
A.Raghuvanshi

Assistive Technology

The Bombay –Vellore Artificial Hand-Recent Developments Suresh Devasahayam, Suranjan 29

Bhattacharji
Bio Feedback Training for The Use of Assistive Devices Suresh Devasahayam, George 33
Tharion
Single Key Omni Directional Pointing and Command System Suman Deb, Sujay Deb 37
(SKOPS)-A Smart on Screen Navigational Tool for
Physically Disabled Persons
Information Technology for The Education of The Persons A.K. Sinha 42
with Hearing Impairment.
Comparison of Normal and Pathological Knee Swing Cycle Suhas S. Gajre, Rajendra K. Saxena, 46
Electrical Impedance Signals for Possible Noninvasive U. Singh, S. Anand
Diagnosis of Osteoarthritis
A Neurolinguistic Analysis of Bengali Hyperkinetic Nilantika Chakraborty, Krishna 50
Dysarthric Speech Bhattacharya, Trishit Roy
Extending Library Facilities To Physically Cahallenged-Case Atin Nandi, Pratap K.J. Mohapatra 54
of IIT Kharagpur
 Title Authors Page No.

Health Information System & e-Health Care

An Approach to Low Cost Wireless Tele-Monitoring System D. Goswami, J. Mukhopadhyay, 59

for Rural Population in Developing Countries S. Mukherjee, S. Chakrabarti
e Health Care India’s Unfulfilled Dream David Raju Manne 63
Med-Centre-Drug Inventory Monitoring and Control System G.S.Bansal,S.S.Duggal, Rahul Jain, 67
Ravinder Geol, Anil Aggarwal.
Data Management System for Community-Based Shivaprasad S Goudar, Margo F 71
Randomized Controlled Trial in India Brinkley, David G Forvendel, Arthur
Pradeep Singh, Amit P Revankar,
Bhalachandra S Kodkany, Richard J
Derman
Developing e- Care Pathways for The 21st Century Health Sheenu Jhawar 76
Information Systems
Online Speech Recognition for Medical Transcription K. Ramamohan Reddy, S. Prajit 79
Nair, Y. Madusudhana Reddy,
S. Dandapat
Referral System Planning Using GIS Chandreyee Das, Jibanananda Roy. 83
Knowledge Management in Health Systems: The Emerging Sanjay Bedi 86
Role of Chief Information Officer (Health) From
Departmental Level To Health System
Computerized Physician Order Entry: Way Forward Pankaj Gupta 89

Telemedicine Systems & Hospital Information Systems

Secured Network Solution for Health Care Industry with Surid Kumar Das , A.Vadivel 93
HIPAA Compliance
Role of Online Thrombolysis in Coronary Heart Disease Atanu Ghosh, Tapas Dutta, Devasis 97
(CHD) Pal, Avijit Dutta
Data Conferencing Over A Low-Bandwidth Communication Anunay Nayak, V. Pallavi 100
Link
Strategies in Assessing Telemedicine Quality M.PalaniNathaRaja, Subash 104
Wadhwa, S.G.Deshmukh
Telemedicine Systems-An Overview P. Hari Krishna Prasad 109
The Design of A Web Based Telemedicine System Amiya Kumar Maji, J. 112
Mukhopadhyay
The Scope of Utilising Telemedicine For Improving The Prasanta Pathak, Sanjeev Bakshi 116
Functioning of The Indian Public Health System in The
Prevalent Socio-Economic and Cultural Scenario
Web Based Patient Record System Y. Madhusudhana Reddy, Dandapath 122
S., Harshal B. Nemade
 Title Authors Page No.

Medical Imaging & Computer Aided Diagnosis

A Diagnostic Tool for Early Predictions of Demyelination S.B. Mehta, Santanu Chaudhury, A. 127
Using M R Images Bhattcharyya, A. Jena
Detection of Blood Vessels in Retinal Images Using Model M. Ravi Kumar, S. Dandapat 131
Based Approach
A Unified Image Processing Based Model for Early Cancer Rajib Mahapatra, P. K. Biswas 135
Detection from Mammogram Images
Lesion Segmentation in Mammograms by Mean Shift Nagaraju Odala, Ajoy Kumar Roy 139
Algorithm
Co-Occurrence Based Clustering for Mixed Bio-Medical Lipika Dey, Amir Ahmad, Vipul 142
Datasets Goel, Rajat Mangla
Improved Detection of Ventricular Ejection Time For V. K. Pandey, P. C. Pandey 146
Impedance Cardiography
A Wavelet Based Method for Identification of Tags From Ajay V. Deshmukh, Tanish Zaveri, 151
SPAMM Cardiac MR Images for HARP MRI. Vivek Shivhare , Alok O. Modak,
Vikram M. Gadre, Deepak P. Patkar,
Sona Pungavkar
Standardisation of Traditional Tongue Diagnosis Aided by Debasis Bakshi, Sujata Pal 155
Computerized Digital Photography - A Study

Medical Instrumentation

ECG Data Acquisition and Monitoring System for Prashant Agrawal,Arundhati Jana, 159
Telemedicine Application Ajit Pal
Characterization of Motion Artifacts Using Wavelet Vijay S Nimbargi, Vikram M Gadre, 163
Transform and Neural Networks Soumyo Mukherji
Eeg Signal Processing for Monitoring Depth of Anesthesia Anil Srivastav, Amod Kumar, Y. L. 167
Narayanan, Sneh Anand
A New Low-Cost Muscle Strength Testing System for P.Thirusakthimurugan, P.Dananjayan 171
Neuroleptic Patients

Medical Standards

Our Own HLX For Telemedicine S. B. Gogia, S. B. Bhattacharya 175

Implementation of HL7 Standards In Telemedicine Pramod Adiddam, Suman Kundu, 178

Environment J. Mukhopadhyay

Healthcare Data Interchange Standard-Health Level Seven Sudhir Agarwal , Ravi Saksena 183
(HL7)
 Title Authors Page No.

Abstracts and Short Papers

Standardization in Telemedicine in India: Initiatives and B.S.Bedi 189

Policy Issues
Taking Telemedicine To Indian Villages The Wimax Way D. Lavanian 190
Teleradiology - Clinical Experience and Education R.B. Gujral , Archna Gupta 192
Telemedicine for Disaster Mitigation in Health A Ayyagari, Ujjala Ghoshal 193
Tele-Consultancy in Specialty Medicine Bidyut Das 194
Tele-Cme for Post Graduate Medical Students B.N. Mohanty, S. S. Mishra, 195
D. Patel, S. K. Mishra
Biomedical Applications of Electroporation on Human Red S.H.Sanghvi, K.P.Mishra 196
Blood Cells
Establishing Healthcare Bridges in The International Arena: Rini Verma 197
An Informatics Company Perspective
Telemedicine in India - ISRO's Initiatives L.S.Satyamurthy & R.L.N.Murthy 198

Sponsors 200
TELEMEDICINE IN INDIA
 IMPACT OF INFORMATION AND COMMUNICATION TECHNOLOGY IN
HEALTH CARE, MEDICAL EDUCATION AND RESEARCH
Professor Kartar Singh
Director, Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow, Uttar Pradesh INDIA

Abstract there is diversity in the standard of quality of teachers,

availability of educational tools, books and journals,
Discoveries in science and medicine, combined quality and the type of patients materials due to
with advances in information and communication geographical difference in the incidences of various
technologies, are leading to a transformation in health diseases, pattern of diseases in rural, urban and semi
systems and services, and in the relations between urban patients and availability of the medical
individuals, institutions, communities and resources, diversity in the quality of the students
governments. These technologies hold great promise because of entry and exit parameters. Medical research
for low-as well as high-income countries, and some activities are limited to only to few advanced medical
countries are already realizing the benefits. These centers and there is hardly any link between these
benefits apply not only to the healthcare delivery researchers except occasional encounters in annual
systems, but also to public health, administration, conferences. Information and communication
finance, education, research and health related markets technology holds a great promise to network people,
more generally. A lot of activities are happening in the organisations, libraries, research laboratories across all
field of telemedicine in India and the benefits are well kinds of barriers including geographical separation.
perceived by the stake holders in all the areas of health Uniformity can be achieved in the areas of health
e.g healthcare, medical education and research. Central care, medical education and research across a state and
government agencies such as Indian Space Research country in general.
Organisation and Information Technology and many
state governments are supporting telemedicine
Telemedicine in India: 1,2,3
initiatives across the country. Though the technology
of telemedicine holds a great promise in all the areas of Healthcare is a state subject which follows a three
health sector, the real impact for our country needs to tier system – primary health centre catering a group of
be measured in scientific terms after the technology is villages, Secondary level health centre located at
used by larger groups. More effort is now being district level and medical college hospitals constitute
directed to develop enterprise based telemedicine the tertiary level located in the big cities. Besides, there
network in a systematic manner which will eventually are few advanced medical institutes of national
lead to a national level telemedicine network. importance having clinical, teaching and research
facilities in many super-specialties. In addition to
Background government run health system, same hierarchical
healthcare services exists in private sector. There is no
India is a vast country of 1. 4 billion population national health insurance system, though the
occupying an area of 3,287,268 sq. kms. It consists of government, public sector and corporate organizations
29 states and six union territories governed by a federal sponsor the healthcare expenses of its employees and
system. There is no national health insurance policy for family. In the recent years few insurance companies are
the country. Government supported healthcare delivery venturing into health sector. Inspite of well networked
follows a three tier system and is the primary health care system access to healthcare in rural areas is
responsibility of each state. A great deal of digital far from satisfactory. In the current scenario, 75% of
divide in health exists between rural and urban areas the qualified consulting doctors practice in urban, 23%
since most of the specialists are located in mid size and in semi-urban (towns) and only 2% in rural areas where
large towns and cities where as majority of Indian as the vast majority of population live in the rural
population live in villages. The same phenomenon also areas. Hospital beds/1000 people are 0.10 in rural as
exists in medical education. There are about 150 compared to 2.2 in urban areas. Further, a vast
medical colleges located in twenty-one states. In recent proportion of north and north-eastern region of country
times many super-specialties and sub-specialties in the lie in hilly terrain and some territory in remote islands
medical sciences have been developed. Medical experts making healthcare reach impossible to such far flung
in these fields are few and are located in large areas. Tele-healthcare concept is no longer new to the
university medical centers. Medical education is unique country. Both government and private agencies are
in the sense clinical material constitutes study material. venturing into it. Few Indian companies are being
Since most of the medical colleges aim at providing capable of providing hardware and software solution
education at graduate and postgraduate level, such for tele-health care. Products of reputed overseas tele-
super/ sub-specialty areas are not covered enough. Also health industry have their presence. Efforts are directed

1
towards setting up standards and IT enabled healthcare telecommunication infrastructure, specialist
infrastructure in the country. All those activities carried availability, geographical considerations, etc, Some of
out by various agencies are collected and summarised these initiatives are briefly presented below:
below. • DIT has supported development of telemedicine
software systems – the prominent ones by C-DAC.
Indian Space Research Organisation Under this ongoing project, technology developed
(ISRO) : has been used for connecting three premier
medical institutions - viz. SGPGI Lucknow,
Indian Space Research Organisation (ISRO), as AIIMS New Delhi and PGIMER Chandigarh-
part of its commitment for social sector development using ISDN connectivity. The technology
have been applying space technology for healthcare developed is now being deployed for setting up
and education, under GRAMSAT (rural satellite) other Telemedicine systems in the country.
programme. It has initiated number of Telemedicine
pilot projects which are very specific to the needs of • Tele-medicine for diagnosis & Monitoring of
development of the society. ISRO’s Telemedicine tropical diseases in West Bengal using low speed
initiative has been broadly divided into the following WAN, developed by Webel (Kolkata), IIT,
areas: Kharagpur and School of Tropical Medicine,
Kolkata has been implemented. The system has
a) Providing Telemedicine Technology & been installed in School of Tropical Medicine
connectivity between remote/rural hospital and Kolkata and two district hospitals. About a
Super Speciality Hospital for Teleconsultation, thousand consultations have already taken place
Treatment & Training of doctors & paramedics. over this network. Another project on setting up
b) Providing the Technology & connectivity for of telemedicine facilities at two referral hospitals
Continuing Medical Education (CME) between and four district hospitals using West Bengal State
Medical Colleges & Post Graduate Medical Wide Area Network of 2 Mbps is also under
Institutions/Hospitals. implementation.

c) Providing Technology & connectivity for Mobile • An Oncology Network for providing Telemedicine
Telemedicine units for rural health camps services in cancer detection, treatment, pain relief,
especially in the areas of phthalmology and patient follow-up and continuity of care in
community health. peripheral hospitals (nodal centres) of Regional
Cancer Centre (RCC) has been established. The
• Presently ISRO’s Telemedicine Network consists Telemedicine network utilizes Internet
of 90 Hospitals – 61 Remote/Rural/District connectivity in addition to leased lines. The project
Hospital/Health Centre connected to 29 Super was implemented by C-DAC, Trivandrum and
Specialty Hospital located in the major cities. RCC. More than 4000 patient consultations have
More than 12,500 patients have been provided been done till date using the network. A cost
with Teleconsultation & treatmen under ISRO benefit analysis has shown that economic benefits
project. to the patients have been far more than the
investment made in this project.
Department of IT, Ministry of
Communication & Information Apollo’s Telemedicine Initiatives:
Technology4 Apollo is a pioneer in the field of Telemedicine in
Information and Communication Technology has India and is credited with being the first to set up a
enabled major potential contributions in a large number Rural Telemedicine Centre in the village of Aragonda
of economic and social sectors. Telemedicine is one in the state of Andhra Pradesh. Even though the first
such area, which utilizes telecommunications for Telemedicine site was launched in 1999, Apollo’s
affecting specialized consultations for diagnosis and initiatives with Telemedicine started well before that.
treatment of diseases at a distance. Realising the Some of the earlier experiences with Telemedicine
benefits and having the capability in terms of the applications are given below:
technical and medical expertise in India, Department of • Apollo experimented with Telecardiology more
Information Technology(DIT), Indian Space Research than 5 years back through the usage of
Organization (ISRO) of Department of Space, and Transtelephonic ECG machines, which were set up
other public and private organisations have started in various smaller hospitals whereby a doctor based
Telemedicine projects in different parts of the country. in the Tertiary hospital could monitor his patients
As a facilitator, DIT has taken initiatives for from a distance.
development of technology, initiation of pilot schemes • Simulations of Teleconsultations between Apollo
and standardisation of Telemedicine in the country. Hospitals, Hyderabad, Chennai and Dubai for cases
The pilot schemes were carefully chosen to take into specific to Cardiology, Neurosurgery &
account the diverse issues related to currently available Orthopedics were done.

2
• Medical Conference - Brussels. Surgical procedure Escort Heart Institute & Research Center
in Cleveland U.S.A – received in Apollo Hospitals Project
Delhi, Chennai, & Hyderabad in 1997.
Telemedicine Centres set up by Apollo Installed in 2002 by Escort Heart Institute & Research
Center, it has been involved in telecardiology service.
Network Foundation:2
Apollo has set up over 45 Telemedicine Centres Mobile Tele-Ophthalmology service:
across different locations in the country and many
more are in the pipeline. Apollo has worked with With the support of ISRO, Shankar Nethralaya at
different kinds of entities in the healthcare industry
Chennai and Meenakshi Eye mission at Madurai have
ranging from large corporate hospitals and Government
hospitals to small clinics and Information Centres. launched Mobile Tele-ophthalmology service.
Apollo has the expertise to execute different kinds of
Telemedicine projects. Online Telemedicine Research Institute,
Ahmedabad initiative:2
Telemedicine Initiatives at Sanjay Gandhi
This industry has been involved in R&D
Postgraduate Institute of Medical Sciences,
activities in telemedicine hardware and software and
Lucknow:4,5,6
has been successfully executing projects of ISRO and
Sanjay Gandhi Postgraduate Institute of Medical DIT with turnkey approach.
Sciences ( SGPGIMS), Lucknow is a tertiary care
referral hospital has been actively involved in Conclusion
telemedicine since 1999. It is connected with three
medical colleges of Orissa ie Cuttack, Berhampur and e-health services are being adopted by healthcare
Burla through VSAT. It follows the systemic approach provider organizations in India gradually. The growth
through proper scheduling of telemedicine sessions of e-health services has given rise to the need for a new
well in advance. Different departments conduct
breed of healthcare professionals, healthcare
regular tele-education, teleconsultation and tele-
followup sessions by exchanging patient’s, history, administrators and healthcare technologists. This
x’rays, pathology slides etc thus the patients are given industry needs people who can understand any two of
the specialist consultation regarding the treatment, healthcare, business and technology. Historically the
follow-up after the treatment at SGPGIMS. In NIC healthcare industry has been the last adopter of
project , CME sessions are conducted monthly towards technology, the same has been the case with healthcare
professional carrier development of doctors of North education. Even till today, many of the medical, dental,
East states. Nine state headquarters of NIC are nursing, pharmacology and other healthcare degrees do
simultaneously connected to SGPGIMS through broad
not have courses on information technology. The
band interactive mode and 450 CICs through video
broadcast mode via DAMA based satellite system. industry has grown to such a size that people from non-
Regular monthly tele-education sessions are held that healthcare backgrounds are being recruited and trained.
enabled the doctors to improve their learning and Interestingly, a lot of youngsters with traditional
quality of patient care. The telemedicine centre of healthcare degrees such as MBBS and BDS are
SGPGIMS is connected via satellite to similar facilities exploring career options in e health. As technology is
in AIIMS, New Delhi and PGI, Chandigarh under a pervading more into our education system, learning is
project of Ministry of Information Technology. The becoming “anywhere and anytime”. E-learning enables
Mercury software for telemedicine was developed in
students to study and appear for exams at their
collaboration with C-DAC as part of Research and
Development. It is under evaluation phase. SGPGI is convenience from any place they want to, as long as
now setting up a School of Telemedicine & they have a good internet connection. Some forward
Biomedical Informatics in its campus with the thinking companies such as Medvarsity have started
objective of meeting the demand of highly skilled offering healthcare courses to healthcare students and
health technologist in this emerging area. practitioners. It is helping medical students to prepare
for higher studies through online courses. It also has
Asia Heart Foundation telemedicine courses directed to both students and practitioners. e.g.-
initiative:2 courses in subjects such as emergency medicine,
Installed in 2002 by Narayana Hrudayalaya, health insurance, etc. E-Health service companies can
Bangalore, Karnataka it has now achieved a figure of get business more easily if their employees’
more than 2000 tele- cardiology consultation through qualifications are recognized by their clients. Since
an enterprise based network. majority of revenue comes from American companies,
many companies are encouraging their employees to

3
enroll for online healthcare courses recognized by
American educational institutions and associations. In
the coming years India is going to have more e health
activities considering the present trend.

References
[1] http://www.whoindia.org/EIP/GATS/10-67.pdf [5] Proceedings of 6th IEEE Healthcom 2004,
[2] http://www.mit.govt.in/telemedicine/home/asp Odawara, Japan
[3] http://www.whoindia.org/EIP/GATS/13- [6] Proceedings of 9th Annual Conference of
Annex2.pdf American Telemedicine Association, Tempa,
Florida, 2004, USA
[4] Proceedings of 2nd APT Telemedicine Workshop
2004, New Delhi

4
 APPLICATION OF TELEMEDICINE FOR TROPICAL DISEASES
Dr. Janaki N. Maiti and Atanu Jana
Webel Electronic Communication Systems Ltd.,
DN-53, Sector -V, Salt Lake, Kolkata-700091, India
E-Mail: jnmaiti@yahoo.co.in ; wecs@giascio1.vsnl.net.in

Abstract distant populations by bridging the rural-urban divide

and distance barrier through use of Communication and
The paper deals with the application of Information technology.
Telemedicine for Diagnosis and Medical consultation
of Tropical diseases at Government Hospitals in West Recognising the important role the Telemedicine
Bengal. The Telemedicine facilities have been set up at can play in providing health care to the people,
one Referral Center- The School of Tropical Medicine, Department of Information Technology, Ministry of
Kolkata and two Nodal Centers- MJN Hospital, Dist- Communications and Information Technology,
Coochbehar and Habra State General Hospital, 24 Government of India have sponsored several projects
Parganas (North). The systems are in operation since for the Development and Application of Telemedicine.
2002 using POTS as communication link which has The Computer Science and Engineering Department,
been upgraded in 2003 with ISDN for higher IIT, Kharagpur and Webel ECS Ltd. Kolkata took the
bandwidth. The medical information of patients in initiative in 1999 to undertake few Telemedicine
different formats are transmitted mostly off-line while projects for indigenous development of technology
on-line sessions are also conducted for conferencing along with its applications for various diseases using
and data transfer as required. More than 800 patients different communication links of varying bandwidth.
(Leprosy, Malaria, Kala Azar, other Blood and Skin Since the communication links available in most
related diseases) of Habra State General Hospital and remote locations in India were ordinary PSTN lines
more than 400 patients of Coochbehar Hospital have and not ISDN or leases lines, the need for developing a
been provided expert medical services from the referral system that would work even in low bandwidth PSTN
STM, Kolkata under the project. A typical lines were strongly felt. The stress expert subject
Telemedicine system, the statistics of the patients initially chosen was Tropical disease for which expert
referred from the two Nodal Centers to the Referral services could be provided by the Specialty center-the
Center and few case studies are presented in the paper. School of Tropical Medicine, Kolkata.
The pilot project sponsored by DIT, Govt. of India has
been implemented by Webel ECS Ltd, Kolkata in
association with IIT, Kharagpur who have developed The Project
the Telemedicine Software (TelemediK).
The project reported here is the 1st Pilot Project
Introduction undertaken for implementation by WEBEL, Kolkata
and IIT, Kharagpur for Tropical diseases. It is on the
Telemedicine is the use of information and development and application of Telemedicine for
telecommunication technologies to deliver health diagnosis, monitoring and medical consultations of
services to those in need by electronic transmission of Tropical diseases (Leprosy, Blood and Skin related
the necessary expertise and information among diseases) using low-speed WAN. The project has been
geographically dispersed parties, including physicians sponsored by Department of Information Technology,
and patients, to result in improved patient care and Government of India. The Referral center is the School
management, resource distribution efficiency and of Tropical Medicine, Kolkata which has been
potential cost-effectiveness. It also includes activities connected with two Nodal centers –one at Habra State
relating to educational and administrative uses of these General Hospital, District-24 Parganas (N) and the
technologies in the support of the health care such as second at MJN Hospital, District –Coochbehar at a
distance learning, continuing medical education, distance of about 700 Km from the referral center. The
training for isolated or rural health practitioners and connectivity between the Referral and Nodal centers is
video conferencing for administrative purposes. ordinary POTS. IIT has developed, tested, upgraded
Telemedicine offers one of the best solutions to health and customized the software. WEBEL has done system
care service providers to deliver health care services integration, equipment installation, commissioning,
particularly to the underserved rural and geographically testing and trial run of the systems at the centers.

5
Software Development The typical system requirements are :

The Telemedicine System developed by IIT, • Hardware Requirements for Referral and
Kharagpur consists of a backend database server and Nodal centers:
several data communication tools. It has been o One Intel Pentium server with
developed on the Windows NT platform with the MS 256MB RAM and 80 GB HDD and
SQL server as backend RDBMS and operating on low One Intel Pentium PC with 128MB
speed WAN. The user requirement analysis has been RAM and 80 GB HDD
exhaustively carried out in consultation with the o Multimedia Kit
doctors of the centers. The developed Telemedicine o Keyboard, Mouse etc
software (named, TelemediK) has different features o Color Monitor
and functionalities. It has the features of both store and
forward technology and on-line conferencing. The • Software Requirements for Referral and
specialty of the system is that it can operate with Nodal centers:
various communication channels of varying o Operating System - Windows 2000
bandwidths such as POTS, ISDN, Leased Line, VSAT Professional or Windows XP
etc. It records patient information in various forms o Network Operating System –
such as Text/Hypertext Image (JPEG, BMP, DICOM), Windows NT/2000/2003
Graphics, Audio, video etc. It supports an integrated o Database- SQL Server 2000
on-line communication module for white board
conferencing with images, graphics and text. It • Peripheral Devices at Nodal Centers:
provides web-enabled customized user interface for o X Ray Scanner,
different diseases. It has an integrated patient o Digital camera,
information management system for efficient storage o Microscope with CCD camera with
and retrieval of medical records with a backend image grabber card supporting JPEG
RDBMS. and BMP image formats
o ECG Recorder
o View Box
Telemedicine System
• Networking Requirement
A schematic diagram of the Telemedicine system is o Telephone lines and Modems (for
given in Fig 1. normal PSTN Connection)
o ISDN lines, Router, NT 1 (for ISDN
connection)
o Two pair copper wire, Router,
Leased line Modems (for leased line
connection)
• Video conferencing Systems
o Web Camera for PSTN links
o Multipoint V.C. systems for the
Referral and point to point systems
for Nodal centers for ISDN/leased
lines

Application of Telemedicine

The Telemedicine facilities have been set up at one

Referral and two Nodal centers in West Bengal and the
technology developed for low bandwidth channel has
been used. The systems have been connected using low
speed POTS when most of the data are transmitted off-
line and on-line conference made with Web camera.
The systems at all three centers are operating since July
2002. A very good number of patients of Leprosy and
blood and skin related diseases at the distant Nodal
centers have been referred and provided medical
services/treated by the specialists of the Referral
Fig. 1 Schematic diagram of Telemedicine system center. The statistics of the patients referred and treated
at the two Nodal centers are given below:

6
 Total No. of patients referred by Habra Centre
to STM, Kolkata : 804
No. of patients under Blood
related diseases : 25
No. of patients of Leprosy : 5
No. of patients under General
Blood & Skin related disease : 774

Total No. of patients referred by

Coochbehar Centre to STM, Kolkata : 452
No. of patients under blood related The prescription from the doctor of the Referral
diseases : 17 center transmitted for the patient at the Nodal center is
No. of Leprosy patients : 13 shown below:
No. of patients under General –
Blood & Skin related diseases : 432

The connectivity has later on been upgraded also

with ISDN at all the centers and Video conferencing
equipment has been added to the system.

Case Studies
The patients treated at the Nodal centers include
Leprosy, and various blood and skin related diseases
i.e. Malaria, Kala Azar etc. The cases of several
numbers of patients appear to be interesting. However,
because of shortage of space and time for the The patient required treatment for a long time.
presentation, we give here only two case studies
relating to two patients of Nodal center referred to The name of another patient is say, YYY whose
STM, Kolkata. clinical document is:

The name of the patient is say, XXX. His case Chief Complaints & History of Present Illness:
history is: The Patient is suffering from generalized skin rash with
itching all over body for last few weeks. Past History:
Before 16-17 years back once the patient applied Recurrent episodes of similar attacks for last few years.
Blacknit oil on his head for blacking of the hair. Then General Survey: Anemia - mild, Drainage Lymph
he noticed allergy on the head. For that he took so Glands -NAD, Jaundice - Nil, Liver & Spleen - NAD.
many medicine including homeopath but all in vein. Systemic Examination Cardiovascular: NAD
After that he used BETNOVATE OINTMENT but it Respiratory: Scattered Ronchi over the Chest.
would not help him. In SKIN CARE Centre they Abdominal: NAD. Neurological: NAD. Investigation
adviced Manolicen Ointment and Tablet and used to General Hematology: Polymormhonuclear (80%)
take sunlight but it would not help. Then he noticed Leucocytosis (10,000/C.C.), ESR -56 mm / Hr. Blood
small spots on the knee, elbow, head and other Biochemistry: Blood Sugar (F)-100mg%, Sugar (P.P.)
different parts of the body. Then DR. P. Banerjee gave -120mg%. X-Ray & Images: Chest X-ray—NAD.
him DIPROVATE Ointment and lotion. It helps him. Special: Not Done

The medical images of the skin disease- patient are The medical images of the patient are shown below:
shown below:

7
 patients desire to have live interaction with the
specialists at the Referral center using video
conferencing facilities The communication links
between the centers were found to be not very reliable
and need to be improved.

Conclusions

The prescription from the doctor of the Referral The Telemedicine technology has been developed
center transmitted for the patient at the Nodal center is and the project implemented successfully at one
shown below: Referral and two distant Nodal locations for Tropical
diseases using even low speed PSTN lines. More than
1400 patients have already been treated and many more
will be benefited under this project. . In view of the
successful operation of the Telemedicine system, two
more projects using ISDN and leased line as links and
covering other diseases like Cardiology, Radiology,
Pediatrics, Neurology etc. have f been taken up for
implementation by Webel ECS Ltd and IIT, Kharagpur
in 13 more centers in West Bengal with financial
support of DIT, Government of India

Acknowledgment
The authors are grateful to Prof. A. K. Majumdar
and Prof J. Mukhopadhaya, Dept. of Computer Science
and Engineering, IIT, Kharagpur for developing the
software and providing the software support in
Lessons learned implementing the project.. The authors are thankful to
Department of Information Technology and
The patients at the distant Nodal end are found to be Department of Health & Family Welfare, Government
eager to use the facilities and get the medical services of West Bengal who were the facilitators for the project
from the specialists of the Referral center. However, and to Department of Information Technology,
some doctors were found unwilling and not interested Ministry of Communications & Information
to use the system or to refer the cases of the needed Technology, Government of India for sponsoring the
patients to the experts of the Referral center. Almost all project.

8
 TELEMEDICINE INITITIATIVE OF ASIA HEART FOUNDATION
Dr Vinod Kumar Gupta, Dr Devi Prasad Shetty, Dr Alok Roy
Rabindranath Tagore International Institute of Medical Sciences
(A unit of Asia Heart Foundation)
124 Mukundapur, E M Bypass
Kolkata - 700099., India
vkg258@yahoo.com

Abstract electrocardiogram done. This data is sent to R N

Tagore International Institute of Cardiac Sciences
Telemedicine is the one of the ways to treat (RTIICS) at Kolkata or Narayana Hruthayalaya (NH)
patients in places where it is difficult to provide at Bangalore through the communication network set
specialist medical advice and care. An effort was made up. A specialist at the hub centre gives his advice that
by Asia Heart Foundation to provide emergency is carried out by the doctor at the remote centre.
cardiac care to patients where such facilities were not Similarly any patient wanting to consult any specialist
present. Connectivity was through a simple telephone, at the hub centers is similarly screened.
fax, ISDN, satellite link. The outcome was satisfactory
and significant morbidity and mortality was A total of nearly 5000 in-patients and around
significantly reduced. 13000 out-patients have been advised and treated
through this network without their having to travel.
Keywords: Telemedicine, e-Health, cardiac This group included new, follow- up , and
emergencies, connectivity, cardiac care unit, distance postoperative patients. Most of the in-patients were
treatment. discharged. Feedback forms from these patients was
positive and the role of telemedicine in helping them
Introduction was appreciated.

India is a vast country where majority of the

population reside in villages and towns. These places Conclusion
do not have access to specialist medical advice and
care. For such a facility they need to travel great India is a vast country where majority of the population
distances, spend a lot of time and money. In cardiac reside in the semi-urban and rural areas. These people
patients early recognition of the problem and its early do not have access to specialist medical advice and
treatment helps to reduce morbidity and mortality. Asia care as such facilities are available only in a few
Heart Foundation had set up a few cardiac care units in metropolitan cities. One of the many reasons for such a
the districts and sub-divisional hospitals in West disparity could be due the high infrastructure costs and
Bengal, Tripura, Karnataka in participation with the the lack of enough number of specialists. A team of
respective governments and ISRO over a period of motivated doctors, nurses, technicians, politicians,
more than three years. bureaucrats along with a reliable telecommunication
link can change the health scenario of our country.
Body of paper
Any person attending the emergency department Acknowledgement
with complaint of chest pain or any other complaint
related to heart is referred to the cardiac care unit The telemedicine project is under the joint partnership
(CCU). Here he is examined by the doctor –who is a of Asia Heart Foundation, ISRO and the governments
medical graduate- history is recorded, and an of West Bengal, Karnataka, and Tripura.

9
TELEMEDICINE APPLICATIONS

10
 NATIONAL INFORMATICS CENTRE’S TELE – CME FOR
PERIPHERAL DOCTORS

Dr (Ms.) Shefali S. Dash, Sr. Tech. Director, Ms. Naina Pandita, Sr. Tech. Director
National Informatics Centre, New Delhi – 110003; email: dash@hub.nic.in

Abstract for providing equity of health care delivery across the

country. These two major applications - teleconsulting
Continuing Medical Education (CME) plays a vital and Tele-education can provide better health care as
role in knowledge up gradation and when medical well as training of doctors in all areas of the country.
professionals are located in remote areas of the
country, and then it becomes vital for them to Community Information Centres (CICs)
somehow participate in these programmes. The
Northeastern part of the country has many prestigious The North-East of India is a land of breath-taking
institutions with a large number of specialists catering natural beauty. The region comprises eight states viz.
to the health care of the common man. Participating in Arunachal Pradesh, Assam, Manipur, Meghalaya,
CMEs becomes a difficult task with poor or average Mizoram, Nagaland, Sikkim and Tripura. The hills
telecommunication networks. National Informatics with dense forest cover make traditional forms of
Centre’s network, NICNET reaches the far and remote communication very difficult. A vital need has been
areas of the country thus solving the problem of felt to extend the reach of modern day technology to
telecommunications. Using NICNET CME the area in order to enable rapid socio-economic
programmes were launched for doctors in the North development and bring the area closer to the national
East and Sikkim under its Community Information mainstream. With this objective in mind, the Prime
Centre (CIC) programme. Using the videoconference Minister announced a plan for the establishment of IT
facilities available, NIC has conducted several such facilities at the block level in the North-Eastern Region
sessions wherein specialists are invited to deliver a in February 2000. In April 2000, the then Ministry of
lecture on a specific topic for the benefit of doctors in Information Technology launched a project to establish
these State capitals. In addition, the specialists’ Community Information Centres (CICs) in all blocks in
lectures are being broadcast to 483 CICs in all the North-Eastern States [1]
North Eastern states and Sikkim. Doctors from
Uttranchal and Lakshwadeep also benefit from these The CICs were set up with the following
programmes. This way doctors in the remotest areas objectives:
are able to benefit from these Tele-CME programmes.
• Establish IT infrastructure at Block level.
Introduction • Create IT awareness amongst local population.
• Provide email and web access.
CME programmes using the concept of
• Conduct computer based training programmes.
“teleconferencing” is a real time and live interactive
program in which one set of participants is at one or • Provide citizen centric services.
more locations and the other set of participants are at • Provide access to socio-economic databases.
another location. The teleconference allows for • Facilitate CME programmes.
interaction, including audio and/or video, and possibly • Use IT tools for sustainable regional development.
other modalities, between at least two sites. In India it
has been found that since many areas are very far away At present Community Information Centres have
from medical facilities, a number of persons do not been set up in 457 blocks of the North East and these
adequate treatment due to lack of contact with were dedicated to the people of the region on 17th
specialists. Also doctors in remote locations do not August 2002 by the Hon’ble Minister Shri. Pramod
have access to the latest information in their field and Mahajan. Under this computer and communication
are unable to discuss cases with their peers as well as infrastructure is being set up in each block with the
specialists due to distances and poor communication centres being located in local schools, colleges or
facilities. Many parts in the Northeast as well as in suitable government buildings. In order to ensure
other remote areas do not have access to adequate uninterrupted communication, the Centres are
medical care. This is especially due to a hilly terrain, connected through a satellite based computer
lack of good roads and the absence of speciality communication network and each CIC is equipped
referral hospitals in the area. Since videoconferencing with a VSAT, computer systems, printers and
effectively shortens distances, it can be a powerful tool networking equipment.

11
 in North East using the existing Videoconference
NIC’s Videoconference Facilities: facilities. These CME programmes sessions were
aimed at small groups of doctors in different areas of
The network used for NIC's videoconferencing specialization and were restricted to a brief
services is a SCPC VSAT link operating at 128 Kbps lecture/presentation by a specialist followed by a
from each of the 206 locations. All the locations are question answer session. In these sessions’ doctors
connected to Delhi in a star configuration. NIC is also from the Northeast were able to discuss different
providing videoconferencing services from many aspects of a disease like the diagnostic aspects,
Central and State Government Ministries using ISDN treatment and management aspects or a drug.
lines. LAN/WAN Gateway is also installed in NIC's
network to transcode between the H.323 environment The 1st such session was organized in October 2001
over LAN and H.320 environment over WA [4]. wherein a senior Gastroenterologist, Dr. Randhir Sud
was invited from Sir Gangaram Hospital, New Delhi.
Portable Videoconferencing Facility -
The specialist, while sitting at NIC Hqrs., New Delhi
"Anytime...Anywhere...Videoconferencing
could address the problem of Diagnosis and
NIC has two portable SCPC VSATs of 1.8 m and Management of Hepatitis A & B, areas of great
1.2 m dish size, which can be easily carried to any interest to the doctors in the North East.
destination by air cargo within few days. Once oriented
towards the satellite wherein NIC has leased Following this, doctors from Sikkim requested a
transponder capacity, portable VSAT provides full session to be arranged with Dr. Naresh Trehan,
duplex network connectivity at higher speeds. With Director of Escorts Heart Research Institute, New
this facility, videoconferencing and high speed internet Delhi. He was requested to speak on Coronary
access can be provided from any city / district / village Artery Diseases. After these two successful CME
of India in a short time. This service can be utilized to programmes sessions a schedule was drawn up to hold
provide temporary videoconferencing service for regular sessions. Some of the specialists invited were:
conferences, exhibitions and other important events.
Dr. U. Kaul from Batra Hospital to speak on “Balloon
Angioplasty Techniques”, Dr. Sita Naik from
Voice, Video & Data Services over SCPC DAMA
SGPGIMS, Lucknow to speak on "Basic Serological
Network
Profile of Hepatitis Infections and the Diagnostic
NIC has established SCPC DAMA network Tests and their Interpretation”; Dr. Bupesh Kaul
Management System at New Delhi and has started from Magee Women’s Hospital, Pittsburgh, USA who
voice, videoconferencing and high speed Internet delivered a talk on “Painless Labor Techniques”, Dr.
access services in an integrated manner using a single Suman Bhandari from Escorts Hospital to discuss
VSAT. Using this technology, any DAMA location by different aspects of “Stress and Hypertension” and
simply using a handset can initiate a voice, video or Dr. A.K. Dutta from Kalavati Saran Hospital, New
data call, without the manual intervention of Hub Delhi to talk on “Perinatal Asphyxia and its
operator. Connections are made in a mesh topology on Management”.
a point-to-point basis between any nodes in the
network using on demand SCPC carriers. A specialist was also invited from Manipur, Dr.
Loukham Manglem Singh, from J.N. Hospital,
The key feature of SCPC DAMA is that bandwidth
Porompat, Imphal to discuss the various aspects of
is provided to the remote VSAT only when it is
required. This helps in better utilization of satellite “Neonatal Sepsis”. The sessions are announced on the
bandwidth amongst various VSATs in a closed user CIC webpage (www.cic.nic.in) and the minutes of each
network. session are posted on this page enabling the doctors to
note down the major issues discussed in each CME
SCPC DAMA NMS is like a telephone exchange in programmes session [2].
the sky, which works on circuit-switching concept. It
sets up Voice circuits at 8 Kbps, Videoconferencing For over one year NIC has tied up with SGPGI,
circuits upto 384 Kbps and asymmetric data circuits Lucknow to bring in specialists from that institute to
upto 2 Mbps within 5 seconds of demand by conduct the CMEs for the benefit of doctors not only in
application in an automatic manner. the North East, but also from Lakshwadeep and
Uttranchal. These programmes are broadcast to all the
CME Programmes: 450 odd CIC’s in the North East enabling the doctors
in the remotest part of the country to participate in
As the CIC’s were being set up NIC decided to these programmes. Questions are asked either
experiment in “bringing” the specialists to the doctors

12
interactively or posted a few days prior on the website “Molar Pregnancy” by Prof. Mandakini Pradhan,
or during the lecture session [3]. SGPGI, Lucknow
Some of the CME programmes that have been
conducted were: “Advances in Laboratory Technology" by Prof. Sita
Naik, SGPGI, Lucknow
"HIV Post Exposure Prophylaxis in Health Care
Workers" by Dr Renu Dutta, Professor of “Pelvic Inflammatory Diseases” by Dr. Anju Rani,
Microbiology, Lady Hardinge Medical College, New SGPGI, Lucknow
Delhi
“Dysfunctional Uterus Bleeding” by Dr. Mandakni
"De-stress Stress" by Mr Atul Gandotra, VP, Pradhan, SGPGI, Lucknow
Morepan Labs, New Delhi

"Depression in General Practice" by Dr Achal Bhagat, Conclusion

Apollo Hospital, New Delhi
These CMEs, covering a range of topics have
"Pelvic Inflammatory Disease" by Dr Aruna Batra, attempted to overcome the problems faced by the
Safdarjang Hospital, New Delhi doctors in the peripheral areas of the country in
knowledge enhancement and upgradation. These
“Congenital Malformations: before and after birth” by doctors without having to leave their cities/towns can
Dr (Mrs.) Shubha Phadke, Additional Professor, participate in these programmes to provide better
Medical Genetics, SGPGI, Lucknow health care facilities to the common man in these parts
of the country with the assistance of NIC’s
“Principles of management of Diabetic Foot and its telecommunications infrastructure.
prevention” by Dr. A.K. Verma, SGPGI, Lucknow.

“Diagnostic Approach to Infective Diarrhea in Bibliography

Community & Rational Antibiotic therapy” by Dr
Archana Ayyagari, SGPGI, Lucknow 1. CIC website : http://www.cic.nic.in

“Recent Trends in Management of Rheumatoid 2. Dash S.S, Kumar, S. National Informatics

Arthritis” by Prof. R.K. Mishra, SGPGI, Lucknow Centre’s initiatives in Distance Education for
medical professionals in remote areas of
“Budd Chiary Syndrome” by Dr. Kartar Singh North East & Sikkim. Biennial Conference of
(Director), SGPGI, Lucknow Indian Association of Health Informatics at
PGI, Chandigarh, 18-19 October 2002
“Birth Defects of heart, its presentation and
treatment” by Dr. Nirmal Gupta, SGPGI, Lucknow 3. Dash S, Chakarborty S, Ray D, Pandita N.
Health services for the common man over
“Respiratory distress in Children” by Prof. A. K. National Informatics Centre’s network
Baronia, SGPGI, Lucknow (NICNET) Asia Pacific Telemedicine
Conference, New Delhi, 25th – 26th February
“Management of Acute Coronary Syndrome” by 2004.
Prof. Nakul Sinha, SGPGI, Lucknow
4. NIC’s Videoconference Services website
“Diabetic Retinopathy” by Prof. Kumudani Sharma, http://vidcon.nic.in/
SGPGI, Lucknow

13
 HEALTH MANAGEMENT INFORMATION SYSTEM: THE CURRENT
INDIAN PERSPECTIVE

Sita Naika, AK Sarkarb, Rakesh Aggarwalc, Rakesh Pandeyd

Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow, India

Abstract developmental activities of health sector in India.

The national health policy statement (1983)
Successful health management programs require recognized that appropriate decision making and
monitoring of the health status of the population and program planning in the health and related fields
provision of efficient health care services. These was not possible without establishing an effective
services should provide state-of-art patient health information system. The need to establish a
management and also cover areas such as drugs nationwide organizational set up to procure essential
stocks and consumption patterns, equipment status health information, which may provide support for
and availability, finances and personnel local management of the healthcare and effective
management. Accurate, relevant and upto- date decentralization of activities, was identified. It is
information is essential for health service managers clear that in the current context, these objectives can
if they are to recognize weakness in health service only be met through widespread implementation of
provision and take actions that will improve service IT solutions at national, state and local levels.
delivery. Accordingly the development of effective
information systems is a necessary precursor to
managerial improvement.
HIS in India

Hospital Information system (HIS) is a process

whereby health data (input) are recorded, stored, According to industry analysts, any hospital with a
retrieved and processed for decision-making minimum capacity of 100 beds is a potential IT-
(output). Decision-making broadly includes buyer. It is stimated that there are over 1,000
managerial aspects such as the planning, hospitals in that category in the country. Nasscom
organization and control of health care facilities and believes that healthcare organizations in India will
clinical aspects including (a) provision of optimal spend around Rs 1,000 crores on IT in the current
patient care, (b) training of medical personnel to year. The major component of this will be hospital
generate appropriate human resources and (c) management systems and networking hardware
facilitate research and development activities in required for processes like telemedicine.
various fields of medicine.

A majority of the hospitals in the country are rooted

The Bhore committee report, release soon after the in manual processes, which are unable to cope with
independence, identified the need for a strong the volume of data generated. In the larger hospitals,
information system as a support to the various patient records remain difficult to access and these

14
adversely affect quality of health delivery. The HIS solutions available
demands of the insurance sector for more efficient
information storage and retrieval are also going to
add to the pressure on hospitals and health providers. Although a large number of products are available in
This may, in fact be the major driving force for the market , the major players in this field, are
modernization of this sector since the health CDAC, Wipro, TCS and Siemens Information
insurance sector is poised for major growth in the Systems Ltd (SISL).
coming decade. The current Government health
policy is also shifting its emphasis towards health
insurance. CDAC, an autonomous government IT organization,
was a pioneer in developing HIS solution in India.
They have developed the first total HIS software in
Automation is the only solution that will can help collaboration with Sanjay Gandhi Post Graduate
hospitals to meet the challenges of modern health Institute of Medical Sciences, Lucknow in 1997.
care delivery. However, IT has been a late entrant in This has been implements at SGPGI and GTB
this field and most hospitals which forayed into this hospital, New Delhi as also at other ongoing sites.
area started with small systems that were developed Wipro, through its division Wipro Healthcare,offers
inhouse. Till the middle 90’s no standardized a reasonably complete range of IT services including
solutions were available and these local innovations domain specific solutions for healthcare
were the pioneers. However, they neither give the organizations. Their end-to-end solution covers the
desired results nor can they be integrated with newer entire spectrum of the healthcare industry’s needs
systems. The major demand for updated solution including a Hospital Information System (HIS),
started with the establishment of the large corporate Picture Archival Communication System (PACS)
hospitals many of which like the Apollo group, and telemedicine solutions. Their Hospital
implemented strong IT solutions in the latter half of Information Resource Planning System (HIRePS) is
the nineties. With the increasing demands of the a new generation enterprise wide software solution
market, many sturdy, standard HIS solutions were that covers all aspects of hospital management. The
developed by the major IT companies. Today, the solution is spread across 60 modules overing
Healthcare segment is, in fact, going through the functionalities like administration, clinical, support
kind of evolution that the banking and financial and back office among others. They have over 200
services sector went through a decade ago. This is person years of experience in executing healthcare
being driven by the huge annual increase in the projects and have worked with over 30 top hospitals
number of hospital beds mostly in the corporate in India and abroad.
sector.

Tata Consultancy Services has specific solutions

However, the government still owns 66% of the and a healthcare portal and its solutions include the
Indian hospital market. The Central Government has TATA HMS and Intensive Care Unit Management
also announced setting up of large new hospitals. Systems (ICUMS). TCS plans to increase their
The public sector hospitals have realized the healthcare domain specialists from the current 450
direction in which the wind is blowing. This can be professionals to about 900 in the next 18 months. It
appreciated by the fact that Delhi Government has is currently implementing the complete suite of 22
recently approved HIS solutions for four of its major modules of its HMS at the Chennai-based Sankara
hospitals. Nethralaya and at Woodlands Hospitals of Kolkata.

15
Siemens Information Systems Ltd (SISL) has been segments in a star configuration, 100 MBps LAN
implemented at the Amrita Institute of Medical switch, UTP cabling totaling 10 km, with a
Sciences, a 900-bed multi specialty hospital based in combination of 2/24 port 10/100 MBPS unmanaged
Cochin. It had one of the first hospital information hubs/ switches and 8-port mini hubs. Currently over
systems in the country developed entirely by an 250 computers are on the network. The 365 X 24
Indian team and launched in 1996. SISL now sells hour network runs on multiple platforms with
the solution across the world. servers running on Sun Solaris with Microsoft
Windows 95/98/XP as front end. The RDBMS is
Oracle 9i. We are currently upgrading the network to
a class B one with a 1 Gb central switch. The
Hospital Information System (HIS) at existing Hubs are being replaced with switches and
SGPGIMS, Lucknow additional 5kb cables are being laid so as to cater for
an additional 300 terminals. This exercise will
prepare the network for PACS as well as for total
Sanjay Gandhi Postgraduate Institute of Medical integration with the telemedicine center.
Sciences is a 100% UP State funded Institution with
a 600-bedded tertiary care hospital. It has 120 senior
consultant doctors, 200 trainee doctors and 500 The Institution has been able to oversee a smooth
paramedical personnel and provides state of the art transfer of a 100% manual system to a 100%
care in 12 clinical specialties. It also has the computerized one over a 2-year period. At the time,
additional responsibilities of training and research. it was the first fully computerized hospital in the
The hospital caters to approximately 30,000 new country with all patient related activities taking place
registrations and 20,0000 hospital admissions per online and continues to be only one in the
year. Government sector. Since 2000, all patient care
functions are computerized with 1200 users having
different security levels of access to the database.
SGPGI has been a pioneer in this country in health The current size of the database is 9 GB and it is
IT. It took a decision as far back as 1990 to totally growing at an annual rate of 2.0 GB. Majority of the
automate hospital functions. At the time there were nurses and technical and clerical staff had no
no indigenous software or hardware expertise in the previous exposure to computers and had to be
country. Hence those initial forays did not yield any trained. Ongoing training is provided to all new
success. A fresh attempt in 1995 brought together medical trainees who join each year. While the shift
the Department of Electronics (as it was then to computer based working increased the self-esteem
called), Dr Vijay Bhatkal at the Center for of many classes of employees, it has, in fact,
Development of Advanced Computing (CDAC) increased the work burden in some cases.
Department of Computer Sciences, IIT (Kanpur) and
SGPGI. Realizing the paucity of available solutions,
CDAC took up the challenge of developing the HIS
software with the domain knowledge input by Picture Archival Communication System
SGPGI. This software was developed and deployed (PACS)
between 1998 and 2000. It has 14 modules covering
all aspects of patient management and patient related
administration. PACS (Picture Archival Communication System) is
an integrated system of digital products and
technology allowing for acquisition, storage,
With the technical expertise provided by Department retrieval and display of pictorial images such as
of Computer Sciences, IIT, Kanpur, SGPGI radiology, pathology, endoscopy etc. Since
established a class C IP enterprise network transmission of graphics is more complex, PACS
consisting of Sun servers configured as Cluster (Sun solutions were expensive and more difficult to
cluster 2.2), 3 km of optical fiber consisting of 11 implement. However, with the advances in

16
communication standards, decreased costs, open a total HIS solution. The major IT companies in this
platforms and possibilities for phased integration field are VEPRO, SIEMENS, WIPRO etc. and they
with existing HIS it is becoming a more widely have multiple installations of PACS and Data
accepted modality for data storage. In fact, the warehousing in India and abroad. SGPGI has also
Patient Medical Records are incomplete without the initiated plans for installation of PACS for its
pictorial data and hence PACS is an essential part of hospital.

17
 INFORMATION TECHNOLOGY IN MEDICINE – MEASURING AND
MAPPING CANCER PATTERNS

Dr A. Nandakumar
Deputy Director General (Sr Gr) and Officer-in-Charge
National Cancer Registry Programme (Indian Council of Medical Research)
No. 557, ‘Srinivasa Nilaya’, New BEL Road, Dollars Colony, Bangalore 560094 (e-mail: ank@blr.vsnl.net.in)

Abstract need to be covered. The NCRP cancer registries cover

Under a WHO supported project on ‘Development selected urban centres and just one rural pocket. The
of an Atlas of Cancer in India’,1,2,3 a cost-effective other population cancer registries also cover essentially
design and plan using advances in modern electronic urban centres except for parts of two districts in Kerala
information technology, was conceived to collate and and part of one district in Tamil Nadu. Consequently,
process relevant data on cancer. The objectives were to the patterns of cancer in several urban centres and rural
obtain an overview of patterns of cancer in different regions remain largely unknown. Setting up of new
parts of the country; and, calculate estimates of cancer registries throughout the country as in some Western
incidence wherever feasible. The Internet through a countries would involve enormous cost in establishing
web-site www.canceratlasindia.org was identified as and maintaining the same. Therefore, under this
the primary communication medium. This was a project, on ‘Development of an Atlas of Cancer in
unique concept being tried for the first time (in India India’ a cost-effective design and plan using advances
and to the best of our knowledge anywhere else in the in modern electronic information technology, was
world) under the project. conceived, to collate and process relevant data on
cancer so as to fulfill the objectives of obtaining an
In the two year period (1 January 2001 to 31 overview of patterns of cancer in different parts of the
December 2002), information on 217,174 cases f from country; and, calculating estimates of cancer incidence
105 centres across the country was available for wherever feasible.
analysis. Some of the cancer patterns that arose were
expected but several new ones emerged. The study Methods
demonstrated the immense potential of the system and
the numerous possibilities for cancer research and
control. The concept in using web-based design and The Internet was identified as the primary
approach with on-line transmission of cancer data has communication medium for collecting the data.
worked – a major advance for using Information Internet as a medium of data collection for patient
Technology in Medicine – Measuring Disease Burden information was a unique concept being tried for the
and Health Informatics. The project was extremely first time (in India and to the best of our knowledge
cost-effective. anywhere else in the world) under the project. The
‘core proforma’ for collecting the patient information
was hosted on the web-site www.canceratlasindia.org.
Key Words The proforma for data entry on the web-site is designed
Information Technology, Measuring Disease, so as to be user friendly and reduce the time taken for
Cancer Surveillance data entry. Internet Browser based data entry
eliminated the need for software to be installed on
Introduction every system and the hassles of administration and
maintenance. Collaborating centres were given an
The Indian Council of Medical Research initiated a individual login-ID and password with detailed
network of cancer registries under the National Cancer instructions on entering the core patient information.
Registry Programme (NCRP) in 1981 and data Care was taken to code/encrypt the data entered so that
collection commenced in these registries from January the identity or the nature of the data could not be
1982. Since then, the registries have provided deciphered by any one except those concerned with the
information on incidence and patterns of cancer that in project. This ensured confidentiality of patient
terms of quality and validity meet international information and security of data transmitted.
standards. Thus, in India, for cancer, and perhaps for The successful working of this concept is reflected,
only this disease, we have a systematic programme of first, in the data that is downloaded on a regular basis
data collation so as to have reliable incidence and for the past two and a half years. The core data of
mortality rates, thereby laying a foundation for approximately 1200-1500 cancer cases is received
scientific research – whether that research be every week. Secondly, in the feedback received from
epidemiological, basic, clinical or in cancer control. the participating centres during the All India Workshop
However, India being a vast country, extensive areas - 95% of the respondents, felt, that the web-site was

18
easy to use and 80% of them have a fairly stable
Internet connection. Thirdly, because most of the
collaborating centres were able to transmit the required Conclusion
information as soon as a diagnosis was made the report
for the years 2001 and 2002 could be brought out fairly Three essential features were observed:
early (in the beginning of 2004). In due course the 1. The results gave a whole new set of cancer
experience should enable us to provide the main tables incidence and patterns demonstrating the
of incidence rates soon after the end of the calendar immense potential of the system and the
year and then on-line. numerous possibilities for cancer research and
control. It identified hot spots of high incidence,
Data was also received through floppy disks and in recognised belts of geographic areas with specific
some places through copies of completed proforma. types of cancer and discerned likely zones for
establishing population based cancer registries.

Results 2. The concept in using web-based design and

approach with on-line transmission of cancer data
has worked – a major advance for using
Based on the above, a total of 2,17,174 cancers for
Information Technology in Medicine – Measuring
the two year period (1 January 2001 to 31 December
Disease Burden and Health Informatics.
2002) from 105 centres across the country was taken
up for analysis.
3. The project was extremely cost-effective.
The results confirmed some known and expected
Sustained support on a long-term basis with scope
features and also brought to light several new ones.
for considerable expansion would be required for
Many districts in the state of Aizawl reported overall
setting up a composite database through what may be
very high incidence rates of cancers in both males and
called a National Electronic Surveillance System for
females. These rates were the highest that are reported
Cancer. That then would constitute a sound platform
to-date anywhere else in India.
for quality research in cancer in all its dimensions. This
could be for observing incidence and patterns across a
Cancer of the lung and other sites of cancers
wide spectrum of populations, evolving analytic studies
associated with the use of tobacco like oral cavity
in molecular epidemiology to seek clues in cancer
(chewing tobacco in various forms is widely prevalent)
aetiology, looking at patterns of cancer patient care and
were extremely common. However, one major new
survival so as to regulate management or in evaluation
finding was the very high incidence of lung cancer in
of cancer control measures.
women in Aizawl (Mizoram state) and Imphal
(Manipur state). The incidence rates of 26.2 per
100,000 was only slightly lower than 35.5 per 100,000
reported in Scotland. References
There were several other newly discovered 1. NCRP – Development of an Atlas of Cancer in
“hotspots” of cancer patterns. The most important was India, 2001-2002. Volume I. National Cancer
the very high incidence rates of stomach cancer in both Registry Programme (Indian Council of Medical
men and women in the states of the Northeast like Research), Bangalore [www.canceratlasindia.org],
Mizoram, Manipur, Nagaland and Sikkim. The high 2004a
incidence rates of Gall Bladder cancer in women in
several districts along the river Ganges was note- 2. NCRP – Development of an Atlas of Cancer in
worthy. The coincidence of high incidence of cancer India, 2001-2002. Volume II National Cancer
cervix in women and penile cancer in men in some of Registry Programme (Indian Council of
the districts of Tamil Nadu state was the other new Medical Research), Bangalore
finding. It is known that both cancers could be caused [www.canceratlasindia.org], 2004b
by the Human Papilloma Virus. These districts have
also reported a high prevalence of Human 3. Nandakumar A., Gupta PC, Gangadharan P.,
Immunodeficiency Virus. The districts of the Visweswara RN and Parkin DM: Geographic
southwestern coast of the country showed a high Pathology Re-visited – Development of an
incidence of cancer of thyroid in females. This is an Atlas of Cancer in India, Int J Cancer (under
area noted for its high background radiation. The Publication).
incidence of cancer of the breast in women in Goa was
higher than expected. Very high incidence rates of
Nasopharyngeal cancer comparable with the highest in
the world were found in the northeastern states of
Nagaland and Manipur.

19
 TELEPATHOLOGY - INDIAN SCENARIO
Manoj Jaina, Ramesh K Guptaa, Rakesh Pandeya and Saroj K Mishrab
Department of Pathologya and Endocrine Surgeryb
Sanjay Gandhi Postgraduate Institute of Medical Sciences, Lucknow, Uttar Pradesh, India

Abstract speed) or via VSAT (384 KBPS). Greater bandwidth

will provide fast and effective data transmission.
Offline telepathology (store and forward) is used for
Telepathology is defined as performance of pathology sending still images for teaching and consultation.
at a distance using available telecommunication links. Dynamic telepathology (real time) usually requires
With the advent of latest hardware and software both way communication for sending and receiving
technology in computer and imaging and boom in live images directly from microscope. This can be with
information technology sector, effective and useful or without microscope control at the remote site. The
telepathoogy practice is now a reality. Potential of practice of telepathology is limited by the quality of
telepathology in health care and medical education is image and its rate of transfer. Real time case discussion
being realized in India as well. We present our is done with audio and video conferencing.
experience of telepathology at a tertiary care medical
centre over last four years. We conducted regular tele- For the beginners and individual pathologist at
eduation and tele-consultation sessions with three remote area, telepathology practice can be started with
medical colleges of Orissa, India. The outcome was just one digital camera attached to the microscope.
satisfactory in terms of participation, experience and Later on images are transferred and stored in a personal
knowledge gains. computer. With the help of the commonly available
Email and internet facility, images are attached to
Key Words: email in JEPG format and sent for expert opinion or for
tele-teaching and tele-discussion. Various dedicated
Telepathology, Tele-education, Tele-consultation, pathology web sites provide facility to upload cases for
Digital images, Store & forward telepathology, expert opinion and for discussion groups.
Dynamic telepathology
Other possible uses of telepathology include
sharing resources in terms of equipments, hardware
and software for the purpose of microscopic and
Introduction macroscopic image optimization, morphometry and
image analysis and also sharing skills among
Telepathology is defined as performance of pathologists. An essential prerequisite for good
pathology at a distance using available telepathology is good quality of microscopic
telecommunication links. It enables pathologists to preparations.
render diagnosis and to consult remotely. It can be
applied to all areas and activities of a pathologist, but Material & Methods
the main role is in anatomic pathology, especially for
intraoperative frozen section consultation, in tele- With an aim to increase the awareness of
education, and teleconsultation in surgical pathology, telepathology among the pathologist community and to
cytolopathology, immunopathology and electron assess the usefulness and difficulties encountered in
microscopy. telpathology practice in India, we conducted first
interactive telpathology session in India between
Instruments required for telepathology include a
PGIMER, Chandigarh and SGPGIMS, Lucknow, in
good quality CCD or digital camera attached to
Feb 2001 during Indo-US symposium and slide
microscope with appropriate mounts, which is
seminar on endocrine surgical pathology. Live images
interfaced with a latest computer having appropriate
from microscope were transferred and both way
hardware with enough storage capacity and facility for
interactive videoconferencing discussions were held.
data archiving and software for real time and still
Comments were received from the Department of
image capturing, image optimizing and image
Pathology, PGIMER Chandigarh, on a formatted
transferring to remote places by available
questionnaire. We also conducted eight static and live
telecommunication links. Another digital camera is
tele-eduation and tele-consultation sessions with
used for gross specimen photography. The computer is
Department. of Pathology, SCB Medical College,
linked with remote location having a similar facility,
Cuttack, during September-December, 2001. Seven
via telecommunication links such as Local Area
postgraduate students and faculty members from
Network or telephone or ISDN line (128 KBPS
Cuttack participated in the sessions. The emphasis was

20
on renal, hepatic, gastrointestinal and endocrine endocrine pathology, gastrointestinal and hepatic
pathology, as such cases are uncommonly seen at pathology and neuropathology were presented from our
Cuttack. Case details of 5- 6 lesions for one session institute to residents and consultants of pathology
were sent by email beforehand. Interactive live images department of these medical colleges and interactive
were transferred and both way discussions were held discussion was done by video conferencing. Cases and
by videoconferencing. Tele-consultation was done in a discussions were well appreciated. We also conducted
similar manner receiving stored and real time tele-education on interesting topics in surgical
microscopic images along with clinical details from pathology with three medical colleges during last one
Cuttack. year. We also received difficult cases for tele-
consultation. Limited opinion could be offered.
Results Limitations were suboptimal quality of microscopic
images and fewer images per case. Our opinion helped
According to the response on a questionnaire them in diagnosing and managing the cases.
provided by Department of Pathology, PGIMER
Chandigarh, the overall perception of video quality was Conclusions
perceived as good, and sound quality as unsatisfactory
by the participants. Telephone line had frequent Telepathology is in a developing phase in our
disconnectivity and slower image frames transfer rate. country. It requires awareness among pathologists and
The course content was interactive and educative and needs improvement in the quality of instruments used
was considered useful by both residents and and in bandwidth of telecommunication links. Store
consultants. and forward telepathology is now becoming common
among pathologists for individual consultation and
The sessions were well appreciated and discussion. With the improvement in the quality of
diagnosis were made in all cases by participants at digital archiving microscopes and in software and
Cuttack and the lesions were considered excellent and hardware technology in the near future, digital
educative by the participants at Cuttack. Image and scanning and archiving whole histopathology glass
sound quality was fair and they appreciated live images slides at higher resolution and transmitting contents of
over still images. In tele-consultation with SCB slide within seconds at higher bandwidth will improve
Medical College, Cuttack, limited opinion could be drastically the telepathology consultation. Finally, the
offered from SGPGI. Limitations were due to quality quality of microscopic preparations and image
of microscopic preparations and inability to see the optimization always plays a major role in making
desired field at desired magnification, due to slower decisions at the remote end.
frame transfer rate.

After our initial experience we started regular

telepathology sessions using either stored digital Address for Correspondence
images or transferring real time microscopic images
followed by video conferencing for discussion. Dr Manoj Jain
Sessions were conducted once or twice a month during Associate Professor
last two years from Jan 01 to Dec 02 with Pathology Department of Pathology
Departments of SCB Medical College, Cuttack, Sanjay Gandhi Postgraduate
MKCG Medical College, Behrampur and VSS Medical Institute of Medical Sciences
College, Burla Orissa. Fully worked up, uncommon Lucknow -226104, India
cases on particular sub-specialties like renal pathology, Email: mjain@sgpgi.ac.in
urologic pathology,

21
 TELERADIOTHERAPY NETWORK IN UTTAR PRADESH TO PROVIDE A
COMPREHENSIVE RADIATION THERAPY CARE FOR CANCER PATIENTS
Prof. Niloy Ranjan Datta*, Prof. Saroj Kumar Mishra+
*Head, Department of Radiotherapy and +Head, Department of Endocrine Surgery and Nodal Officer, Telemedicine
Sanjay Gandhi Postgraduate Institute of Medical Sciences, Rae Barelli Road, Lucknow – 226014, India
*nrdatta@sgpgi.ac.in ; +skmishra@sgpgi.ac.in

Abstract telecommunications and information technology, has

Radiation therapy forms an important enabled the emergence of Telemedicine in modern
component in the multimodality management of nearly health care. Telemedicine is utilized by health
80% of the cancer patients. Establishment of a providers in a growing number of medical specialties,
comprehensive radiotherapy facility needs substantial including oncology.
financial input that often becomes the major bottle
neck in creating new centres in government The role of telemedicine in radiotherapy has yet
institutions. With the projected global increase of 50% to be explored in India. Radiotherapy is an essential
in the cancer incidence by 2020, there would be an component of the multimodality management of
impending crisis world wide and specially in the cancer, especially in the today’s era of organ and
developing countries to match the increasing demand functional preservation of the diseased part. However,
with the limited radiotherapy centres. Teleradiotherapy the establishment cost of a state of the art radiotherapy
network, based on sharing the resources and centre is enormous and may not be always feasible in
optimization in its utilization could form a very all government medical colleges and hospitals. An
effective media, that could provide teleconsultation, effort should therefore be made to pool the limited
make the state-of-the-art technology available to all resources, device its optimal utilization so that the
patients, enable teaching and training through the benefit could be percolated to a majority of the cancer
“virtual class room” and facilitate conduct of multi- patients [1]. Added to this, the training of the
centric trials. This unique network with various manpower at all levels - nursing personnel, technicians,
radiotherapy departments of the State Medical Colleges medical physicists and doctors should be feasible for
of Uttar Pradesh (UP) is being planned for the first effective patient care.
time in the country under a project funded by the
Department of Science and Technology, Government Cancer is the second leading cause of death in
of India. Department of Radiotherapy at Sanjay Gandhi most developed countries today. Developing countries
Postgraduate Institute of Medical Sciences would act as are gradually following similar trends in cancer
a primary hub and link the six other radiotherapy incidence as the western world. These changes are
departments of the State of UP to share its occurring due to changing life styles and increasing life
comprehensive state-of-the-art technology in expectancy. Today, worldwide, cancer affects at least 9
radiotherapy, technical expertise both in manpower million people each year and kills 5 million. Its
development and conduct multi-institutional clinical economic as well as health consequences make the
trials in common cancers of this region. disease an alarming problem. At present, cancer
accounts for one tenth of the deaths world wide, and
Key Words the burden is bound to increase further due to an
increase in the population exposed to tobacco and a
Teleradiotherapy, cancer, telemedicine, neural gradual higher of proportion of elderly population in
network, eHealth most countries. The projected number of new cases by
the year 2015 is estimated to be around 5 million for
“Cancer is another area that would require special the developed countries and 10 million for the
attention ….. the programme is limited to procurement developing countries, while the number of deaths
of the equipment in a few urban centres and in a few estimated could be 3 and 6 million respectively [2,3].
other areas to prove that we have it elsewhere too. This
narrow vision should be changed” With such a magnitude of problem, it becomes
- Dr. A. P. J. Abdul Kalam, the need of the hour to lay stress on the development of
Honourable President of India, centers dedicated towards this cause, which would
In “India 2020: A Vision for the New Millennium” have comprehensive treatment facilities namely -
radiotherapy, surgery and chemotherapy along with
Introduction ancillary support. Apart from treatment of cancer
patients, these centers need also to be well equipped for
The rapid strides in the computer technology generating trained manpower, carry clinical and basic
aided by concurrent developments in modern

22
research in the contemporary areas of oncology a vital role in the providing radiotherapy services as a
especially relevant to the Indian subcontinent. Since referral centre, teaching, training and carry out clinical
radiotherapy forms a primary form of treatment and translational research of the common cancers seen
required in around 80% of cancer patients, an effective in the state of UP. This could be effectively possible
cancer management programme would need with the establishment of tele-radiotherapy links
establishment of centres having good radiotherapy between the various medical college departments
facilities. According to International Atomic Energy having limited facilities for radiotherapy. The present
Agency, the establishing these facilities in developing proposal of tele-radiotherapy is an effort in this
countries would cost around US $ 2.5 billion [4,5]. direction, perhaps for the first time in India.
Considering the annual cancer incidence in
India to be 0.5 million each year and the existing
cancer load to be 1.5 million, a total pool of 2 million
Different centres proposed for tele-
cancer patients annually seek help in India for
radiotherapy link
treatment at various levels of health care system.
Projected Indian statistics reveal that 1 in 9 males and 1 The project envisages to form a tele-
in 8 females will develop cancer during his or her radiotherapy link up with the various radiotherapy
lifetime [6]. The population of Uttar Pradesh was departments of the medical colleges of the state of UP.
estimated at 110 million in the 1981 census and the These would be linked with the department of
population projected for the year 2000 AD was 150 Radiotherapy at SGPGIMS (central coordinating unit:
million. It is estimated that annual new cancer patient primary node). These departments as secondary nodes
load would be around 100,000 to 120,000 by the year include: Departments of Radiotherapy at King
2010 and there would be a rise of around 56% in the George’s Medical University (KGMU), Lucknow; J. K.
incidence of cancer from 2001 to 2020 (Fig.1). The Cancer Centre, Kanpur; L.L.R. Medical College,
radiation treatment facilities exist in a limited number Meerut ; S.N. Medical College, Agra; Banaras Hindu
of medical colleges, although lately there are few University, Varanasi, Aligarh Muslim University,
private cancer centres, totaling to around 22 Aligarh and Kamla Nehru Hospital, Allahabad (Fig.2).
radiotherapy units, while the present requirement is
around 160 units. Tele-radiotherapy network linking all the
nodes will be set up. A consortium of medical teachers
from all the departments of Radiotherapy in the above
medical colleges and the university will be formed to
develop interactive teaching material.

The primary node at Department of

Radiotherapy, SGPGIMS would also be networked
with the hub located at the School of Telemedicine at
SGPGIMS. Each department of Radiotherapy of the
medical institutions will form a local tele-radiotherapy
committee comprising of the faculty members of the
department and one member of the committee will be
identified as the Nodal person who should be
authorized to co-ordinate the day to day running of the
Sanjay Gandhi Postgraduate Institute of project. In the first phase, the link up would be between
Medical Sciences (SGPGIMS) is a well-equipped super SGPGIMS and KGMU, Lucknow. Later the remaining
specialty tertiary care teaching hospital with an active medical colleges would be linked up in two phases
Radiotherapy department. The Department of over a total period of 3 years.
Radiotherapy, SGPGIMS, came into existence in 1987

and has ever since been serving and providing quality

oncology services to the state of Uttar Pradesh,
Uttaranchal, Madhya Pradesh, Bihar and the adjoining
regions including Nepal and Bangladesh. The
Department with its present facilities and those coming
up in the next few months with the state-of-the-art
facilities namely high energy linear accelerator, micro-
multileaf collimator, multislice CT scanner, could play

23
Application and Utility be feasible at other centres, these patients could be
directly pooled and referred to SGPGIMS.
The tele-radiotherapy link would focus on the
following areas: Similarly, patients requiring brachytherapy
either during or after teletherapy - in form of
Clinical services interstitial, intracavitary, intraluminal, moulds etc.
could be evaluated prior to the start of treatment
Tele-consultation: Patients attending the through tele-consultation. This would help to avoid
various departments of radiotherapy could be unnecessary delays and patient inconvenience at a later
considered for a tele-consultation at a mutually agreed date. These patients could be therefore be given prior
scheduled time wherein the treatment policies could be appointments at SGPGIMS for the specialized
discussed based on the exchange of clinical details, treatment, if these are not available at the referring
radiological images and other reports. Live Medical College.
consultation should be feasible in case the clinicians
would like to see and interact with the patient. Tele-follow up: Follow–up consists of an
important component of cancer management and needs
to be conducted throughout the remaining life period of
Radiation therapy planning for external beam
the cancer patient. Even though, the patient may
radiotherapy and brachytherapy: This could involve a
continue to have no disease related morbidity at a given
detailed discussion on the radiation therapy delivery,
point, but since the patient always carries a finite risk
treatment planning, plan evaluation, simulation film of disease relapse locally or at a distant site, a periodic
verification and on line verification. In case the patient follow up is essential. Even though this would be
requires treatment using linear accelerator with carried out by the referring department, these could be
specialized techniques like three dimensional discussed through the tele-radiotherapy network in case
conformal, stereotactic radiotherapy or radiosurgery, patient develops any untoward and unexpected
intensity modulated radiation therapy which may not problem following the primary treatment.

24
Teaching and training reviews could be conducted on line through video-
conferencing.
Virtual class room: The tele-radiotherapy
linkage between the various departments of Central warehousing of the data. This would
radiotherapy would help in developing a virtual class include processing, quality control and review for
room concept whereby one can formulate a common publication. All participating institutions could be co-
teaching programme as per the predefined course authors of such publication.
syllabus for the postgraduate residents. The faculty
from the various centres could be identified based on Periodic review of the trials through video-
their expertise and interest and these interactive conferencing. To ensure strict compliance, the progress
teaching could be beamed on to as many centres as of the trials could be periodically reviewed by the
permitted and supported by the hardware. group and take effective steps to make certain smooth
conduct of the trails as designed in the protocols.
Tele-conferencing and workshops: The tele-
radiotherapy network between these centres would also Such coordinated activities are very much
help in tele-conferencing and enable the staff members required to frame treatment protocols that would be
to take part in workshops. This would help to involve feasible to carry out looking into the various
all the members to take part in such conferences and infrastructure facilities available at various centres of
workshops that could be organized from time to time the State and also taking into consideration the
focusing on a specific topic. Online transmission of the financial implications of the treatment offered. This
procedures like brachytherapy should be of great help would be first step towards the formation of a UP
to the residents, technicians and nursing staff of those Radiotherapy Oncology Group (UPRTOG).
departments who don’t have brachytherapy treatment
facilities. This would help in the training of such
Salient technical features of the proposed
personnel to a wider application of the modern
teleradiotherapy network
radiotherapy procedures.
In the first phase, the SGPGIMS and KGMU would be
Research
linked, while in the next phase, the other departments
of Radiotherapy would be tele-linked with SGPGIMS.
Research activities could be undertaken in a
The functional features of the system would include:
wide range of disease conditions predominant in this
part of the country. These include cancer cervix, breast, - Multi-point video conferencing from the
head and neck, gastrointestinal malignancies and remotest place for face-to-face consultation.
others. The research activities could be focused on - Live–data transfer through integrated
clinical research and translational research. The telemedicine software.
department of Radiotherapy at SGPGIMS is one of the - Incorporated multi-media.
few centres in the country to have a tumour biology - Telemedicine solutions with interfaces for
laboratory in the department and this could be of various medical equipments.
immense use for translational research. The research - International standards like HL 7 & Dicom 3.0
activities would be therefore consists of:
- Built-in-data and image manager.
Multicentric trials: This could be conducted - Upgradation and maintenance servicing time to
between the participating institutions co-ordinated by time based.
Department of Radiotherapy at SGPGIMS. These - Medical devices like CT, MRI, SPECT,
could be in form of phase III randomized clinical trials radiotherapy treatment planning systems which have
targeted towards the most common cancers seen in the DICOM compatible output or having digital output
State - namely cancer of the head and neck, cancer would be directly integrated to the telemedicine
cervix, cancer gall bladder and cancer breast. system ready for plug and play use.
Sharing of online data and its evaluation and - Special hardware include: Multi Communication
interim analysis: The data generated would be stored Card for connecting multiple devices having digital
online at the data processing centres at SGPGIMS and output to the system, Multi relay for multiple video
could be available to all centres following its switching and controlling it through software,
processing and other statistical analysis. Periodic Medical Image capture card for capturing medical
images from medical

25
 Table 1: System hardware for server and nodes

Telemed 2000 Server Telemed 2000 Node 1 Telemed 2000 embedded node
o Pentium P-IV 2.2 GHz processor o Pentium P-IV 2.2 GHz o Embedded GXI-300 MHz processor
o Windows 2000 Server processor o Bios-Award 256 KB Flash memory
o 810D Intel Chipset motherboard o Windows 2000/Xp o Chipset NS CS5530A
o 256MB RDRAM o 810D Intel Chipset o 128 MB SDRAM
o 80 GB Hard Disk Drive motherboard o Serial/Parallel/USB/Infrared Port
o 52X CD ROM Drive o 128 MB RDRAM o 1~4 MB Share memory
o Graphics Adapter with 32MB o 40 GB Hard Disk Drive o Support Compact Flash Cards
RAM o 52X CD ROM Drive o PCI 10/100 Mbps Ethernet
o Sound card with four channel o Graphics Adapter with o Keyboard, Mouse, Speaker &
o Network Interface Card 10/100 32MB RAM Microphone
Mbps for Networking o Sound card with four o SB Telemed 2000 Telemedicine
o Keyboard, Mouse, Speaker & channel Node Software
Microphone o Network Interface Card o Display unit
o Network Switch 10/100 Mbps for o Video Conferencing System having
o Server Based Telemed 2000 Networking High Resolution PTZ Camera
Integrated Telemedicine License o Keyboard, Mouse, Speaker o 29” Flat Color TV
Software & Microphone o Switcher to attach different cameras
o 50” Plasma Screen o SB Telemed 2000 and Medical Equipments
o Video Conferencing System Telemedicine Node o Inkjet Printer
o Switcher to attach different Software o Cables, connectors and accessories
cameras and Medical o Laser Printer
Equipments o Cables, connectors and
accessories

devices, Communication card for communicating to References

different centers using communication media and Multi
Conferencing Card for multiple conferencing using 1. N. R. Datta, D. Rajasekar, Improvement of
different communication media. radiotherapy facilities in developing countries: a
The communication media would be ISDN, LAN three-tier system with a teleradiotherapy network,
or WAN with 128 kbps or above bandwidth with the The Lancet Oncology, 5(11), 2004, 695-698.
type of data being either electronic medical records 2. J. Ferlay, F. Bary, D.M. Parkin, P. Pisani (eds).
(EMR), vital sign, video conferencing, still & moving, GLOBOCAN 2000; Cancer Incidence and
audio and radio images (in live, store and forward Mortality Worldwide (IARC Cancer Bases No.5,
manner). The details of the system hardware are Lyon, IARC press; 2001).
summarized in Table 1. 3. WHO. Global cancer rates could increase by 50%
to 15 million by 2020. Available from: URL:
Conclusion http://www.who.int/mediacentre/releases/2003/pr2
The proposed UP teleradiotherapy network is a 7/en/.
maiden attempt to integrate the various radiotherapy 4. IAEA. Millions of cancer victims in developing
departments of the state of UP with an endeavour to countries lack access to life saving radiotherapy
improve the quality of patient care, teaching, training, IAEA press release 2003/11, 26 June 2003
manpower development and conduct multicentric Available from:
clinical trials in sites commonly seen in this State. The URL:http://www.iaea.org/NewsCenter/PressRelea
proposal with a teleradiotherapy link up should enable ses/2003/prn0311.shtml.
not only the patients to have access to the highest state
of cancer care, but is also built on a philosophy of 5. H. Ashraf. Poor nation need more help to slow
sharing the expertise and facilities in a specified growing cancer burden. Lancet 2003; 361: 2209.
geographical area and provide a comprehensive cancer 6. K.A. Dinshaw. Addressing the cancer agenda: how
care within the limited resources. have we measured up? In: S.P. Aggarwal, Y.N.
Rao, S. Gupta (eds). Fifty Years of Cancer Control
Acknowledment in India. National Cancer Control Programme,
Directorate General of Health Services, Ministry
The UP teleradiotherapy network project has of Health and Family Welfare, Government of
been funded by the Department of Science and India, 2002, 6 – 13.
Technology, Government of India.

26
 EVIDENCE OF BENEFIT FROM TELEMEDICINE.
Dr. Sanjeev Garg, Dr.Vinod Gupta, Dr. A.Raghuvanshi.
Rabindranath Tagore International Institute of Cardiac Sciences
[A Unit of Asia Heart Foundation]
124, Mukundapur, EM Bypass
Kolkata – 700 099, India
drgargsanjeev@rediffmail.com.

Abstract referred to the Tele-CCU. There he is examined by the

doctor, who after taking history, records an ECG,
Telemedicine is a new concept to provide quality which is then sent to Kolkata or Bangalore. A specialist
medical care to people at remote places. RTIICS at the hub then gives his advice, which is executed at
initiative was to treat patients with various cardiac the remote Centre.
emergencies in remote places through setting up
CCU’s in these places and linking them to Kolkata and A total of 3198 patients with cardiac emergencies have
Bangalore. In the last 3 years, more than 3000 patients been treated successfully in the CCU’s of Siliguri,
with cardiac emergencies have been treated with Bankura, Udaipur and Tinsukia. A majority of these
significant reduction in mortality and morbidity. patients presented with Acute Myocardial Infarction. A
total of 388 patients were successfully thrombolysed at
Keywords these Centres with minimal complications. Serious
complications like Tachyarrhythmia were also
Telemedicine, Cardiac emergencies, successfully reverted to Sinus Rhythm. [2% of total
Tele-CCU’s, Acute Myocardial Infarction, admissions have these complications] The major
Tachyarrhythmia. reason for patients not receiving timely thrombolysis
was late presentation.

Introduction
Conclusion
Killip et al in 1967 advocated the role of C.C.U. in
management of cardiac emergencies in reducing Treatment of cardiac emergencies requires properly
mortality. India has a vast population and an ever- equipped and adequately manned CCU. With proper
increasing rate of cardiac aliments. Are we equipped care and planning, we can provide medical care to our
enough to handle this increasing load? Most of the vast population with minimum constraint on our
district, sub divisional hospitals and remote places limited resources.
either lack CCU’s or have ill-equipped ones. Mortality and Morbidity in these CCUs are comparable
AHF initiated a novel project of setting up properly to those in Kolkata and Bangalore .
equipped CCU’s in remote places telelinked to Kolkata
and Bangalore.
Acknowledgement
Body of paper
Asia Heart Foundation, ISRO, Govt. of West Bengal,
Any person presenting at the emergency department Tripura and Karnataka.
with symptoms of cardiac disease is screened and

27
ASSISTIVE TECHNOLOGY

28
 THE BOMBAY-VELLORE ARTIFICIAL HAND – RECENT DEVELOPMENTS
Suresh Devasahayam and Suranjan Bhattacharji
Department of Bioengineering and Department of Physical Medicine and Rehabilitation
Christian Medical College
Vellore 632002, Tamil Nadu, India
surdev@cmcvellore.ac.in, s_bhattacharji@cmcvellore.ac.in

Abstract Electrically powered prostheses have been designed

to overcome the shortcomings of mechanical
No prosthetic hand available today can fully replace prostheses. Most commercially available electrically
trhe motor and sensory functions of a natural human powered prostheses are expensive and most models are
hand. The number of upper limb amputations in India not well tropicalised and tend to malfunction in hot and
is estimated in the tens of thousands, and there is a humid conditions, especially when sweating occurs.
pressing need for a reliable, affordable prosthetic hand. In designing an artificial hand it is useful to
A low-cost prosthetic hand has been developed at IIT- consider the functions of a normal human hand in order
Bombay and CMC-Vellore. It has been tested on 32 to prioritise the functions to be included in the design
amputees in the last 2 years. The current design of the prosthetic hand. Fig. 1 shows a simplified
comprises, finger grasp/prehension, forearm rotation schematic of a natural hand, emphasising the feedback
and elbow flexion/extension. Each of these functions nature of the control involved. The motor output of the
uses a DC motor driven by MOSFET H-bridges hand is obvious. We often take for granted the sensory
controlled by a microcontroller. The hand is operated inputs that help us to control the hand for a variety of
by means of micro-switches placed inside the socket tasks. Indeed the sensory feedback enables us to use
fitted over the residual limb of the amputee. The our hands with little or no visual attention. This is in
switches are pressed by subtle muscle movement inside sharp contrast to the use of any machine, like a motor
the socket. One switch is used for selecting the vehicle, for example, which requires so much visual
function and two other switches are used to operate the and auditory attention as to make its operation an
selected motor. This allows very fast response of the exclusive task for all practical purposes. Sensation of
hand to the user commands. These three switches can touch, pressure, joint position, etc., are so refined as to
be used to control two artificial hands in a bilateral enable us to do a variety of tasks ranging from lifting
amputee. heavy objects to very fine tasks.
The initial version of the hand consisting of only Fig. 1 Schematic of musculoskeletal
forearm and finger control has been fitted and tested on control
32 amputees at CMC-Vellore and studied over a period B r a in
of 3-18 months. The current version of the hand is
expected to be fitted on amputees in early 2005. S p in a l

+ cord
M u s c le s Loa d
Key Words
Tou c h
Artificial hand, prosthetic, amputees, low-cost hand, s e n s a t io n
embedded control.
P o s it io n
s e n s a t io n
1. Introduction
The loss of a hand decreases the ability to carry out Prosthetic hand designs in laboratories range from
activities of daily living, impairs functioning at school sophisticated space-age robotic hands to simple
and work, makes social interactions awkward and devices. Among the commercial devices the most
overall reduces the quality of life. The options commonly used are the Otto Bock myoelectric hand
available to upper limb amputees are rather limited. made by Otto Bock, Germany, and the Boston elbow
Many amputees settle for a simply cosmetic prosthesis. made by Liberty Mutual, USA [1]. These have
Functional prostheses are mostly mechanical and are embedded electronics and small batteries. A three-
moved by cables powered by the user's body. These are finger prosthetic hand with miniature electromagnetic
cumbersome and difficult to use especially for children motors has been described by Carrozza, et al [2].
and elderly people. Comparison of different artificial hands is usually done
by measuring forces at different points on the fingers,

29
during, lateral pinch, power grasp, and precision grip electronics, (c) rechargeable and easy to replace
[3]. batteries, (d) low cost, and (e) ease of use.
Sensory feedback is an important aspect of research.
Pressure sensors, and vibration detection based slip 2.1 Mechanical design
sensor have been tested. After the electronic detection
of these physical parameters the information has to be The artificial hand uses commutated (brushed) DC
transferred to the user in an unobtrusive manner. For motors for each movement. These motors were chosen
this purpose vibration transducers, low level electrical on account of their easy availability and reasonable
stimulators and auditory buzzers have been used. A reliability. The motors are commercially available with
novel method recently described has been the use of an integrated gear box (ratio=1:80) that give a final
interferential current stimulation to impart sensation to shaft speed of 24 rpm. Such motors are used in a
the residual limb. Two sinusoidal currents, one at variety of applications like automobile wipers,
5KHz and the second varying from 2KHz to 4.9KHz airconditioner vane oscillators, etc. Therefore, they are
gave a net interference frequency ranging from 100Hz widely available at very low cost. One motor each is
to 3KHz. The stimulation was modulated by force used for the finger grasp/prehension, forearm
sensors picking up the force on the load. This pronation/supination, and elbow flexion/extension. The
reportedly gave a biomimetic feedback, giving the finger grasp/prehension motor, M1, is coupled via a
feeling of increasing compliance with decreasing pair of bevelled gears to drive the pincer-like finger
mechanical force [4]. movement. The forearm motor shaft supports motor
M1 with the fingers. Therefore, operation of motor M2,
The control of the artificial hand by the user is a very
directly rotates the hand. The third motor, M3, drives
important aspect and it has received a lot of attention.
the movement at the elbow, by a worm gear
The most common control signal used is the
arrangement. This gear arrangement, allows the load to
myoelectric signal picked up from the residual
be taken by a supporting structure (not shown in the
muscles. The myoelectric signal is rectified and
figure) when the elbow is not being driven. This
averaged before thresholding in the simple schemes.
protects the motor's internal gears from excessive
More sophisticated processing techniques have been
stress.
proposed, including time series modeling [5] and
neural network based systems. Graupe used
autoregressive (AR) models of the surface EMG, and The current version of the hand uses an internal gear
reported that 4 AR parameters are sufficient. These
parameters are then mapped into decision space for Fig. 2 Schematic of the mechanism
multi-function prosthesis control [6, 7]. Huang et al
have described a neural network based system using
features obtained from variance, bias, zero-crossings,
autoregressive model and spectral estimation. The
features are extracted during simulated eight types of M3
hand movements, such as three-jaw chuck, lateral
hand, hook grasp, power grasp, cylindrical grasp, M2 M1
centralized grip, flattened hand and wrist flexion [8].
Keeping in mind that disabled people have reduced ratio of 1:60 for M1 and M3, and a ratio of 1:120 for
earning capacity, the cost of any prosthetic device is an the forearm. The fingers have an additional external
important factor. Low cost, robust performance, and gear
ease of maintenance were important considerations in
the design of the artificial hand described here. ration of 1:5, and the elbow has an additional gear ratio
Prompted by a felt need at the All India Institute for of 1:12 (fig.2). The time for full excursion of the
Physical Medicine and Rehabilitation, Haji Ali, fingers is about 5s, of the forearm about 2s, and the
Mumbai, a motorized prosthetic hand was designed at elbow 12s.
the Indian Institute of Technology, Bombay. The
Bombay-Vellore hand initially designed at the Indian 2.2 Control electronics
Insitute of Technology, Bombay, was subjected to The three motors are controlled by MOSFET H-
patient trials and substantial modifications at the bridges, which are in turn controlled by an 8-bit
Christian Medical College, Vellore. Production microcontroller (Microchip). The microcontroller
versions of the Bombay-Vellore hand were receives digital control signals from three switches.
manufactured by WORTH Trust, Katpadi, Tamil Nadu. The microcontroller also monitors the current drained
from the battery. The current is monitored by readng
2. Design of the Artificial Hand the voltage across a 0.5 ohm resistor into one of the
A/D ports in the microcontroller. If the current exceeds
The main design constraints were: (a) simple and a threshold for a time greater than 2s, the motors are
robust mechanical design, (b) high reliability turned off. The 2s time interval is to allow high starting

30
current and temporary high load. By changing the loop, with four functions called. The first function is a
current limit and the time limit the strength of the grip timer call to ensure that each loop time is exactly 50ms.
can be adjusted – this is currently fixed at a single The second function polls the three switches, S1, S2,
value. The following figure (fig.3) shows a block and S3. The third function reads the A/D #1 to get the
schematic of the control electronics. battery voltage. The fourth function reads A/D #0
Fig. 3 Schematic of the electronics (current).
For training purposes, the micrcontroller also pulses
M3 M2 M1 two tricolour LEDs, one for the function selected
(red=fingers, orange=forearm, green=elbow), and the
second LED for the battery level (green=okay,
3 x M otor orange=medium, red=low battery, charge
c o n t r o lle r s
( H -b r id g e ) immediately). This section of the code is not shown in
S1 L i-I o n
the listing for the sake of simplicity.
8 -b it b attery
S2
m ic r o c o n t r o l l e r
If the overcurrent condition is satisfied (current
exceeds “Imax” for a duration greater than “Tmin”),
S3 A/D # 1 A/D # 1 C u r r e n t
sen se Batt then the motors are turned OFF, and the switches have
le v e l to be released before any motor can be operated again.
This allows the motor to be operated in the same
direction if desired. Unlike cycling sequential schemes
2.3 Embedded Software where operations cycle between clockwise (CW) and
counter-clockwise (CCW) rotation of the motors, this
The micrcocontroller program is shown in pseudo- scheme ensures very quick response from the artificial
code in the listing below. Since timing plays a crucial hand. In this control scheme, only one switch is used to
role in the control of the hand, the built-in timer select any particular function (any one motor). The
interrupt is used. The main program is an indefinite advantage of this is that even if the number of residual
// Initialize muscles is small (as in the case of a bilateral above
Setup clock interrupt elbow amputee, for example) a large number of motors
Setup I/O ports can still be operated. The use of two separate switches
for CW and CCW rotation of the selected motor
// Loop indefinitely ensures very quick response of the hand. Although, the
Loop: quick response is possible only for one particular motor
Call Timer_Interval in such a scheme, it is found that this is acceptable to
Call Switch_Press most users. Of course, if more switches can be used,
Call Battery_check then multiple parallel operations can be incorporated.
if any switch pressed: Call Current_check
goto Loop
2.4 Clinical Trials
Switch_Press:
Check Switch 1 Thirty two patients were fitted at CMC-Vellore.
if true: Motor CW
Three of these patients were bilateral amputees who
Check Switch 2
received two hands each. Thus a total of 35 hands were
if true: Motor CCW
Check Switch 3: fitted in CMC. After fitting, each patient was requested
if true: Check duration T to undergo 2-3 weeks of physiotherapy and
if (T<T1) select M1 occupational therapy before going home. This was to
if(T1<T<T2) select M2 ensure that they were properly trained in using the
if(T>T2) select M3 artificial hand. The Occupational Therapists evaluated
return the patients in Activities of Daily Living (ADL) before
discharging them from the hospital. Subsequently, the
Battery_check: prosthetic team paid several visits to the patients'
Read A/D #1, set battery level LED (output pin) homes to evaluate the extent of use at home.
The patients whose quality of life experienced the
Current_check: greatest improvement were the bilateral amputees.
Read A/D #0 Next, were the above elbow amputees. Only two
if (value>Imax) start time_count among the 32 patients were found to be not using the
if (value<Imax) reset time_count hand during the home visit. All the others were using it
if (time_count>T1) set flag & stop motors regularly. Regular functional use was observed in most
if (flag) wait for switch release of the patients. The following two pictures show an
return open view of the artificial hand (left) and an above
elbow amputee wearing the artificial hand with
Timer_Interval:
cosmetic covering.
wait for one timer interrupt completion
return
31
 References
[1] J.G.Webster, A.M.Cook, W.J.Tompkins and
G.C.Vanderheiden, Electronic Devices for
Rehabilitation (Chapman & Hall, London, 1985)

[2] M.C. Carrozza, B. Massa, S. Micera, R.

Lazzarini, M. Zecca, P. Dario, The development
of a novel prosthetic hand - ongoing research and
preliminary results., IEEE/ASME Trans
Mechatronics, Volume: 7 , Issue: 2 , June 2002 ,
108 - 114

[3] F. Sebastiani, S. Roccella, F. Vecchi, M.C.

3. Conclusion Carrozza, P. Dario, Experimental analysis and
performance comparison of three different
The most common source of failure of the artificial prosthetic hands designed according to a
hand was by breaking wires. Next was damage to the biomechatronic approach. Proc. IEEE/ASME Intl
internal gears. The wire are now secured more firmly Conf Advanced Intelligent Mechatronics,
inside the shell. Different motor manufacturers have Volume: 1 , July 2003, 64 – 69.
been evaluated by WORTH Trust and a new set of
motors have been identified. In preliminary trials these [4] Y. Sasaki, Y. Nakayama, M. Yoshida, Sensory
motors appear much more reliable. They are awaiting feedback system using interferential current for
full user trials. EMG prosthetic hand. Proc. 24th Annual IEEE
The most important enhancement desired by Conf EMBS [Second Joint EMBS/BMES Conf],
amputees after the elbow flexion, is wrist movement – 2002. , Volume: 3, Oct. 2002, 2402 – 2403.
specifically, wrist flexion/extension. The next is some
sensory feedback of grip strength, and accompanying [5] D. Graupe, Time Series Analysis, Identification
modulation of grip force by the user. and Adaptive Filtering (Krieger Publishing,
Florida, 1984).
4. Acknowledgements
[6] D Graupe, EMG Pattern Analysis for Patient
We would like to acknowledge our colleagues Prof. Responsive Control of FES in Paraplegics for
Rakesh Lal and Prof. Pandey of Indian Institute of Walker Supported Walking, IEEE Trans. BME,
Technology, Bombay, who have been involved in the vol.36, No.7., 711-719, July 1989.
design and testing of the artificial hand from the
beginning. [7] D.Graupe, J.Magnussen & A.A.Beck, A
microprocessor system for multifunctional
The invaluable work of the prosthetists, control of upper limb prostheses via EMG signal
physiotherapists and occupational therapists in the identification. IEEE Trans. Autom. Cont.,
fitting, training and evaluation of the artificial hand is vol.AC-23, pp.538-544, 1978.
gratefully acknowledged.
[8] Han-Pang Huang; Chun-Yen Chen, Development
Funding sources for this project include: (a) The
of a myoelectric discrimination system for a
Ministry of Social Justice and Empowerment, India, (b)
multi-degree prosthetic hand, Proc. IEEE Intl
IRCC, Indian Institute of Technology, Bombay, (c)
Conf Robotics and Automation, May 1999,
WORTH Trust, Katpadi, and (d) PMR Special Fund,
Volume: 3, 2392 – 2397.
Christian Medical College, Vellore.

32
 BIOFEEDBACK TRAINING FOR THE USE OF ASSISTIVE DEVICES
Suresh Devasahayam and George Tharion
Department of Bioengineering and Department of Physical Medicine and Rehabilitation
Christian Medical College
Vellore 632002, Tamil Nadu
India
surdev@cmcvellore.ac.in, tharion@cmcvellore.ac.in

Abstract functions in new ways. This is rather like learning to

drive a motor vehicle. Unless sufficient proficiency in
Biofeedback system have been widely used for using these assistive devices is achieved, the user will
relaxation, sports training, rehabilitation in disability, not be able to use them to their full potential, and will
etc. In this paper we describe an EMG biofeedback quickly become frustrated and discontinue using the
system for training disabled people to use assistive devices.
devices.
In the case of traumatic accidents resulting in the
There are two main classes of patients who benefit loss of movement or sensation, there is usually a long
from biofeedback training for the use of assistive interval of hospitalization for wound healing and
devices. The first are amputees using powered recovery from the injury. The actual motor loss
prosthetic hands. The second is the set of severely resulting from neurological damage is often
disabled patients who need to use communication exacerbated by the long period of acute illness which
assistance devices to partially compensate for their may result in weakened muscles and poor motor
disability. Both groups of people use residual muscle control. Biofeedback can be used to restore the motor
movements to transfer information to their assistive function closer to its potential. Biofeedback is a
devices. The better the strength and control of these collection of methods to amplify sensory awareness of
residual muscles, the better is the SNR of the various body functions [1]. It is a form of operant
information picked up for control of the assistive conditioning that is used to suppress or enhance
devices. Therefore, biofeedback based reinforcement is physiological responses of which the user is not usually
useful. consciously aware [2]. In this paper we discuss the use
The system described here was designed to use a of biofeedback specifically for the purpose of
standard PC with minimum additional hardware. The reinforcing and strengthening motor function in
software is written in the Java programming language disabled patients who use various assistive devices.
and can run as an applet in a Java compatible web- The following figure shows a schematic of a general
browser; it is thus cross-platform. The EMG signals are biofeedback system.
digitized by the PC sound-card. Two channels of EMG
D is p la y
can be picked up and digitized in the computer. The
EMG is then processed and visual and auditory Fig. 4 General schematic of a biofeedback system
Au d io
feedback is given to the user. For young children a
simple video game was devised to sustain interest and Lo g ic
improve the performance of the biofeedback system. Ta c t ile

Key Words Tr a n s d u c e r s /
se n sor s
Biofeedback, physically disabled, EMG, Java,
rehabilitation, assistive devices.

1. Introduction The biofeedback system consists of a set of

transducers and sensors which pickup the physiological
Severe disablity following various kinds of spinal activity to be reinforced. This may be a single
cord injury and brain injury is a combination of motor, physiological parameter or multiple parameters related
sensory and cognitive impairment. Several to a certain task. The signals are pre-processed
technological solutions are available to compensate for (filtered, amplified) and then further processed by an
the motor and sensory loss. The interface between the analogue system (Schmitt triggers, comparators, etc.),
user and the assistive device is an artificial one and digital system (digitization, boolean logic), or a hybrid
therefore imperfect. The user needs to learn to use the system. The output of the logic operations is delivered
assistive device using the available senses and motor to the subject by means of visual inputs (lights, video

33
display), audio inputs (buzzer, computer speaker) or
tactile inputs (vibrator, stimulator).
Biofeedback has been reported in the literature to be
In contrast to amputees, severely disabled people
used for the rehabilitation of physically disabled
with paralysis have very limited motor control and may
patients in a variety of ways [3]. Biofeedback of the
be able to move only a finger or foot. This small
pressure from sphincters using a pressure transducer
voluntary movement is used to control a
has been used for rehabilitation of spastic anal
communication device to interact with the external
sphincters [4]. They measured pressure using thin
world. Frequently, patients with spinal cord or brain
piezo-electric sensors placed on a probe with EMG
injury have spastic muscles which have uncontrolled
(electromyogram) electrodes. This probe directly
spasms of contraction. Any voluntary control present in
measured rectal pressure and therefore proved to be
such muscles cannot be used unless the spasticity is
better than using the EMG. The EMG or myoelectric
reduced. The sensory input of such disabled people is
signal, is the most common signal used in biofeedback
also usually limited. Therefore it is very important to
systems. The EMG may be regarded as a zero mean
maximise the control and strength in this residual
stochastic process. The energy in the EMG is non-
movement.
linearly related to the muscle force. Most EMG
biofeedback systems use the mean value of the The biofeedback system described below has been
rectified EMG as the biofeedback parameter. designed for training such patients at CMC-Vellore.
Occasionally more complex features like the median
spectral frequency, zero-crossings, turns/sec, etc., have 3. Design and Testing
been used [5]. EMG biofeedback can be used not only
for muscle strength training, but also for training in the The biofeedback system was designed around a
performance of co-ordinated higher complexity tasks. personal computer, using open source software. The
It has been used for gait training where several muscles open design is useful for further development and low
must work in harmony. Joint position or muscle force cost. This section describes the hardware and software
can also be used for movement training. For example, design, as well as the clinical testing of the biofeedback
biofeedback of joint position has been used to correct system.
genu recurvatum in adult stroke patients [6]
2.1 Hardware Design
Computer games have been frequently used to
sustain interest and improve biofeedback performance
In order to keep the system as portable as possible,
[7]. An interesting computer game driven by a
the hardware is kept to a minimum, and signal
biofeedback system will remove the boredom of
digitization is done using the sound-card in the
repetitive training, and works especially well for
computer. This has two channels in most computers
children.
(line-in input), and has AC coupled inputs (high pass
Biofeedback can be used not only for strength filtered at about 15Hz) in order to keep the DC offset
training by reinforcing increased muscle activity, but it problems to a minimum. Therefore, low frequency
can also be used to train relaxation in spastic muscles. signals cannot be digitized, but the surface EMG which
A biofeedback system to reinforce movement of has signal power in the band 15 Hz to 600Hz can be
children with cerebral palsy has been reported [8]. easily digitized. The EMG amplifiers consist of single
Biofeedback has been used to train patients with facial chip instrumentation amplifiers, followed by band-pass
palsy to improve their control of facial muscles, filters; the overall gain is 400. The electronics is
asthmatic children to improve diaphragm control, etc. powered by a single 9V battery with a switched
[9]. Biofeedback has been used for training paralysed capacitor circuit to obtain a negative supply for full
patients to use communication devices [3]. bipolar operation of the electronics.
There are two main classes of assistive devices used In order to comply with medical safety standards, the
by patients at the Rehabilitation Institute in CMC- electronics in electrical contact with the user/patient
Vellore that benefit from the use of biofeedback must be electrically isolated from the mains powered
training. The first is a motorized prosthetic hand (the computer. This can be done using an isolation amplifier
Bombay-Vellore Hand), and the second is a set of and a second battery to power the isolated side. An
communication devices developed at CMC for patients alternative method of achieving isolation is use a radio
with severe motor and sensory loss. link between the battery powered EMG amplifiers and
the computer. We used a commercial wireless stereo
Amputees who are fitted with the motorized
headphones transmitter/receiver for this purpose in one
prosthetic hand use muscles in the residual limb above
of our systems. This also makes the system very
the amputation to control the electronics for the
convenient to use as it reduces the clutter of wires
prosthesis. These movements for switch activation
around the user/patient. Since the voltage levels,
have to be consistent and as automatic as possible in
bandwidth, etc., of the wireless headphones are
order to achieve a near-natural control of the prosthetic
hand.

34
designed to be compatible with the audio input of the adult biofeedback training. Mode 3 is used for special
computer, the interfacing is straightforward. cases and research experiments where EMG spectral
s p e a k e r (s )
changes are expected to occur with fatigue, muscle
Fig. 5 Schematic of the hardware
fiber type, etc. Mode 4 is used for biofeedback training
EMG
preamp
of children.
In mode 1, the raw signal from both channels is
F ilt e r s Au d io displayed continuously on the screen. The display
lin e -in sweep duration can be varied for convenience. The old
s u r fa c e
data is shifted to the left and the new data is drawn on
e le c t r od e s
P ow e r the right of the screen giving the impression of
s u p p ly continuously scrolling paper.
In mode 2, the EMG signal is rectified and averaged
in blocks of 200ms (continuously sliding window) and
F ilt e r s displayed. The displayed trace is again scrolled
continuously. A threshold can be set for each EMG
EMG channel. For muscle strength training, the average
pream p EMG must exceed the threshold, and for relaxation
training the EMG must be less than the threshold.
Combinations of contraction and relaxation can be used
for training. When the set criterion is met, the
Further processing of the EMG, like rectification, biofeedback user is given both visual and auditory
average value calculation, etc., traditionally done in notifications. The visual feedback has a flashing image
hardware have been implemented in software. of a light bulb (in addition to the waveform trace). The
The computer used in the setup is a standard auditory feedback is a beep output on the PC speakers.
multimedia personal computer (circa 2000). All the The box below shows a brief outline of the program
audio input and output use the sound-card A/D and written in pseudo-code.
D/A converters as well as the multimedia speakers.
Sampling rates available are 8000Hz, 11025 Hz, 22050 Initialize
Hz and 44100 Hz. For surface EMG, 8 kHz is setup AudioLineIn Input
adequate. Although sampling frequencies as low as 2 Build GUI
kHz can be used, most PC sound cards do not allow the Start input on 2 audio chans, sampling rate=8KHz
use of arbitrary sampling rates.
Thread Run
2.2 Software Design Read AudioIn data buffer into arrays
Calculate moving window average
The software uses the Java programming platform Shift old data and write current data to display
which is cross-platform. It has been used on both Calculate FFT, add magnitude to weighted array
Microsoft Windows 2000 machines as well as Compare signal levels with threshold: conditions?
computers with Linux (Mandrake 9.2, kernel 2.4.22). if (conditions=true)
advance game display
if (endurance time) “congratulation”
output audio beep
output “bulb” image flash
return

Display:
Test display selection:
if (EMG trace) display average EMG
else if (spectrum) display STFT
else if (game) display “rocket” new position
if (endurance time) display message
return
end
Fig. 6 Screenshot of the program in "Game" mode In mode 3, the spectrum of the raw EMG is
displayed. It is continuously updated every 500ms. In
order to improve the spectrum estimation, an
There are four modes of operation: (1) Raw EMG,
exponentially weighted average of the Fourier
(2) average EMG, (3) EMG spectrum and (4) Rocket
magnitude is used.
Game. Mode 1 is only for initial testing of the signal
quality by the therapist. Mode 2 is used for normal

35
 Mode 4 has a simple computer game in which a
rocket moves across the screen from left to right as acknowledge the work in Rehabilitation Engineering
long as the muscle contraction/relaxation criteria are done by Mr Sivasakthi Murugan in making the
met. In 15 seconds the rocket reaches the goal, and a communication systems used by the patients.
rewarding message is displayed and output on the
speakers. References
2.3 Muscle Strength Training [1] J.D.DeLisa & B.M.Gans (ed.), Rehabilitation
Medicine: Principles and Practice (Lippincott-
Two copies of the biofeedback system described Raven, Philadelphia, 1998)
above has been fabricated and installed in the
occupational therapy unit of the department of Physical [2] E.R.Kandel, J.H.Schwartz, T.M.Jessell,
Medicine and Rehabilitation at CMC. They are being Essentials of Neural Science (Prentice Hall
used in the rehabilitation of patients with spinal cord International, New Delhi, 1995).
and brain injury. They are also used to train children
with cerebral palsy to achieve better control of their [3] J.G.Webster, A.M.Cook, W.J.Tompkins and
muscles. G.C.Vanderheiden, Electronic Devices for
Rehabilitation (Chapman & Hall, London, 1985)
3. Conclusion [4] R. Magjarevic, B. Ferek-Petric, K. Lopandic,,
The clinicians using this biofeedback system have Biofeedback in rehabilitation of anal sphincter
expressed a need to have additional modalities for muscles, Proc. 22nd IEEE Annual Intl Conf
training. The most important are muscle force EMBS, July 2000, Volume: 1 , 423 – 426.
biofeedback, sphincter pressure biofeedback and
bladder pressure biofeedback. [5] J.J. Im, D.H. Rho, Y.J. Jeon, N.B. Lee, J.I.
These modalities all need to have steady levels Chung, Extraction of parameters from EMG
monitored during the biofeedback. Therefore, the signals for the biofeedback electrical stimulation.
analogue electronics has to process DC signals up to Proc. 24th Annual Conf EMBS, Second Joint
the digitization stage. This can be done by using EMBS/BMES Conf., 2002. , Volume: 1, 133 –
dedicated data acquisition computer add-on cards with 134.
A/D converters. This will increase the cost of the
hardware substantially. An alternative that has been [6] E. Dursun, E. Ceceli, A. Cakci, O. Tuzunalp, O.
tried in our laboratory is to frequency modulate (FM) Ozturk, Z. Telatar, Proceedings of the 1992
the low frequency signal and then input it to the PC International IEEE Biomedical Engineering Days,
sound card. The demodulation of the FM signal in 1992: 246 – 248.
software will recover the original signal with low
frequencies extending to DC. A problem that must be [7] G.M. Lyons, P. Sharma, M. Baker, S. O'Malley,
kept in mind with all DC signal circuits is the offset. A. Shanahan, A computer game-based EMG
Imbalances in the initial value of the force or pressure biofeedback system for muscle rehabilitation,
transducers causing DC offsets can often cause Proc 25th IEEE Annual Intl Conf EMBS, Sept.
amplifier saturation at high gain levels. The DC offset 2003, Volume: 2, 1625 – 1628.
can also be due to the resting force or pressure level.
Therefore, removal of the initial DC level is important. [8] Yu-Luen Chen, Chia-Ling Chen, W.H. Chang,
Automatic DC level adjustment circuits are very May-Kuen Wong, Fuk-Tan Tung; Te-Son Kuo,
useful. An analogue DC offset adjustment circuit uses a The development of a biofeedback training
sample-and-hold circuit to take the DC value at the
system for cognitive rehabilitation in cerebral
beginning and then subtract it during the rest of the
palsy, Proc 19th IEEE Annual Intl Conf EMBS,
session. Sample and hold circuits are usually designed
Oct. 1997, Volume: 5 , 1919 – 1920.
to hold only for periods of a few seconds at the most. A
digital “sample and hold” can be devised using an A/D,
latch and D/A in hardware. This system after complete [9] L.C. Carvalho, H.F. Albuquerque, C. Pontes,
laboratory testing will be given for clinical trials. M.T. Maia, D.P.R. Mangueira, L.V. Batista,
Computerized biofeedback tool: application in
4. Acknowledgements electromyogram-biofeedback, Proc. 25th Annual
IEEE Intl Conf EMBS, Sep. 2003, Volume: 2 ,
The enthusiasm of the physicians and occupational 1609 – 1612.
therapists in the department of Physical Medicine &
Rehabilitation in using the system and giving valuable
suggestions is gratefully acknowledged. We also

36
 SINGLE KEY OMNI DIRECTIONAL POINTING AND COMMAND SYSTEM (SKOPS)
-A SMART ON SCREEN NAVIGATIONAL TOOL FOR PHYSICALLY DISABLED
PERSONS
Suman Deb* , Sujay Deb#
Dept. of Computer Science & Eng., Indian Institute of Technology Kharagpur*
Dept. of Electronics & Communication Eng., NERIST#
sumand@cse.iitkgp.ernet.in, sujayne03@rediffmail.com

Abstract establishment of basic protocols governing on-screen

object selection and manipulation.
This paper describes an inexpensive computer
access system for physically disabled persons. It is SKOPS operates: under the Windows operating system
based on micro switches rather than electro- and can be used in conjunction with any application
oculography (EOG) or very expensive reflectance software. When combined with an on-screen keyboard,
based method. We have built the hardware and the user has complete accessibility to Windows
software to demonstrate the variability of specialized features. The system is designed to run in the
very low cost micro switches and software precision background while the user is working with standard
for human–computer communication. Our experiments Windows applications. Voice dictation software can
indicate that SKOPS provides the basis for an adequate also be used with the SKOP System resulting in
input interaction device supplementing the need of both efficient, hands-free operation for both text entry and
keyboard and a pointing device with a single access pointing operations.
key. Being very inexpensive and versatile the system is Accuracy: Cursor placement on toolbars and menus is
applicable for persons who cannot use a hand-operated consistent and easy. At most two-step translation can
mouse, including those who are experiencing soreness select any arbitrary point on the screen. To ensure
in their hand or wrist to those who have no hand more accuracy on selecting some small and quick
function at all. It can also be used for many virtual launch icons a zonal cursor movement speed control
reality system as well as video games and algorithm is also incorporated. It automatically slows
entertainment. down the pointer speed in some predefined screen
zones, e.g at task bar, windows minimize, maximize
Key Words and close button etc at the upper right corner of
Assistive Technology, access system, alternative window which needs some precision pointing.
communication, single key pointing system, mouse
supplementary. Customizability: The SKOPS program has all it’s
parameters adaptable. These can be set by the user or a
1. Introduction constant performance-monitoring algorithm can also
tune the strictures to ensure optimal performance. Not
Many computer interface hardware and software only the software but the hardware is also amendable
designs, such as mouse and touch screens are a nice according to an individuals requirement, e.g. some may
improvement over keyboard for some tasks, but cannot operate the same device with tongue other may operate
be utilized by quadriplegics[12]. Yet, it is the with toe with a very next to nothing change.
physically disabled that have the most to gain from and
the greatest dependence on computer and electronic Internet accessibility using SKOP System: The user
aids for work, recreation, environmental control, and can independently go on-line, and browse up to any
even for the most basic communication needs. extent. Usually net surfing incorporates most of
Although several computer interfaces have been clicking on hyperlinks and this system does everything
devised for the handincapped (e.g., Erica [10], EWP by a single key needed to accomplish internet
[11]), there are not any inexpensive and non-intrusive operations, e.g e-mailing, online news paper reading
systems that deliver the true power and ease of today’s etc.
computers. Switch Options: To "select" or "click" on items, a
Our main goal is to develop an inexpensive hardware- range of switches been designed to meet variety of
software system for use in the most challenging cases. disabilities. Ease of use, durability, flexibility, and
above all very low cost are the prime emphasis of these
Most of the disabled persons are able to control at least switches design.
either of any body parts to a small extent. This
encompasses the construction of an hardware and its Who can Benefit: SKOP System users include adults
fine-tuning in software, as well as the definition of and children with cerebral palsy, spinal cord injury,
acknowledgeable muscle behavior and the brain injury, ALS( Amyotrophic Lateral Sclerosis),

37
multiple sclerosis, brainstem stroke (Locked-In
Monitor_Timer
syndrome), muscular dystrophy, and Werdnig-Hoffman
syndrome[13]. It can be used in homes, offices,
schools, hospitals, and long-term care facilities.
The paper is organized in some prime segments Is No
as following. Section 2 gives a description of the keystate==Kasc

system. There are two subsections in it. The subsection

2.1 describes the software sub-system, subsection 2.2
describes the hardware sub-system and the Section 3 Yes
provides the concluding remarks besides outlining the
future enhancements and modifications planned for this No Is Yes
system. flagaloc==TRUE

flagevent==TRUE
2. System Description flagaloc==FALSE

flagaloc==TRUE
2.1 Software Sub-system
The software sub-system is a Microsoft Windows Selection_Timer=FALSE
based PC application. Execute_Timer=FALSE

Execute_Timer=TRUE
Selection_Timer=TRUE
Command
pointers
Is NO Select: CASE
Directional CASE> option reposition(CASE)

pointers
Anti clockwise YES CASE=CASE+1

scan progress CASE= 1

Is
CASE<direction

NO
Fig(2.1a): Mouse Pointer Locater YES

move(CASE)
Select: CASE
highlight(CASE)

NO
Is
flagevent==TRUE

YES

execute_event(CASE)

Fig(2.1b): Operational Control panel

flagevent==FALSE

The principal software modules in the system and their Kasc : ASCII code of the equivalent keystroke. It is set to F9 (120).
functions are: flagaloc : BOOLEAN type global variable.
a. Graphical user interface: This includes flagevent: BOOLEAN type global variable.
control algorithms to manipulate cursor CASE : Integer variable ranges from 1 to 12.
motion and decision algorithms to drive the 1 to 8 stand for direction and 9 to 12 for events as
overall interface. It automatically decides clicking, double-clicking, dragging, right clicking.
when the user is actually engaged/disengaged option : Integer type global variable.
in interacting with the system. highlight(CASE): Function graphically highlights options.
b. On screen action prompt, like movement reposition(CASE): Function moves the pointer and
direction or clicking event. graphical directors.
c. Adaptive direction scanning and pointer speed
control panel. N.B. at any point of time the system can exit from it’s associative
d. Automatic multiple screen resolution setting menu. It is not shown in figure.
adoption.
Fig(2.1c):SKOPS Flow diagram
e. Performance monitoring and automatic
pointer speed control.

38
Working principle: successive translations occur with out any action, then
probably it means target location is missed. And as a
As the task of the system is intended to map
timer controls the translation of pointer so this target
the entire pointing device (more precisely mouse)
missing may be due to the speed of pointer. So as soon
functionality so let’s explore a classical system.
as the performance monitoring function encounters
If we think of pointing on the screen by any pointing
more than one successive translation cycle it reduces
device, we encounter two obvious goals as
the pointer reposition speed by increasing ts% of the
i. In which way and how long to move?
existing timerexc delay and this delay persists until next
ii. What to do at the point?
action occurs. On the occurrence of next action the
Answers of these two questions are conveyed to the
regular translation speed (set by user) is restored.
system by different movement of our hand associated
to interface peripheral. But for physically challenged
There are eight directions and four command
persons this ‘movement’ and ‘what to do’ both these
(vise left click, right click, left double clicking and
need to be communicated by a single switch. So the
dragging) selectors. All these selections are treated as a
system is designed with two mutually exclusive timers
case and are donated by a unique case number. So in
to answer the two questions and another timer above
total there are twelve cases. One timer controlled
them to decide which one them will be answered.
function gives option to select any case. One selection
These three timers together map all physical
of any case the other timer does the job defined by that
movements and commands of user.
case no.
The following diagram depicts the working of system
more clearly.
For example, one color indicator navigates
which option to choose. If one selects to move left, the
Direction Selection System:
selection timer will suspend and the execute timer will
SKOPS is having eight direction selections,
go on moving the pointer as well as the navigator pane
which instead could be modeled as clock wise or anti
to wards selected direction unless next pressing of
clock wise rotational selection(Fig. 2e).
switch.

Step:2 Step:1
The color indicator showing which
On selection left translation.
command line is active. This
Repositioned cursor and navigator indicator jumps anticlockwise
Fig(2.1g): Translation
Fig(2.1d) Fig(2.1e)
But rotational selection needs continuous traversal of Software System Specification: SKOPS operates under
360º, which consumes more time. So to reduce the the Windows operating system(98, 2000, ME, NT4.0,
operational latency, only a set of eight directions at 45º and XP). The Graphical User Interface of the software
relative regular intervals chosen. and functional logics has been developed using
Automatic Pointer Speed Control and precision Microsoft Visual Basic 6.0.
calibration: Computer output screen is the fourth
quadrant of the two-dimensional projection system.
Geometrically to locate pointer over any intended icon, 2.2 Hardware subsystem
button or canvas either of X or Y axis or (X,Y) both
together translation is needed only one time. We have produced switches from inexpensive, house
hold components. The hardware sub-system is an
(0,0) (0,X) external power free assembly of switches (push to on),
which may be connected to a common controller
device to make a single system versatile and reduce the
●( x′, y′) production cost. The hardware is not equipped with any
fragile or costly circuitry. It is a keyboard controller,
(Y, 0) producing a key press instance. And set of access
switches developed with some geometrical ergonomic
X=X+x′ x′ to facilitate the limbs movement as well as switching.
Y=Y+y′ T= y′ Depending on the problem individuality any of them
may be chosen and plugged to the controller. This
Speed=T/ milliseconds controller is interfaced to the computer via USB port.
Fig(2.1f): Translation matrix
Pointer translation is accompanied by anyone
of four basic actions (clicking, double clicking,
dragging or right clicking). If more than two

39
Its main functionalities are:
• Acquire switching signal.
• Generate keyboard interrupt. Description (2.2c): This globs is equipped with micro
The activity chart of hardware light switches at finger tips which can be operated from
next noting movement of fingers and here five optional
system: switches can be used depending on the users ability.

Connected to a Controller is
interfaced to
Range of switches common controller computer

Fig(2.2a): Activity diagram

Fig(2.2d): Bite Command

Some switches:

Description(2.2d):This is a nipple equipped with a

micro switching system inside it. It is so comfortable in
use and during our experiment different level of
disabled persons worked best with this.

Fig(2.2b): Tortoise button

Description (2.2b): This is a semi circular tortoise body

shaped switch that can be attached to palm by a very
comfortable roll crow. It is so designed to make
switching with less effort and hands free positioning.
Actual pressing location is inclined to a side so that to
push it the hand needs a small move on any surface an
to release the switch no effort needed as the semi circle
automatically roll backs the hand.
Fig(2.2e): Common Controller

Description (2.2e): It is the common controller where

different kind of switches can be plugged; it is made so
to minimize the cost of system. This comprises of a
keyboard controller circuitry. This controller is
interfaced to computer via USB port.

Fig(2.2c): Help Hand

40
 Reference:
[1] Win32 API Programming-Steven Roman
Performance
monitoring
(O’REILLY ISBN: 81-7366-136-7)
[2] Visual Basic 6 Black Book(Publisher: The
Coriolis Group)Author(s): Steven Holzner ISBN:
1576102831
Generate Options
Hardware interrupt Software
Subsystem
User [3] http://www.spd.org.sg/atc_alternativemouse.html
Subsystem Reposition/
Execute Command
[4] http://cameramouse.com/buy.htm
Switching
[5] IICP, Kolkata.
Point
Reposition [6] Dr. ir. Edward E. E. Frietman ,Faculty of Applied
Sciences Quantitative Imaging, Lorentzweg1 NL
- 2628 CJ, Delft, The Netherlands
[7] http://www.oxinst.com/MDCAPP288.htm
Fig: Context Diagram
[8] http://wmed.narod.ru/w_ophth/diagn/d_20.htm
Appendix A - Data Dictionary
3. Conclusion [9] http://www.stormingmedia.us/keywords /
electrooculography.html
There are many ways to detect muscle movement, [10] [Hut89] Hutchinson et al., "Human-Computer
some far more accurate than mechanical switches, but Interaction Using Eye Gaze Input", IEEE Trans.
these are expensive furthermore, the motion or pressure on Systems Man & Cybernetics, 19, 6 (1989)
tracking method is just a means, one in which pinpoint 1527-1534.
accuracy is not really necessary; the provided service
and ease of use of the system controlled interface is the [11] [YaF87] Yamada, M. and Fukuda, T., "Eye word
true goal. Our experiments have shown that SOP Proessor and Peripheral Controller for the ALS
system is a viable and inexpensive method for human- Patient",IEEE Proceedings A, 134, 4 (1987) 328-
computer interaction. 330.

We have been able to implement all the basic systems [12] Quadriplegics: People who have partial to total
anticipated. We are working towards improving the paralysis of both arms, both hands, both legs, and
system by adding the following features and both feet. They may be either "Quadriplegics" or
implementing the modifications to: "Quadriparetics."
For more: http://www.abledata.com/Site_2/NIRE/Quad_AT.htm
a. Cross plat form supportive. [13]Werdnig-Hoffmann Disease: It is a rare
b. Button identification technique. progressive neuromuscular disorder of infancy. Also
c. Precision drawing and gaming. known as Infantile Spinal Muscular Atrophy (SMA)
Infants with Werdnig-Hoffmann Disease may lack
d. Mobile and PDA compatible.
head control, may be unable to roll over or support
e. Wheel chair direction and control. their weight, and tend to lie relatively still, with little or
f. More intelligent scan timer and pointer speed no movement (flaccid paralysis).
setting.
g. Eye movement and eye blinking command T : Translation Matrix
system. timerexc : Execute timer delay.

41
INFORMATION TECHNOLOGY FOR THE EDUCATION OF THE PERSONS
WITH HEARING IMPAIRMENT
A.K.Sinha
Assistant Director, AYJNIHH, ERC, Kolkata.

Abstract Information related to the education of the children

with hearing impairment too is available on the
Early identification and adequate intervention internet.
prepare the children with hearing impairment to be
enrolled in regular schooling appropriate to their Information technology is being used in the
age. Parents, special educators, speech pathologist, education of children with hearing impairment in
audiologist and all who cares for the children with two ways. First it provides ample information
hearing impairment require regular input of through various web sites about current knowledge
information. One such source of information is of the experts/ parents/educational
internet and various web sites meant for that. institutes/government as well of the researchers.
Institutes working in the area, Association of the Secondly it provides use of different programmes
professionals, Educational technologist, Association such as word star/power point to be flexibly used for
of the parents, Manufactures of aids/appliances the teaching of the children with hearing impairment
provides ample in formations which are of immense both language and speech. Research has also
importance and useful to all of us. Websites like indicated that information technology not only helps
master teacher.com, inclusive-education.com to identify and improve speech and auditory skills
provides weekly Tip List on inclusive education, but also give children with hearing impairment
discipline in classroom, teaching language etc. immediate access to a vast wealth of information as
Information on auditory training materials, stories, well as expanding their knowledge and improving
teaching books can also be obtained. Free their written communication, such programs
downloadable modules are also available. stimulate the students' young, inquiring minds and
Parents/professionals can directly interact with the engender in them the self-confidence which they
experienced professional through e-mail/chatting. will need to move comfortably into secondary and
Research has shown that most children with hearing- tertiary education and to be in the workforce of the
impairment tend to use shorter, simpler sentences new millennium. In its capacity as a tool, Internet is
than those used by their hearing peers. As a result but a ripple on the surface of our educational
their general written expression appears immature institutions. In its capacity as intellectual challenge,
when compared with that of hearing peers of the it is without precedent in the educational history of
same age. Information technology not only helps to mankind. Enrollment of children with hearing
identify and improve speech and auditory skills but impairment in the regular school or special shools
also give children with hearing impairment has been limited. Education for all can’t be without
immediate access to a vast wealth of information. As them. Inclusion is not a privilege we are going to
well as expanding their knowledge and improving give to the children with disability, but it is the
their written communication, such programs natural consequence of a humane society.
stimulate the students' young, inquiring minds and Community, management, principal, teacher, parents
engender in them the self-confidence which they and professional must be equipped to face and
will need to move comfortably into secondary and inhibit the attitudinal barrier that prevents the
tertiary education and to be in the workforce of the children with disability to be enrolled in regular
new millennium. classroom. We are backed by legislation, motivated
people, awareness in community and scientific
Introduction information, therefore we may make inclusion
practical than theoretical process. This will make us
Information technology allows knowledge to children with hearing impairment in regular
spread near-instantly, and makes it accessible to classroom more visible (Sinha, 2003).
everyone, i.e., breaking down the walls of education.
A huge amount of information is available on the The importance of Internet in education,
internet. All the information is present in the form of particularly using its Web is a well-recognized fact.
web pages and websites. Moreover this number is A wealth of resources and techniques now exist
increasing consistently at unbelievable rate. which serve as a source both for exciting examples
of new teaching practices, as well as easily

42
accessible methods for adoption into various formats on inclusive education, maintaining discipline in
of teaching and learning. Internet technology allow classroom/providing answers to student’s query in
teachers and students keep up with their minds. It let non offensive manner.
them try their ideas as soon as they come up with
Another web site known as i4c (Internet
them. Internet access learners can learn anything,
4classroom) is of nature which useful for the
anywhere, and anytime. The miracle of the Internet
teacher, parents as well as students. In this various
was supposed to let great teachers reach any student,
programmes like power point are used to teach
any subject, any time, and anywhere.86th
mathematics/science /language to children with
Amendment to the Constitution of India has made
hearing impairment of different grades. Teacher can
Education a Fundamental Right for Children.
make her own lesson plans and can use their
Universalisation of Elementary Education and
imagination for the purpose of learning. Free
Education for All is one the most cherished goals of
downloadable module on teaching
Government of India.(Advani and Anupriya, 2003).
science/mathematics/spelling/ are also available. i4c
Information technology can play important role in
is website for helping teacher use the internet
achieving this goal.
effectively. It provides daily dose of the web/Links
for K-12 teacher/On –Line practice module.i4c.
Searching The Web
Listen-Up web provides free downloadable
If one wants to get information on a particular modules of “Learning to listen sounds”. Parents can
topic one should go to Search Engine or a Search interact with the professional using internet. One
Directory. A search engine/directory may be such web site available is elln@Auditory Verbal
considered as a catalogue of information on the Training .com. Apart from family consultation it
internet. The famous and efficient search engines are also provides “Use fast track Auditory Verbal
Google/ Altavista/ Wisenut/ Teoma/ program based on Ellen Rhoades work “road map”
HotBot/Lycosand Althweb. The important for infants, preschoolers, and elementary school
directories include Yahoo/LookSmart/Open aged kids. Professional/parents can also take basic
Directory. If we search for “Use of internet for the course for certification as Auditory Verbal
education of the children with hearing impairment”, Therapist.
we get 54,400 web results. “Activities for Auditory
Training” provides 72,000 web results. Internet provides all information related to
amplification/hearing test/abstract of articles/review
of books/research articles and various professional
Internet For Classroom and parents associations. Indigenously web site of
Parents, teachers and students can take AYJNIHH provides site to check your hearing, jobs
advantage of internet directly. There are two types of for the deaf and linkages to other web sites.
programmes available through internet. One is
interactive and other is non-interactive. Interactive
programme provides opportunity to Personal Computer Application Course
parents/teacher/students to interact with distant Personal computer application course is being
teacher/expert through online programmes and also conducted by the ERC of AYJNIHH since 2000 in
provides ways and to express their emotion and collaboration with WEBEL, Government of West
thoughts through chatting/e-mail. One of the major Bengal for the persons with hearing impairment.
advantages of the online programs is the expert Fourth batch has been in progress. So far 26 males and
opinion available to many persons at a time making 9 females have been registered. 29 candidates belonged
it a cost effective method. Teacher/parents can get to general category, 3, 2, and 1 belonged to SC, ST and
regular updates for certain problems faced by them. OBC category respectively.
One of such web sites provides weekly updates
known as Masterteacher.com, which provides Tiplist

Age range distribution is shown in Table I.

SN Age Range(Years) No. of candidates

1 20-25 13
2 26-30 14
3 31-35 8
Table I

43
WEBEL Computer Applications required minimum class X pass candidates. In the beginning of the course an
apprehension was that if sufficient candidates will be available or not. Table II indicate qualification of the
candidates enrolled.

SN Class X Class XII Graduate Post Graduate

1 13 13 7 2

Table:II
Achievements of the candidates with hearing impairment are definitely appreciable in spite of having hearing loss
as shown in Table III.

S Bil. Bil. Bil. Mod- Severe/ Mod-

N Profound Severe Mod- Sevre/ Profound Severe/
SNHL SNHL Severe Severe SNHL Profound
SNHL SNHL SNHL

1 24 7 1 1 1 1
Table: III
Examination was conducted by WEBEL and so far three batches have been passed out with the grades as shown
in Table IV.

SN Grade A Grade B Grade C Grade D Grade E

(85% and (70-84%) (55-69%) (40-54%) (<40%)
above)
1 1 17 8 NIL NIL

Table: IV

After completion of the computer course all students felt more confident to face the challenges of the life and had
elevated esteem. Table V indicates the employment status of the candidates who successfully undergone the
course.

SN Unemployed Self Private Govt. Sector Information Continuing

Employed Sector Not Higher
Available Studies
1 2 8 5 2 8 1
Table: V

Candidates also improved in communication skills using painting program very successfully. However
while they were practicing sending e-mail, designing program could not be sustained due to lack of
page, making birthday/new year card and surfing manpower. It is time to revolutionized use of
different web sites. The results are encouraging and information technology in the education of the
in agreement to the research which indicate that children with hearing impairment in our country as it
information technology helps in improving will enhance our effort of inclusion.
communication skills for the children with hearing
impairment. Exposure to computers also facilitates
higher education for the persons with hearing Conclusion
impairment. We must ensure that the knowledge for
To educe means to bring out a potential existence.
the given course should be at par with their hearing
Education, therefore, is a process of intellectual
peers.
growth. The biggest impact of the Information
Introducing computers to young children with technology is to change the point of view that
hearing impairment was also introduced at the education is something that can and should be
center. Students showed lot of interest and were delivered. Education comes from learning, not

44
teaching. The world's best teachers are not tertiary education and to be in the workforce of the
repositories of knowledge, but skilled navigators new millennium.
who lead young minds to discovery and
understanding. Learning is about reinventing the Reference
wheel, and may all learners have the opportunity to 1. www.masterteacher.com
do so. The educator is merely a midwife in this 2. www.i4c.com
process. Information technology not only helps to 3. www.Listen-Up.com
identify and improve speech and auditory skills but 4. www.ayjnihh.org
also give children with hearing impairment 5. Advani,L., Chadha,A.(2003):You and Your
immediate access to a vast wealth of information. As Special Child. UBSPD, New Delhi.
well as expanding their knowledge and improving 6. Sinha,A.K.(2003):Inclusive education of the
their written communication, such programs children with hearing impairment. Abstract
stimulate the students young, inquiring minds and
Volume of IASE,
engender in them the self-confidence which they
University of Kalyani.
will need to move comfortably into secondary and

45
COMPARISON OF NORMAL AND PATHOLOGICAL KNEE SWING CYCLE
ELECTRICAL IMPEDANCE SIGNALS FOR POSSIBLE NONINVASIVE
DIAGNOSIS OF OSTEOARTHRITIS
Suhas S. Gajre1@, Rajendra K. Saxena1, U. Singh2, S. Anand1
1Centre for Biomedical Engineering, Indian Institute of Technology, Delhi and
Biomedical Engineering Unit, All India Institute of Medical Sciences, New Delhi
2Department of Physical Medicine and Rehabilitation, All India Institute of Medical Sciences, New Delhi

arise. Normal articular cartilage acts as a shock

Abstract absorbing material, which prevents knees from injury
in mechanical loading. OA represents an imbalance in
Osteoarthritis (OA) of knee joint is the most common the destructive and synthetic processes of the cartilage
disabling disorder in Asian countries. It affects that leads to erosion of the cartilage. In addition, there
significant population, mainly the elder people and is a decreased concentration and viscosity of the
results in enormous financial costs of therapy and later synovial fluid in osteoarthritic patients, and this may
surgery. It is a degenerating and irrecoverable decrease the lubricating and cushioning properties of
condition of articular cartilage. The causative factors the joint.
responsible for genesis of disease are not well
understood. Various etio-pathological factors such as Different noninvasive techniques have been tried by
mechanical stress, obesity, abnormal postures and many researchers [6_10], but none so far has achieved
aging have been found for the early onset of the the final goal. Although such methods as MRI scan,
disease. The progress of disease can be contained if it Computed tomography (CT) and ultrasound are being
is diagnosed early. Several diagnostic techniques have tested at present, the only validated indicator is still a
been developed, but none gives cost-effective, simple, flat x-ray image [6, 7]. X-rays report the degeneration
and useful for mass and repeated use solution for the of articular cartilage when that has reached to the
early diagnosis. Attempts have been made for irreversible stage. There are many cases when, invasive
diagnosis and monitoring knee OA for other novel tests, such as arthroscopy, were found unwarranted
techniques. Analysis of noninvasive signals around when done merely on symptoms. We can not use
knee offers one such interesting way of diagnosis of arthroscopy for monitoring cartilage changes due to its
knee osteoarthritis. In the present study electrical invasive nature and cost. Arthrography is a semi-
impedance plethysmographical signals of knees are invasive procedure, and has its disadvantages. Lack of
used to compare normal and osteoarthritic subjects a cost-effective non-invasive method, which can be
with the possibility of using them for noninvasive used to measure progression of disease, limits the
diagnosis. In knee swing cycle experiments on 7 development of treatments for OA.
normal and 9 pathological subjects (six cycles for each
subject), we found that the mean of variances and mean Electrical conductivity and its distribution change with
of RMSs of OA subject signals after normalization OA [11_13]. Considering this, a control group and OA
with respect to base impedance are 45% and 27% group were used to study the effect of loading on the
lower than that of normal subjects. The results show knee electrical impedance in a study [14]. Results
that electrical impedance knee swing signals may be elucidated significant changes in the fluctuations
used for the noninvasive diagnosis of the disease. during different loading states and the area under load
duration curve in the control and OA group. In this
Keywords _ Osteoarthritis, Knee, Diagnosis, Electrical current study we assessed electrical impedance signals
Impedance Plethysmography, Knee Swing Cycle. quantitatively in a different way, without using the
loading protocols, trying for a more convenient
environment of diagnosis, which would be suitably
INTRODUCTION adapted to the common clinical environment.

Osteoarthritis (OA) of knee is the most common

synovial joint condition in Asian countries [1, 2]. Its MATERIALS AND METHODS
prevalence in the population of elderly is more than
almost 50% [3_5]. OA often causes disability resulting A. Data Acquisition
in dependency and later enormous financial costs of the We studied tetra-polar electrical impedance
therapy, including final surgical treatment, if the need plethysmography (EIP¥ or IPG) signals of knee swing

46
cycles (defined later) to explore the possibility of using and after amplitude normalization wrt base impedance
them for noninvasive diagnosis. Figure 1 shows the of 33 ohms. Static impedance (BI) value of the subjects
schematic of our study. A control group (NN, with was used as a reference for normalization.
normal healthy knees) and an OA group (OA, patients
with pathology of the knee diagnosed) participated in Following quantitative assessment and comparison
our study. Guidelines given by American Rheumatism were made:
Association (ARA) were followed for the selection (i) Mean of means (averages) of all the signals for
criteria [15]. The subject profile is shown in Table 1. each subject,
All subjects were asked to relax for 10 min. before (ii) Mean of variances of all the signals for each
each recording. Training sessions were held before subject, and
recording of the actual data. Specially prepared silver (iii) Mean of RMSs of all the signals for each subject.
coated braided ring type electrodes were placed, with Scatter plots of the parameters were also studied,
the ECG gel on its surface, above and below knee-cap since the aim of the study was to see whether we
(patella). The distance between two voltage electrodes can use these parameters for the classification or
was almost fixed (mean 15 cm, SD 0.7 cm). Current not.
electrodes were made wider for better current insertion.
RESULTS AND DISCUSSION
Each subject sat on a rigid table with his/her legs
suspended over the table's edge. Then the subjects were Table 2 and 3 gives the numerical values of mean and
asked to repeatedly extend his/her leg from 90° to 0° SD for each group without and after normalization. All
during a roughly fixed time, then flex from 0° to 90° in the three statistical parameters are compared without
the same time (see Figure 2). A full swing is defined as and after normalization for NN and OA group. A
the extension and flexion combined. Subjects were standard base impedance of 33 ohms was used for
asked to swing the knee as normally as possible normalization. Figures 5-10 give bar and scatter plots
without bothering about their speed of each swing of the above.
because later all the swing cycles were normalized in
the time domain. The digitized recordings of _Z, a We observed that the base impedances of the subjects
standard 1 Hz clock signal, and a marker signal were are overlapping and can not be used as a feature for
made using DAQ system for ach subject for offline classification (see Figure 11). This is the same result as
analysis. The sampling frequency was chosen as 1000 in [14].
Hz, which was later down-sampled by in offline
analysis. Pathological information of each subject was The results show that there is substantial difference in
also noted down along with other important the statistical properties of the control group and the
physiological parameters. A reading of the static OA group. On and average, the parameters for the
impedance, called base impedance (BI), was also made control group are higher than the OA group, except for
when the leg was suspended normally (at 90°). This the mean of averages. This is because of the almost
was used for normalization in second part of the study. pure ac nature of the signals (see mean of means). This
The recording of 20-30 seconds for each subject was clearly indicates that mean of averages can not be used
done. as a feature for classification. There is no effect on the
values of the same before and after normalization for
obvious reasons. The mean of variances and mean of
B. Signal Processing and Analysis RMSs are showing different values for both the cases.
In case of without normalization, the mean of variances
Some noise might be introduced in the signal in the was 40% and mean of RMSs was 22% less that of the
data acquisition phase, such as instrumentation noise, NN group. This reduction was further enhanced to 45%
ADC noise, myographic noise, powerline noise, etc. and 27% respectively after normalization. Although
We were interested only in low frequency components there is some overlapping in the groups, this can be
because all knee swing cycles were normalized to 4 used as some of the features for the diagnosis of OA. In
seconds. Hence high frequency components were the case with and without normalization, normalization
filtered out outside our band of interest. Six knee swing results are less scattered for both NN and OA group.
cycles were chosen using zero crossing technique for Hence normalization should be preferred for the
each knee, which were used for the analysis. Typical classification.
zero crossing knee swing signals for normal and
osteoarthritic knees are given in Figures 2 and 3 This study was using very limited set of the subjects.
respectively. As far as the classification efficiency is considered,
small number of signals (42 and 54 for each
We studied 42 normal and 54 pathological knee swing group) may not give good results. Hence the
cycles in two ways: without amplitude normalization,

47
48
49
 A NEUROLINGUISTIC ANALYSIS OF BENGALI HYPERKINETIC
DYSARTHRIC SPEECH

Nilantika Chakraborty*, Prof. Krishna Bhattacharya*, Dr. Trishit Roy**

University of Calcutta* and Bangur Institute of Neurology, Kolkata**
India
nilantikachakraborty@rediffmail.com

Abstract limbs, hand, face, mouth and neck. Such movements

interfere with speech resulting in articulatory
Dysarthria is a motor speech disorder. Till date no imprecision, prosodic abnormalities etc.
study on Dysarthric Bengali speech is available from
the neurolinguistic perspective. In this ongoing study This paper deals with the perceptual analysis of the
an attempt has been made to trace the various features Hyperkinetic dysarthric speech from the
of hyperkinetic dysarthric speech of the patients with neurolinguistic angle. This study has been conducted
Bengali as native language from the point of view of on Bengali language and is the first of its kind. The
articulation, respiration, phonation, prosody and other studies, which are worth mentioning in the area
resonance and also to find out the Bengali sounds, of dysarthria, are the Mayo Clinic dysarthric study by
which are, affected the most. The data of 12 Darley, Aronson and Brown, 1975 [1]; the Frenchay
hyperkinetic dysarthric subjects (10-75 years) were Dysarthria Assessment by Pam Enderby, 1983 [2];
analyzed. The participants were subjected to Assessment of Intelligibility in Dysarthric Speakers, by
‘Articulation test’ using words and sentences. These Yorkston and Beukelman 1981 [3]. In this paper main
were recorded in a tape recorder and later on perceptual emphasis has been given on the Mayo Clinic dysarthric
analysis was also done. Perceptual analysis could study, as the 22 dimensions, which have been used here
identify the specific defective features in most cases. in the identification of the type of dysarthria, are part of
Identification of the distorted speech sounds will be of that study. There is a need for the study of these
definite help in speech therapy. dimensions for the Bengali speech sounds as the
phonetic structures and prosodic structures are
Key Words language dependent phenomena. And with the help of
this study we can identify the dimensions, which are
Dysarthria, Hyperkinetic, Perceptual analysis, affected in Bengali hyperkinetic dysarthric speech. The
Imprecise consonant, Strained voice, Irregular study was conducted on 12 hyperkinetic dysarthric
articulatory break patients. Their speech was recorded. With the help of
the perceptual analysis various dimensions, which were
grouped under the five heads of articulation,
1. Introduction respiration, phonation, prosody and resonance, were
analysed. Further the speech sounds that are distorted
Dysarthria is a motor speech disorder that is during speech have been identified. This finding will
caused by the weakness or in coordination of the be of help in speech therapy. The results are given in
muscles associated with speech. The speech gets slow, detail in the later section of the paper.
imprecise or weak. It becomes difficult to comprehend
what the dysarthric individual is attempting to say. A
normal human being speaks as easily as he breathes. 2. Objectives
However, to achieve this, the underlying subsystems,
which are at work to generate speech, needs proper There are two objectives of this paper. First of all,
coordination. The subsystems that work hand in hand the objective of the paper is to trace the various
are the respiratory subsystem, laryngeal subsystem, dimensions in Hyperkinetic dysarthria from the point
velopharyngeal subsystem, and articulatory subsystem. of articulation, respiration, phonation, prosody and
A weakness in any one of the systems can result in resonance of voice so far as the Bengali speech is
dysarthria. concerned. Last but not the least, it also attempts to
Hyperkinetic dysarthria is caused by the find out the sounds affected in Hyperkinetic dysarthria
dysfunction in the basal ganglia, related portions of the and thereby aiding in the future management of therapy
extrapyramidal system or sometimes by the cerebellar for dysarthria. On the basis of the perceptual analysis
control circuit. Any kind of lesion in either of these the defects in terms of the different dimensions could
areas produces unpredicted, involuntary movements of be confidently ascertained which have the potential to
diagnose the disorder to a major extent.

50
 The other dimensions like High pitch, Low pitch and
Monopitch have been omitted from this study. It will
3. Materials and Methodology be dealt separately in future communication.
Subjects
Let us now discuss the definitions of the above-
12 subjects of 10-75 years age (male 24, female 4) mentioned terms.
displaying dysfunction of the basal ganglia participated
in the study from the Neuromedicine Out-patient Imprecise consonant- During production of the
department of Bangur Institute of Neurology, Kolkata. consonant sounds when there is a loose contact
They are all right handed and have Bengali as their first between the articulator and the point of articulation or
language. sometimes when the articulator touches the point of
Procedures articulation slightly forward or backward.
An ‘Articulatory Test material’ consisting of two Distorted vowel- The vowel sound lack precision
categories i.e. 184 words and a text consisting of 9 throughout their total duration.
sentences were used as a stimulus to elicit response
from the subjects for the study. The test material Prolonged phoneme- The duration of a continuant
contains words, which have all the Bengali vowels and phoneme is lengthened while production.
consonants in the initial, medial & final positions of Forced inspiration expiration- Extra effort is used for
the words. Further, it also has a list of words with the incoming and outgoing air during speech
consonant clusters in the initial and medial positions. production.
These words help to capture the extent to which the
speech production is impaired in a particular Audible inspiration- During speech there is audible
individual having hyperkinetic dysarthria. inspiration.
Orthographic representations of the stimulus material Grunt at the end of expiration- There is a grunt like
were also provided in front of the speakers. The sound at the end of expiration.
dysarthric subjects were instructed to produce the
target words at a conversational rate. Subjects’ Voice tremor- During speech the voice displays
productions were audio recorded using a Sony TCM- tremor.
150 cassette recorder with built-in microphone at a Mono loudness- Voice lacks variation in loudness.
constant mouth-to-microphone distance of 5
centimeters. The recordings of each of the 12 subjects Alternate loudness- The loudness of the voice
were then submitted to perceptual analysis. The alternates between high and low during speech
perceptual analysis mainly focuses on finding the production.
phonetic capabilities of the dysarthric subjects and Loudness decay- The voice shows diminution of
also helped to identify the various dimensions of loudness as the speech progresses.
articulation, phonation, respiration, prosody and
resonance as mentioned in Table 1 below. Strained voice – Voice is weak and effortful.
Table 1: Dimensions used in the assessment of Breathy voice- Voice is weak and thin.
dysarthric speech
ARTICU
RESPIR- PHONA- PROS- RESON-
Hoarse voice – Voice is harsh and raspy.
-
ATION TION ODY ANCE Increase in overall rate – Speed of the speech
LATION
IC FIE VT RS HN increases progressively from beginning to end.
DV AI ML SR NE
Inappropriate silence – Speech is interrupted by
PP GEE LD SHR
AL IOR
silence.
BV IS Excess or equal stress- There is stress of equal degree
SV EES on each word of a sentence and even on unstressed
HV IAB parts of speech.
Irregular articulatory break- There are unnecessary
IC-Imprecise consonant, DV-Distorted vowel, PP -
pauses in between syllables of a word. Reduced stress-
Prolonged phoneme, FIE – Forced inspiration
Speech lacks in proper emphasis during the time of
expiration, AI – Audible inspiration, GEE -- Grunt at
production.
the end of expiration, VT -- Voice tremor, ML – Mono
loudness, AL- Alternate loudness, LD -- Loudness Slow rate- Speed of the speech is slow.
decay, SV – Strained voice, BV – Breathy voice, HV --
Short rushes- Speech is separated by pauses.
Hoarse voice, IOR –Increase in overall rate, IS --
Inappropriate silence, EES -- Excess or equal stress, Hypernasality- Voice is excessively nasal.
IAB -- Irregular articulatory break, RS – Reduced
stress, SR -- Slow rate, SHR -- Short rushes, HN -- Nasal Emission- There is emission of airstreams
Hypernasality, NE -- Nasal Emission. through nose during speech production.

51
Statistical analysis • palato-alveolar sounds like c, ch, j, jh become
imprecise as the tongue fails to touch the place
The data was entered into Microsoft Excel Spread- of articulation.
sheet. Statistical analysis was done by Statistica
version 6.0 (2001). Different linguistic dimensions of • the nasals and lateral sound like m, n, l are
the hyperkinetic dysarthric speech were assessed by imprecise.
Chi-square test or Fischer’s Exact test, as appropriate.
With the help of the statistical analysis the distribution Figure 1- Distribution of dimensions among the
of the various dimensions under articulation, patients of hyperkinetic dysarthria
respiration, phonation, prosody and resonance was
tested. The dimensions whose distribution is
HYPERKINETIC DYS
statistically significant were included as identifying
criteria for the hyperkinetic dysarthria.
100.00%

4. Results of the analysis 50.00%

Out of these 22 dimensions discussed under the 5 head 0.00%

groups such as, articulation, respiration, phonation, IC DV PP FIE AI GEE SV BV HV

prosody and resonance, only 6 dimensions are

statistically significant as identifying criteria for
hyperkinetic dysarthric speech. The figure within the
parenthesis shows the percentage of dimensions’ Figure 2- Distribution of dimensions among the
distribution for the hyperkinetic dysarthric speech. In patients of hyperkinetic dysarthria
addition to this the distribution of all the dimensions
has been given in Figure 1, 2 and 3.
Imprecise consonant (91%), Strained voice (75%), HYPERKINETIC DYS
Irregular articulatory break (58%), Voice tremor
60.00%
(51%), Short rushes (46%) and Increase in overall
rate (43%) 40.00%

These results have been derived by analyzing the 20.00%

speech of 12 hyperkinetic dysarthric patients. Since it 0.00%
is an ongoing study, so it should be kept in mind that VT ML AL LD RS SR
with the increase in the number of the patients the
results may differ to certain extent. From the study it
has been observed that the dysarthric individuals have
problem with their phonetic output. That is, while Figure 3- Distribution of dimensions among the
producing the words they have difficulty with the patients of hyperkinetic dysarthria
sound production (specially the consonants). The
patients speak effortfully and have a strained voice
quality. Their speech is interrupted by pauses (IAB) HYPERKINETIC DYS
and exhibits tremor (VT). The voice in many cases lack
tonal variation. In some cases there may be excessive 80.00%

loudness variation. The rate of the speech may increase 60.00%

from beginning to the end. There are short rushes, as if 40.00%

the patient is in a hurry to end the conversation. 20.00%

Articulatory dysfunction is revealed in imprecise 0.00%

SHR IOR IS EES IAB HN NE
consonants. The consonant sounds that are distorted at
the time of production are as follows:

• aspiration is missing in the voiced and

unvoiced aspirated stops like ph, bh, dh, gh, 5. Conclusion
and kh.
It has to be kept in mind that the paper is a part of an
• s sounds are generally dropped in the clusters ongoing research work. More number of patients will
word initially, e.g.: spriSSo ‘touchable’ be included in the study. And as a result the
becomes pricco. observations may slightly differ after further study.
Identification of the distorted speech sounds will be of
• r sound is dropped in the clusters, e.g.:
definite help in speech therapy and future management
prokriti ‘nature’ becomes pokiti
of dysarthria.

6. Acknowledgements

52
 Motor Speech Disorders.
I gratefully acknowledge the help of Dr. Amar Mishra (W.B. Saunders Company, Philadelphia, 1965)
and all the other doctors of Bangur Institute of [2] Enderby, P.
Neurology. I am also grateful to Dr. Abhijit Hazra of Frenchay Dysarthria Assessment.
(College Hill Press, California, 1983.)
Dept. of Pharmacology, University College of
Medicine for his ardent help. [3] Yorkston, K.M., & Beukelman, D.R.
Assessment of intelligibility of dysarthric speech.
References: (C.C. Publications, Tigard, OR, 1981)
[1] Darley, F. L., Aronson, A.E., and Brown, J. R.

53
 EXTENDING LIBRARY FACILITIES TO PHYSICALLY CHALLENGED –
CASE OF IIT KHARAGPUR
Atin Nandi
Technical Assistant
(Central Library, Indian Institute of Technology Kharagpur)
E-mail:atinnandi1@yahoo.com
Ph. No.: 03222-282442
&
Pratap K. J. Mohapatra
Professor
(Department of Industrial Engineering & Management)
E-mail:pratap@hijli.iitkgp.ernet.in
Ph. No.:03222-283738

This paper presents a needs assessment study on Disability: Any restriction or lack (resulting from
the physically challenged users of the Central Library impairment) or ability to perform an activity in the
of IIT Kharagpur. A questionnaire survey has been manner or within the range considered normal for a
made for this purpose. Motor disability and minor to human being…
moderate form of visual disability are the main
challenges for these users. The paper proposes a set of Handicap: A disadvantage for an individual, resulting
short- and long-term measures and various forms of from an impairment or disability that limits or prevents
assistive technologies that the Central Library should the fulfillment of a role (depending on age, sex and
take in order to give adequate services to this special social and cultural factors) for that individual.
category of users.
Broadly speaking, there are four types of disabilities
Keywords: Disability, Assistive Technology, Library [7]:
Service, Questionnaire Survey.
i) Visual Disability (e.g., Low vision, Blindness,
and Color blindness)
1. Introduction
Traditionally, only those students who were ii) Hearing Disability (e.g., Hard of hearing, Deaf,
and Pre-lingually deaf)
medically fit were considered for admission into
engineering institutions in India. IIT Kharagpur was no iii) Cognitive Disability (e.g., Attention Deficit
expectation. With the passage of time, ‘The Persons Hypertensive Disorder, Developmental delays,
with Disability (Equal Opportunities, Protection of Alzheimer’s disease, Brain injury, Stroke, and
Rights and Full Participation) Act 1995’, [8] it became Learning Disability. There are wide varieties of
mandatory to reserve 3% seats for persons with Learning Disability namely Dyslexia,
disability for admission into IITs. Today, the total Dysgraphia, Dyscalculia, Disnomia, Memory
number of students with disability at IIT Kharagpur, Disability, Visual Perception Disability, Auditory
stands at 25. Unfortunately, an important service like and Perception Disability)
Central Library has not geared itself for giving equal
opportunities to this set of users. The present study iv) Motor Disability (e.g., Muscular dystrophy,
aims at finding out the nature of disability of physically Amputation, Arthritis, Cerebral pulsy, and Spinal
challenged users of Central Library of IIT Kharagpur, cord injury).
the difficulties they face in getting the library services, v) According to “The Persons with Disability (Equal
their suggestions with regard to privileges and the Opportunities, Protection of Rights and Full
nature of Assistive Technology they need to make the Participation) Act 1995” [8], disability means
best use of the Library. blindness, low vision, leprosy cured, hearing
impaired, locomotor disability, mental retardation
2. Concepts of impairment, disability and handicap and mental illness.
World Health Organization (WHO) offers the There are different degrees of disability [4]:
following definitions [1 p.13]: - Mild <= 40%
Impairment: Any loss or abnormality or psychological, - Moderate >= 40%
physiological or anatomical structure or function… - Severe >= 75%
- Profound / Total = 100%

54
3. Assistive Technology 3.3. Assistive hardware support
With the advancement of Information Technology, A large number of hardware supports relating
The volume and the range of digitally stored to related to Assistive Technology are also available
information are increasing while at the same time the today. [6][9a]. Some examples are: Audisee (A good
number of people accessing such digitally stored listening device), Boost Tracer Head Mouse (made for
information is increasing throughout the globe. Now persons who lack hand or figure movement but don’t
the question is: How does a physically challenged user find any problem regarding head movement), Braille’n
make use of the Web based information services? It is Speak (note taking device), Quick Glance Eye (move
possible with the help of Assistive or Adaptive the cursor according to user’s eye movement), Reading
Technology. Pen II (a tool with the help of which any word on
computer screen can be spoken through built-in-
speaker), and TTDs (Telecommunication devices for
3.1. What is Assistive Technology? Deaf).
Assistive Technology or Adaptive Technology (AT) 3.4. Accessible web designing
is the application of technology to assist an individual
with disability. People with disabilities may use Millions of victims of different types of disabilities
Assistive Technologies to (a) improve ability, (b) round the globe may miss to extract the benefits of
achieve self-sufficiency, (c) facilitate care, and for (d) digital revolution unless accessible Web sites are
education, (e) sports, (f) communication, and (g) designed for them. Important considerations for
environmental control [9b]: accessible Web sites are given below [7]:
ATs are available in many forms. It may be as (a) Scrolling text, excessive use of color, and small
simple as a magnifying glass or may be a highly font size should be avoided.
sophisticated tangible or intangible product like screen (b) Provision should be made for Style Sheets for
reading software, screen magnification software, or color blind people, Closed Captioning for visual
braille printer. disable people, Open Captioning for hearing
disable people and for Standardize icons for
3.2. Assistive software support Dyslexic persons.
A large number of Assistive software aids are in (c) Testing of accessibility of Web site should be
existence. [3][9a]. A short list of such software aids is carried out with the help of accessibility testing
given in Table 1. software aids such as A-Prompt, Bobby,
Macromedia, and WAVE.
Table 1: Disability software aids
Type of Examples of specific software 3.5. Steps for implementing Assistive Technology
Disability Selection of most suitable Assistive Technology is not
Visual (a) Speech recognition software a very easy task. It is a multi-step process as given
disability Dragon Naturally Speaking, IBM below [7]:
Via Voice, Speech viewer, Jaw
Bone (a) A fraction of budget of an organization should be
(b) Screen reading software reserved to discharge specialized service to the
Out Spoken, Text Help, Supernova, patrons with disabilities.
JAWS, Screen Reader/2 (b) The organization should have modern Technical
(c) Screen magnification software Information Centers, providing Assistive
MAGIC, Zoom Text, inLarge, Technology based service to the users with
Lunar disabilities.
(d) Braille software
Duxbury Braille Software, (c) To organization should determine the target or
Winbraille, Sparsha focus group.
(e) Note taking software (d) It should also determine the kind of Assistive
Aris Technology based service needed for such target
Hearing iCommunicatorTM System, Personal groups.
disability Communicator
Cognitive PEATTM, ISSACTM , Brain Train, (e) The organization should also carefully plan the
disability Speech Viewer III, Kurzweil, Talking location of the Disability Service Unit.
Keys Pro (for Learning disabilities) (f) It should acquire the latest information on AT
Motor Evatuware, Half-Qwert, Magic products and services.
disability Cursor2000, Talking word processors
like IntelliTalk, Write-OutLoud. (g) Selection of proper AT vendors

55
(h) Automatic wheel chair, automatic door openers, Table 2: The Respondent Profile
lifts, suitably designed handrails, elevators etc. Year 2002 2003 2004
should be available. No. of students PG= 8 PG=10 PG=12
(i) It should offer Information Literacy Program for admitted UG= 1 UG=2 UG=NIL
the target users. No. of Students PG= 4 PG= 6 PG= 8
responding to UG= 1 UG= 1 UG=NIL
questionnaire
4. A case of Central Library of IIT Kharagpur Percentage of 55.56% 58.33% 66.67%
response
In the past, neither the administrators of IIT
Kharagpur nor the Central Library administrators had
Summary of response to questionnaire survey is given
made any consideration for persons with handicap
in Table No. 3.
while designing the Library building and the location
of its collection and services. The Library did not have
any idea as to the size of the population of persons with Table 3 Summary of responses
handicap. We have made an attempt here to get to Type of Disability Motor 55%
know the size of the population, the nature of problems Visual 40%
they face, the type of facilities required by them. This Hearing 5%
has helped the Library to make a very modest Frequency of Library >=2days/week 55%
beginning towards delivering its services to these visit Others 45%
unfortunate mass. We generally followed the steps Stay in the Library per >=1hour 50%
given below: visit <1hour 50%
Getting up the staircase Yes 15%
Step 1: We decided to limit the scope of the study to difficult No 70%
only the student users with handicap. To some extent 15%
Step 2: We collected a list of names of students with Facing difficulty in Yes 20%
handicap from the Academic Section of the bringing books from No 70%
Institute. racks To some extent 10%
Step 3: We designed a questionnaire to know (i) the Demanding for separate Yes 50%
type of disability each one of them suffers section Physically No 50%
from, (ii) the extent of use he makes use of the Challenged students
Library and its services at the present, (iii) the
difficulties he is facing while searching for the Two thirds of the students use reading and borrowing
desired documents, and (iv) his suggestions facility whereas only a handful (about 25%) have used
for making the Library services accessible to the facilities (such as photocopying, electronic library)
persons with handicap. that are housed in the Annexe Building.
The PH students also made several
Step 4: We analyzed the responses.
comments/suggestions:
Step 5: We have developed a set of recommendations • They need a separate textbook reserve section.
to address this problem.
• They face great difficulty to climb staircases
Step 6: We have implemented a couple of in the Library.
suggestions.
• Database searching is a bit difficult.
A questionnaire survey was conducted among the
• It is very difficult to search the shelves having
physically handicapped students of IIT Kharagpur who
to stand continuously for long.
were admitted into the UG and the PG programmes of
the Institute during the years 2002, 2003, and 2004. • To trace out a book is extremely difficult.
The questionnaire consisted of 14 questions of which • Illumination is poor in many places.
the majority were closed barring a few that were open-
• Provision for wheel chairs should be made
ended.
within the Library.
In most of the cases we had to meet the students • Nearly 80% users wish that the Library has
personally in their halls of residence and their reference and general collections concerning
classrooms to get the responses. While some hesitated resources on disabilities.
to respond to our survey, a few even did not bother to
respond to the survey. Interestingly, friends of a few • A few Physically Challenged students want
victims did not even know that the latter had any form that the Library should make provision for
of handicap. Assistive Technology namely – Screen
magnification software and hearing disability
Out of 33 (PG and UG) students whom copies of software.
questionnaire were distributed, only 20 students (i.e.,
60.61%) responded (Table 2).

56
5. Conclusions (10) Help Desk should give special assistance to PH
users, including assisting them in searching online
Pitifully, the design of the building of the Central catalogue and other databases.
Library and its resources were never planned keeping
(11) Photocopier facility should be available in Hall 1.
the persons with handicap in view. Our study clearly
brings out the following: (12) Fraction of our Library budget should be kept
aside for providing service to physically
(1) The population of student users with handicap is challenged persons.
rather small with motor disability and mild to
(13) Librarians should be up-to-date with modern AT
moderate visual impairment affecting most. The
products.
lack of a critical mass, perhaps, is the reason why
it has not yet attracted the attention of the IIT and (14) Special training should be offered to this type of
the Library administrators. Time, however, is ripe users in the use of the assistive technologies.
to solve the problem. (15) More intensive studies should be done to know the
(2) Central Library should unequivocally spell out, in type of disabilities our students and staff are
its mission statement, the goal of making its suffering from in order to plan its policies to serve
collection and services accessible to the persons the PH community more meaningfully.
with handicap. The Library has implemented two of the above-made
(3) It should appoint a coordinator of the Library suggestions. The PH users now can take a maximum
services for disabled users. of eight books for a period of three months. Further,
they are now allowed to borrow books from the SC/ST
(4) Library needs to be housed in a barrier-free
Section.
building. Till such a building is made available,
renovation of existing building design must be
carried out and the Library must plan its activities 6. Reference
purposefully to serve the PH community. (1) Barnes (Colin) and Mercer (Geof). Disability.
(5) Renovation can take the form of making Cambridge: Polity Press, 2003.
wheelchair accessible ramps to reach different (2) Country laws index.
floors or making special type of lifts that can carry http://www.dredf.org/symposium/lawindex.html
person on the wheelchair on to the upper floors. (3) Gillete, (Daniel). Designing adaptive technology
(6) A separate section should be created for the PH for those with learning disabilities.
students. It should be housed in Hall 1. The http://www.cisp.org/imp/june-
Section should contain all the textbooks required 2001/0601gillette.htm
by the PH students and various forms of assistive (4) Guidelines for evaluation of various disabilities.
technologies. Kalia: quarterly journal for independent living by
(7) For the motor-disabled users, disabled people. xii(2), 1997, p.1.
- Books should be shelved in this section at a (5) Moore (John R). Technology and access for people
wheel-chair height. with disabilities.
- Toilet and water fountain in Hall 1 should be http://www.txla.org/pubs/t1j76_1/access.html
wheelchair accessible. (6) Nandi (Atin). Web based services for the victims
- A computer available for searching online of disabilities. In the 22nd Annual Convention and
catalog and other databases should be placed Conference on Digital Information Exchange:
on an ergonomically designed table for use by Pathways to build Global Information Society, IIT
persons on wheelchair. Chennai, January 22-23, 2004. pp.473-478
(8) For the visually impaired users the Library should (7) Nandi (Atin), Satpathi (J N) and Bhattacharjee (S
provide K). Impact of assistive technology for providing
library and information service to the users with
- Large text access in the form of either screen
disabilities. Bulletin of the National Institute of
magnification software for electronic texts
Homeopathy, 6(2), Apr. 2003. pp.36-39.
including searching for OPAC and online
databases or scanned material access that (8) The Persons with Disabilities (Equal
combines the features of scanning print Opportunities, Protection of Rights and Full
material and projecting it on the screen using a Participation) Act, 1995: model rules for state
computer projector facility. governments. Kalia: quarterly journal for
independent living by disabled people. xii(1),
- Computer projection facility for online
1997, pp.5-6.
materials including video programmes in the
digital library. (9) (a) http://www.abilityhub.com
(9) The maximum number of books each of them can (b) http://www.nau.edu/ind/ATCenter/whatisa
borrow and the maximum period of loan should be t.s html
raised.

57
Health Information System & e-Health Care

58
 AN APPROACH TO LOW COST WIRELESS TELE-MONITORING
SYSTEM FOR RURAL POPULATION IN DEVELOPING COUNTRIES

D. Goswami1, J. Mukhopadhyay2, S. Mukherjee3 and S. Chakrabarti3

1: B. C. Roy Technology Hospital; 2: Dept. of Computer Science & Engineering;
3: G. S. Sanyal School of Telecommunications, Indian Institute of Technology, Kharagpur-721 302, India
E-mail: mahadgos@yahoo.com; jay@cse.iitkgp.ernet.in; meshiblu@yahoo.co.in; saswat@ece.iitkgp.ernet.in

Abstract • Density of population in rural area is

relatively low (India: about 900 per sq.Km in
This short paper focuses on the need for effective urban area and 275 per sq.Km in rural area)
telemonitoring scheme for maintaining rural healthcare
system in developing countries. A low-cost • Low average income and hence is the need for
telemonitoring system has been discussed briefly. strong public health care system
Sample results for an ECG signal acquisition and
• Insufficient road and transportation system
processing unit of a prototype system, under and so tele-monitoring can play an important role
development, have also been presented. in rural health management
Keywords: Telemonitoring, rural healthcare system, • Poor Telecom infrastructure: Historically, due
wireless communication, ECG signal processing to several reasons including economic, benefits of
a new technology reach rural population relatively
1. Introduction late compared to the time of successful
Phenomenal technological developments in recent commercial deployment of the technology
years have opened up new possibilities of improving elsewhere. Telecommunication technology is no
public health care system [2]. Developments in exception (teledensity in India: urban 18% and
computing, signal processing and communication rural < 2.0%, though plain old telephone system
technologies have resulted in new diagnostic [POTS] has been deployed in public domain
instruments with better precision and their networking several decades back. The scenario, however, has
with computing machines, communication gadgets and been improving considerably over the last decade)
databases [1]. In short, physicians and supporting staff, [5].
responsible for health care system, now have the option • Lack of efficient electricity distribution
of availing networked resource bases and closer system: This is another impediment in the
interaction within the community. However, many of deployment of various other technology-based
these technical marvels are young by age and are not services including health care service.
yet available to all potential users at affordable cost.
• Scarcity of specialist care in the primary and
In this paper, we specifically focus on rural areas secondary health care system necessitates the
in developing countries and discuss the possibility of establishment of telelink to enable skilled or
designing a low-cost tele-monitoring system for semiskilled health workers at the periphery to
facilitating rural health care system. It is imperative consult the tertiary care centers in emergency
that no public health care system in a country can be situations. This will also help reduce referrals and
complete without a strong and well-knit rural (primary) lessen the burden on the tertiary centers by
health care system [8]. developing more confidence amongst the health
In this context, the following observations and workers at the periphery.
views are pertinent for appreciating the design • The above features and views amply indicate
approach explained later. that a meaningful and efficient health care system
Targeted Service Area: should have a health tele-monitoring system as one
of its components. This is specifically because of
• The rural population in a developing the following reasons:
(including India) is usually spread over a vast
a) a tele-monitoring system helps in collecting
portion of the landmass; some areas are remote or
and systematically maintaining health data base
inaccessible.
without massive deployment of manpower. This
• Majority of the national population may be is useful for obtaining various statistics related
classified as rural (e.g. India have about 28% to public health. Further, this database
urban population and 72% rural population) [4]. eventually will be of significant interest to
researchers.

59
 b) Common health problems of rural India are: on a real-time basis. However, because of larger
communicable diseases, nutritional deficiencies, volume of data , some other physical signals such as
sanitation problems, medical care problems and chest sound and colour image may have to be
population problems. transmitted on a non-real-time basis using the
A major use of a telemonitoring system is for suggested transmission rate of 16 Kbps. We propose
screening the rural population for various adaptive differential pulse code modulation [ADPCM]
“hidden” diseases which will in turn help in @ 16Kbps for coding audio signals because of low
early case detection (perspective screening), complexity.
control of diseases (prospective screening),
research purposes and to open up educational Fig.1 gives a schematic description of the
opportunities. proposed telemonitoring unit. An operator for both-
c) A tele-monitoring system may also be useful way voice communication with the central station will
in providing preliminary but timely information use the headphone indicated in the figure. The operator
in treating patients. For example, the decision to thus will be able to provide support services to the
operate a victim of a road accident can be taken patient, any participating doctor and the data base
more prudently if the personal health details of administrator.
the patient can be retrieved from a central
database within a few minutes.
3. Some Experimental Results
d) In the year 1977, WHO adopted HEALTH
FOR ALL, and accordingly, principles of Design and implementation of the scheme
Primary Health Care for third world countries described in Section II is in progress. The signal
include: equitable distribution, community processing [Fig.1] will be implemented on a suitable
participation, intersect oral coordination and field programmable gate array (FPGA). Commercially
appropriate use of technology.
available IC-s will be used for the design of radio
Telemonitoring of rural health by the tertiary transceiver at 2.4 GHz. First phase of the design will
health care centers is a pragmatic approach to only involve real-time signal processing and
achieve the above goals. transmission.

Let us note that a tele-monitoring system ECG is an important signal for any telemonitoring
primarily provides a one-way service, from the patient /
system [1]. A low-cost analog ECG amplifier has been
hospital to the monitoring base and does not include
sophisticated and critical features of ‘telemedicine’ or designed. A typical ECG signal obtained from the
‘tele-diagnosis’. As the costs of computing, storage and amplifier is shown in Fig.2 and its frequency spectrum
communication have been decreasing steadily, design is shown in Fig.3. As evident, no involved precaution
and deployment of a tele-monitoring system may now has been taken to remove the 50 Hz interfering signal
be possible at a relatively low cost. Wireless from 220 V AC power line. This problem is not an
communication technology, because of its spectacular important one in rural areas. However, this interfering
proliferation in recent years in the form of cell phones signal can be conveniently processed in the digital
and otherwise, is especially an attractive choice for domain to obtain cleaner ECG signal as and when
transmission and reception of signals. Wireless systems necessary [6,7]. As an example, Fig.4 shows a cleaner
are potentially flexible, easily deployable and varieties ECG signal obtained by the use of a digital filter.
of technology solutions are commercially available [2].
Transmission of such ECG signal using wireless
In the next section, we present an overview of a
transceiver modules [24Xstream from MaxStream] has
tele-monitoring system highlighting the special
features. In Section-III, we present some details on the been successfully demonstrated in laboratory
design approach that has been taken up and some environment.
indicative experimental results.
4. Conclusion
In this short paper, we have highlighted the need
2. Brief Description of The Proposed
for an effective telemonitoring system for rural
Telemonitoring System
healthcare maintenance in developing countries. An
overview of a telemonitoring system for use in rural
Table 1 lists several physical parameters to be and remote areas has been presented. Design of a
monitored, assessment of their storage requirement, prototype system is underway and sample results for a
mode and rate of transmission and necessary low-cost ECG signal acquisition and processing unit
observation duration. It may be noted that several have also been presented.
physical parameters such as temperature, blood
pressure, respiratory rate and ECG can be transmitted

60
References
1. Multi-purpose Health Care Telemedicine Systems
with mobile communication link support, E.
Kyriacou, S. Pavlopoulos, A. Berler, M.
Neophytou, A. Bourka, A. Georgoulas, A.
Anagnostaki, D. Karayiannis, C. Schizas, C.
Pattichis, A. Andreou and D. Koutsouris;
BioMedical Engineering OnLine, BioMed Central,
March 2003; http://www.biomedical-engineering-
online.com/content/2/1/7
2. Wireless Local Loops; Ed. Peter Stavroulakis;
John Wiley, 2001
3. Wireless Communications - Principles &
Applications, 2nd Ed.; T. S. Rappaport;
Pearson Education; 2002
4. http://planningcommission.nic.in/
5. http://www.dot.gov.in Fig.2: ECG signal [Lead II] obtained from a simple
6. Power Line Interference Removal from ECG amplifier circuit. The 50 Hz signal overriding the
Electrocardiogram Using Simplified Lattice Based ECG signal is because of interference from 220 V AC
Adaptive IIR Notch Filter, Santpal Singh Dhillon power line.
and Saswat Chakrabarti, Proc. of the 23rd Annual
IEEE EMBS International Conference, Oct’25-
28,2001, Istanbul, Turkey, pp. 3407 – 3412.
7. On Structural Recognition and Analysis Methods
Applied to ECG Signals, Antti Koski, Ph.D Thesis;
Research Reports R-97-1, ISBN 951-29-0885-9,
ISSN 1235-6727, Computer Science, University of
Turku, Finland.
8. Park Text Book of Preventive Medicine, 15th Ed.

Fig.3: Frequency spectrum of the ECG signal of Fig.2.

The vertical line on the left side of the graph indicates
the strong 50 Hz interfering component. This
Fig. 4: ECG signal (Fig. 2) after canceling the 50 Hz
component can be almost eliminated by using
component. The noisy ECG signal was passed through
appropriate adaptive band-reject digital filter.
a digital band-reject filter to reject the interfering
component at 50Hz resulting in a substantially cleaner
ECG signal.

61
 Sl Physical Signal Typical Observation Avg. Data Estimate Mode of
No. Duration gen. rate of bits Transmission
1. Temperature 1.5 min. 1 bit/sec 96 bits Real time
2. Pulse oxymeter 1.5 min. 1 bit/sec 96 bits Real time
3. Blood Pressure 1.5 min. 400bps, @50sps 18 Kbits Real time
4. Respiratory Rate 1.5 min. 800bps, @100sps 72 Kbits Real-time, cont.

5. 3-lead ECG 20sec. X 3 1.6Kbps, 200sps 96 Kbits Real-time, cont.

6. Chest Sound 10sec. X 2 16Kbps, @8Ksps 1.28Mbits Non-real time

7. Colour image 2/4 snaps 16Kbps 2.56Mbits Non-real time

8. Document scanner 2-4 pages 16Kbps 4 Mbits Non-real time

9. Audio microphone 3 min., 5 spells 16Kbps ADPCM 3 Mbits Real time

Table-1: Various physical parameters to be monitored and assessment of their storage requirement. The assessed
values are indicative only and will reduce considerably if suitable digital signal compression schemes are incorporated
in the system. It will take about 15 minutes to transmit complete information of one subject @ 16 Kbps.

Remote Station Central Station

Telecomm.
Network
Respiratory Database
Blood Rate
Pressure
Comm.
3-lead ECG Signal Condition Interface
Temperature and processing
unit
oximeter Secondary
storage
Audio
Microphone

Headphone

Scanner
Monitor Camera

Fig. 1: Schematic description of the proposed tele-monitoring unit

62
 e Health Care – India’s Unfulfilled Dream
David Raju Manne MA, MHM
Team Leader, Impl-IT,Apollo Health Street,Hyderabad,AP,India
davidraju.manne@gmail.com

Abstract which caters to the needs of all the health care

professionals. Apollo Health Street implemented its
Healthcare is an industry that has for years software not only on the domestic front but also on
remained lackluster towards investing in information international platforms, which are rated as high
technology. eHealth is an egalitarian way of consulting proficiency implementations. The success behind all
your doctor. Technology supports ehealthcare but law this goes to the Apollo Health Street’s platoon, which
and people’s mindset restrict it. What made us to is supported by three wings of specialties like
restrict it? Concepts like telemedicine, eappointments, Healthcare Administrators, Developers and Network
EMR opened the doors for dynamic work systems, but
Administrators.
we entertain them only as a business tool. We do have
a way to utilise the technology. We should develop Asia is the highest Internet user with 31.7% at a
technology and use it in a right way so that the growth of 7.1% growth (Sept 2004). In Asia India
healthcare excels with the support of informatics. stands third with 7.2% of Internet usage [2]. Checking
the health care portals online interactions WebMD.com
Key Words: has 6 million unique visitors per week, 4 of the top 10
Hospital, Patient, Doctor, ehealth, telemedicine, health sites are diet- or fitness –related [3]. India stands
Electronic Medical Record (EMR), Hospital on the limelight and acts as a major health care player
Information System (HIS) in Asian countries. Global eye turns towards us for a
health care aspect. Many claim that the reason for this
is only that we provide medical Excellency at cheaper
Introduction prices. But the reason behind this is not only the cost
“They say that time changes things, but you actually factor but also even the quality of care provided by the
have to change them yourself” Andy Warhol skill full hands of Indians.93% of adults online in
France and US who use Internet to access health
The Internet has often been described as a information find it trustworthy. Americans generally
disruptive technology, defined as having the ability to give higher comprehensibility and quality ratings than
break through barriers and interrupt the normal course French and Germans [4].
of activities. Healthcare is an industry that has for years That’s the global perspective of ehealthcare. Many
remained lackluster towards investing in information keep referring about ehealthcare but dealt in a
technology. Therefore, the arrival of the Internet in superficial way. This paper “E Health Care – India’s
healthcare appears very much to be the meeting of the Unfulfilled Dream” will focus on the concept of
irresistible force and the immovable object. ehealth care its roots in India, its features and
Generalities aside, we believe the Internet has the advantages, why its been overlooked.
potential to make fundamental improvements in
today’s healthcare system. The pace of activity in Concept of ehealth care
eHealth over the past year has been truly dizzying. The
sheer volume of announcements has made it difficult at e-Healthcare is defined as a way of delivering and
times to separate true potential from hype, to draw achieving better health outcomes through effective and
distinctions between competitors, and to empirically innovative use of health information. It includes in the
value companies. Over the near-term, we expect the health sector "the use of digital data-transmitted, stored
pace of activity to continue and perhaps to accelerate. and retrieved electronically-for clinical, educational
Over the longer term, we believe the sustainability of and administrative purposes, both at a local site and at
eHealth sectors will depend entirely on advancing a distance. eHealth initiatives are designed to present
fundamentals [1]. proactive, coordinated, and evidence-based healthcare
where the clinical, social and technological issues are
Apollo Hospitals Group stands as a pioneer in the combined to create a flexible patient-focused health
health care delivery in South Asia. Apollo Health care system[5].
Street Ltd, a subsidiary of Apollo group is pioneering
in health informatics. Apollo chose to develop its own The aim of eHealthcare is to provide high quality
health informatics center as it has domain expertise in health care to all health consumers; to increase home
health care delivery and it knew what makes the care by remotely monitoring chronically ill patients at
delivery process a quicker, qualitative and professional their homes; and to reduce the need for hospital care
by embarrassing information technology. Apollo for patients. It is also to develop preventive health
Health Street develops software for healthcare entities, education and through the use of information

63
technology reduce errors, waste and cost. This is being information mismanagement by the hospital staff could
achieved through the interchange of collaborative be omitted or discouraged.
multiple health care teams across regional, interstate
and international boundaries and by the projection of In India e appointment is highly discouraged.
specialist medical and surgical expertise to rural and Though many of the HIS support this but the utilisation
remote areas. It is also being developed by the instant is minimal. It’s an interesting issue that though it has a
access to comprehensive, secure, reliable and lot of potential advantages both to the hospital
standardised health records; the integration of hospital, administration and the doctors this is being overlooked.
community, insurance industry, pharmacy, By adopting this technology we could reduce 60% of
government, home and educational health management the complaints in relation to the outpatient services in
systems and the provision of computer based training concern to the management, patient or the doctor.
programs to health professionals
Electronic medical record
Why ehealth for India? For instance Mr. X, an old patient with cardiac
complaint collapsed in his office and ambulance was
Indian hospitals are on a march towards electronic called for a rescue. Fighting the heavy traffic
technology. Health care providers are looking towards emergency technician landed at the spot. Knowing that
an information system, which give them a track of their he is a cardiac patient EMT was hesitant to start the
daily operations. Every day a lot of discussion keeps treatment, as he has to know the present course of
happening on the health care systems. But maximum drugs and whether the patient is allergic to any of the
talks only the operational part of the system with full drugs that he is going to administer. When he calls up
concentration on their administrative advantages. None his emergency center for help in regard to his medical
of the healthcare discussions have taken a view of history, they are on their toes in tracking his medical
quick and effective treatment to the patient. All the record with the help of his registration number, rushed
major hospitals are equipped with the cutting edge to the records room quickly, pulled out the record and
technology hospital information systems but to the rushed back to emergency unit to give it to the doctor.
maximum they support your minimal operations only Emergency doctor analyses the record and refers back
i.e. billing and in patient services. None of them have a to his EMT in regard to the treatment. Just calculate if
broad support to the inventory management or to the Maurice green, worlds fastest sprinter runs for us to
doctors systems or to the patient systems. It’s used as fetch this medical record which usually stored at level
only after service entry system. –1 in many hospitals, minimum time for all this
In India maximum no of patients are aware of the process he would take 10 –20mts.so what happens
diagnosis their doctor is going to reveal. On the above during this time? What happens if he develops another
they will be shooting questions to the doctor regarding stroke? This same scene might happen even in the
the treatment plan. All this is happening because the emergency ward in the hospital itself at 9:30pm, where
way information and communication technology is the patient’s consultant is away from the hospital and
developing in a rapid way. Health care customer is only one staff is posted for the night proceedings at
knowledgeable about the disease he is suffering from. medical records department. We claim that Indian
health care is supported with the state of art
As told earlier we are looking e health care only as a technology. We are doing the most advanced surgeries,
tool to keep us on track regarding the daily operations. but none of us have a count of how many patients we
But give less importance to the aspect of how best and are loosing daily like this. How many of them have
fast we can deliver the care to the patient. developed severe drug reactions if a pharmacist
mistakes the doctors scribbling from .50mg to
Features of eHealthcare 50mg.what happens if we fall sick when we are
traveling? Nothing but needs to eat the radioactive
eAppointments
material again for diagnosis. So what’s the solution for
e Appointment technology is widely used all this? Electronic medical record is the answer.
system in the US. This is a vibrant technology that
helps both the doctor and the patient in fixing their What is EMR?
consultation appointments. With the support of this A provider-based electronic medical record that
technology the doctor who practices at multiple clinics includes all health documentation for one person
or hospitals would be able to view his day scheduler on covering all services provided within an enterprise [6].
line. It’s a hassle free system to plan his clinic and All practitioners within a provider organization can
omits any discrepancies of favoritism among the panel access and input all information through computers on
of the doctors. a paperless basis. The system is the basis for decision
Patient gets benefited, as they would be able support and other computer-base functionalities (e.g.;
to view the doctor’s appointment chart and pre-plan order entry, charge capture, etc.) adopting uniform
their appointment the said time and place to their standards.
convenience. Mistakes like data missing and

64
 Adapting to EMR technology we can pace up our collect the sample and email your investigation report.
daily operations in hospitals. This will provide an win Sitting in your office, you can do an online chat or
win strategy for hospital administration, doctor, patient email your doctor for advice on the results. This is all
and to the third party players like insurance or medical possible if we opt and encourage ehealthcare.
billing institutions. Legal issues are definitely a
hindrance for this aspect, but has a solution for that.. At e QUERIES
the global arena, US started EMR in 1985,being
debated for 20 years, it won he heart of millions of “The mind is its own place, and in itself can make
voters by enrolling EMR as electoral agenda. If we also heaven of hell, a hell of heaven.”John Milton
claim our self that we need to have 20years for this
then we following “Those who cannot learn from I agree with Milton if we have a small doubt
history are doomed to repeat it”. We agree with the troubling our mind, it’s a risky affair to keep it cool. If
global health players and try to bring out their JACHO its concerned to health, you just can’t concentrate on
to our land for the business purpose. on the same if we any work till we clear the doubt whether the pain
look for HL-7 & HIPAA I think we meet the security developed in your chest is related to heart or the effect
standards for patient confidentiality and rewrite the of the spicy food you ate last night. Do you want to go
law. to the doctor and consult him for this? You can get the
Paperless hospital can bring you effective and doctors advice by pressing that killer button on your
timely services. With the Dicom Eye technology there’s keyboard and be happy at no time. Not to raise your
no need for a patients to carry his bulky films every eyebrows and say, is this possible? We are already
time they travel outstation. Work for hospital staff, practicing this method but not in a rapid way.
TPA’s & Doctors becomes a cakewalk. Apollolife.com, doctorndtv.com are good electronic
medical query Indian players. We just need to stretch
More than any one central health department can our services a little.
have an accurate statistics of census and disease
graphs, which would give them a chance to plan for a You can converse with your doctor happily with full
disaster in a better manner. security by logging into these sites. If you request for
privacy, they honor your request. They even provide a
E prescribing chat platform with the specialist doctor free of cost.
E prescribing refers to the use of computing devices
to enter, modify, review or communicate drug e Payments
prescriptions and sign prescriptions .E prescribing
opens doors to the physicians to communicate directly It’s a known fact that the health insurance
to the pharmacist. Missing, duplication or wrong revolution had started in India. Even a common man
interpretation of prescriptions by the pharmacy staff as can dream of entering into corporate hospitals. He can
I mentioned in the EMR example will come to a halt .It just flash his insurance card, get his treatment and walk
even widens the scope of business and customer care off. Then starts the struggle to retrieve his payments
by the pharmacy as physician can transfer the from the insurance provider. If any thing goes wrong
prescription directly from his system, so by the time hospital management will have a tough time in
patient reaches the pharmacy medicines packet waits pacifying the doctors regarding payments for the
for him. Through this we could put a control on services he offered to the patient. All this paper work
scheduled drugs and non-scheduled drugs. process runs to a minimum of 20 days if any hitch
All the pharmacies could be linked through ASP arises 30-45 days. Also on a fair side many of the
technology and advice the patient to buy the medicine health institutions especially medium sized discourage
from a particular pharmacy, if it’s not available with this insurance process, as this is a time killing process.
them by looking at the inventory stock list. Patient can
log on and request for drugs by referring his What if we develop a form like UB-92 or HCF
prescription number and drugs could be delivered at 15000? Why not? Even our health care market is as big
home with medication advise notes. Looking at the as the US. Even if we look at the medical tourism
issues of law, pharmacy would be able to view only the statistics we see that the graph shoots up like a rocket.
prescription of the patient not his EMR. The reason for this is nothing but insurance providers
smelled the fragrance of low cost at same quality.
e Labs
If we also go for epayment for insurance claim
Everyone started realizing the importance of time. If
settlements not only our health care market increases
one needs to check their blood sugar levels, doctor
but also the claims processing time comes down
suggests to give sample early in the morning, this is the
drastically as the insurance company directly interacts
busiest hour for any working personality to keep away
with the doctor on line instead of faxing or couriering
from diagnostic center or hospital. Apart from this,
the documents or he personally makes his presence. It’s
collection of report and consulting your physician for
a happy environment for doctor, hospital and insurance
his review and advice going to strenuous. Why do you
company, which will open doors to the major
want to waste so much of time when you are able to do
population to access the cutting edge medical
all this by sitting at one place .Do an internet booking
technology at the lowest premiums.
with your local labs so that they come to your home,

65
Telemedicine interpretation, medical coding, medical billing and so
on.
“Words are, of course, the most powerful drug used by
US currently follows ICD 9 codes but Indian BPO’s
mankind” Rudyard Kipling
(Apollo Health Street,hyderabad) predicting the future
If you panic, that ehealthcare only provides you of the Us we work on ICD –10 for them. We work on
with the mails, you are absolutely no the wrong side of HL-7 and HIPAA standards for them but hesitant to
the shore. Your doctor sitting in New Delhi can speak adopt them. Apart from this for the US &UK we play a
to you if you are attending a conference in New York major role in developing software for hospital
and discuss your health and take advice from him. All information systems, eAppointment, econsultation
this is possible with the help of telemedicine. You &EMR’s with high security standards but we don’t
might reply back that telemedicine is a costly affair and design for our self. Think its time for us to address this
equipment and configurations problem. Now your quotes, “we cannot hold a torch to light another’s path
problem is solved. You can converse with your doctor without brightening our own”. Ben sweetland.
happily if you have a web cam. It’s a basic thing any
Lastly “Let us not look back in anger or forward in
net cafe would give you. At the cheapest happily
fear, but around in awareness” James thurber. Check
discuss your issues with your doctor.
the rate of literacy growth in India? Check the way the
To list it out the advantages of ehealthcare it runs to technology sprinkling around. Check the way
pages. Healthcare in India is highly discouraged by telecommunications reached to the nook and corner of
focusing on reasons like – law doesn’t permit, privacy India where a government bus can’t reach. What makes
of the patient is at risk and India is a poor country it our dream of ehealthcare to sleep in silence? Why do
doesn’t have facilities .let me address this issues. you need only a personal computer to access the
ehealthcare? All the mobile network providers have a
Lets go back to 1980’s,only business tycoons, cine
facility of Internet. With entrance of BSNL &
actors and rich class could afford a cardiac bypass
Reliance every third person in India holds a mobile.
surgery in developed nations. Upper middle class and
They read their mails, chat and shop through Internet
others have to survive till your medicines support you.
from their mobile phones. Leave alone mobiles 94% of
That time
Indian homes are equipped with television sets with a
Dr Pratap C Reddy, Chairman of Apollo Hospitals cable connectivity and maximum cable operators have
Group, India, fought with the government and at last the internet options in their net work, you just need a
was able to get the green signal from parliament and technological support from your television to access
engraved an affidavit in the Indian law. He made the internet on your television sets. We might end up in
corporate health care possible for Indian population saying it’s costly. Lets remember the basic laws of
that is saving thousands of lives. Added to this Dr A P economics of demand and supply. Year back DVD
J Kalam, President, India, came up with the stents for player used to cost Rs15000-20000/-, users increased,
cardiac patients and prosthesis implants and calipers in now cost stands at Rs3, 500/-. So now its time for us to
orthopedic surgery brought the cost drastically down. ‘Think big, think fast, think ahead. Ideas are no one's
Joining these insurance agencies started Mediclaims, monopoly” once u have an idea crack it so that the rest
which made the finest health care to the doorsteps of proceeds.
many. Dr Pratap C Reddy started telemedicine at his
So now "Our dreams have to be bigger, our
village Arogonda, connecting all the Apollo’s with
ambitions higher, our commitment deeper and our
Arogonda where the government road hardly
efforts greater…” so the unfulfilled dream of
connected with highway. This all happened because
ehealthcare for Indians becomes a reality. Let us march
they knew the importance of health and fought for this
ahead towards excellence in ehealthcare with
better today which should shine tomorrow. Why can’t
informatics touch.
we take a lead for the e healthcare law? How far we
keep spending on things, which could be overtaken by References:
technology at less cost which will make the system
1. Richard D. Lee, Healthcare and the Internet in the
dynamic.
New Millennium, ehealth 2000, WIT Capital,
Addressing the second issue of privacy of the January 31, 2000.
patient is at risk. Let us confess our self, how far are 2. Internet Usage Statistics - The Big Picture, World
we able to protect the privacy of the patient? We Internet Users and Population Stats,
follow the ISO standards, how far has it addressed the http://www.internetworldstats.com/stats.htm
issue of patient privacy? Right now we are again 3. Jupiter Media Metrix, “The National Institutes of
looking at JCI or if our Indian health care federation Health”, May, 2002.
comes up with a common standards for India. How far 4. Harris Interactive, June 11, 2002.
are we going to address these issues? Funniest part is 5. Deloitte Consulting and Deloitte & Touche, “The
that we claim that we don’t have technology for all this emergence of the e-health consumer”, US, 2000.
but through out the globe we hold the major stake of 6. Waegemann C Peter, Strategic Considerations,
BPO’s which process the medical reporting, EHR Summit,2003

66
 MED-CENTRE - DRUG INVENTORY MONITORING AND
CONTROL SYSTEM
G.S Bansal, S.S.Duggal, Rahul Jain
National Informatics Centre, Haryana State Centre
&
Dr. Ravinder Geol,Anil Aggarwal
Directorate of Health Services, Haryana
Haryana Mini Secretariat,Sector-17
Chandigarh – 160017, INDIA
rahul@hry.nic.in

Abstract It is observed that there is perceptible increase in

proper utilization of medicines and number of patients
Usage of Information Technology (IT) remained receiving affordable treatment at Government Health
comparatively very less in Health sector than other Institutions after the implementation of the software
sectors despite having more potential. Health package
Institutions, which is important sector, should
encourage IT usage resulting better productivity, 2. About Med-Centre
effectiveness, efficiency and economics leading to
better health care to all. This paper tells a success story Med-Centre of Haryana is an integrated software
of implementation of ICT in monitoring of medicine in project to capture utilisation of medicine inventory data
Government health Institutions in Haryana and analysing consumption pattern of various medicine
implemented by National Informatics Centre, Haryana location wise to monitor disease occurrence pattern,
State Centre for Directorate of Health Services, pilferage and any other exception in the functioning of
Haryana. System works on low-end resources and E- health institution. Reports generated at various level act
mail based data transfer from District HQ to State HQ. as helping tools for multipurpose analysis and
Implementation of Med-Centre in all the districts of inferences in order to provide affordable health care to
Haryana results in check on pilferage of medicine, common man by monitoring the flow of Medicines
increase in availability of medicine at Government
Institution, increase in attendance of patient/doctors in Med-Centre capture the medicine data for 316
health institution, optimal utilization of medicine and medicines under 23 various categories from 619 health
data capturing at source, resulting in availability of institutions (54 General hospitals, 74 CHCs, 401 PHCs
error-free data at State Head quarter. and 89 others health institutions)

Key Words
BASIC FEATURES
Medicine, Inventory Monitoring,, NIC, DGHS, • Based on Medicine flow, cost, consumption,
Haryana and Med-Centre OPD/IPD and utilization of Medical Staff
analysis is done
1. Introduction
• Easy to install, operate and maintain.
Distribution and utilisation of medicines
(Annexure – I ) in Government Health Institutions is a • E-mail based data transfer.
complex procedure, involving lot of functionaries at • Runs on low-end resources.
various levels. Monitoring and controlling the flow of
medicines in such an environment is a challenging job. • Useful tool for Health Administrators,
Doctors.
It is observed that due to improper monitoring of
medicine flow, there remains a big chance of pilferage • Close monitoring of medicine flow increases
and under-utilization of medicines. availability of medicine resulting in increase
of inflow of patients.
Monitoring of medicine flow is of prime
importance for the authorities concerned. NIC-HRSC • Includes Generic and non-generic medicine.
has developed a software package Med-Centre for this
purpose, which has been highly appreciated by the • Affordable and quality treatment within the
Health Authorities and has been implemented existing resources
successfully in all the districts of Haryana and State
Head Quarters.

67
2.3 IMPACT OF MED-CENTRE 3.3 Exceptional Reporting
• Check on pilferage of medicine. During the analysis of various reports, statistical
• Increase in availability of medicine at parameter like average of OPD consumption, average
Government Health Institutions. patient per doctor etc are compared with these figures
• Increase in attendance of patients/ doctors in of institution and in case of large deviation, the reasons
health institutions. for the same are analysed. From the actual data
• Optimal Utilisation of Medicine. analysed from the system it was found that the reason
• Data capturing at source, resulting in error- of large deviation were :
free data at State HQ.
• More transparency in the functioning of 1. In some cases, there was pilferage of medicines in
Health Institutions. the institution showing very high
consumption/OPD while in some cases the reason
for the same was outbreak of particular disease in
3. Methodology for checking pilferage and that locality.
performance of health institution
2. The institution showing very low consumption
3.1 Data Collection and Compilation ( See picture) /OPD were found to be due to absentism of doctor
in that institution.
Software is installed in all the districts Civil
Surgeons offices and State Health 3. The institution showing very low patient/doctor
ratio shows under utilisation of institution which
Directorate. Medicine Inventory data is collected could be due unpopularity of the Govt Institution
on prescribed format on monthly basis. This data is fed and availability of Private Institution.
into system in first week of every month at Civil
Surgeons office by data Entry operator and using the 3.4 Correctional Action based on exceptional
export utility of the system a file is generated, which is reporting
later sent to Directorate of Health Service (Head
Quarters) by using Email. 1. Defaulter Institution in sending medicine reports
th
At State Headquarter, on 10 of every month the was sending the reminder to send the reports. Civil
data is compiled and defaulter institution is sent a Surgeon’s concern in district was informed about
reminder to send the data. On every 12th of every the defaulter of the districts.
month the final compilation is done.
2. Cost Analysis, Consumption analysis and
3.2 Report Generation IPD/OPD analysis are another tools to monitor the
performance of the institution. It is observed that
Various reports available with software are each district civil surgeon need ½ hour to analysis
generated and put up to Doctor, attached with IT cell and draw inference regarding the performance,
for analysis purpose. Doctor draws various inference which is otherwise never done manually.
from these reports by comparing figures like
consumption of health institution with state average
and similarly other statistical parameters are observed 4 Figure and Tables
and inference are drawn from the data of Institution.
Figure 4.1: OPD Pattern After implementation of Med-
Centre

68
 Jan Feb Mar Apr May June July Aug. Jan Feb Mar Apr May June July Aug.
Ambala 39738 34813 46099 32266 42106 51173 59318 55924 Ambala 8.3 7.88 5.94 7.47 7.77 6.43 6.53 8.33
Bhiwani 50204 54187 60794 39087 90676 84554 93453 51629 Bhiwani 9.1 5.63 5.28 5.67 4.67 4.92 5.36 4.23
Faridabad 40487 53046 63574 58462 63742 78623 91348 82039 Faridabad 8.2 4.71 4.02 4.38 6.5 5.16 5.17 4.94
Fatehabad 10.09 6.36 18.99 6.89 6.86 7.11 6.46 7.95
Fatehabad 30520 31605 39515 32618 39118 38976 46118 52858
Gurgaon 4.55 5.51 5.68 6.12 8.63 6.11 7.6 4.4
Gurgaon 41215 42344 63908 65061 49691 61690 79305 92812
Hisar 4.42 4.47 10.7 6.08 6.1 5.88 5.3 5.39
Hisar 51181 50590 53409 47918 52692 63106 79098 73917
Jhajjar - 7.48 9.76 8.48 8.3 6.17 6.52 5.91
Jhajjar 0 23694 26168 25542 25949 32505 35294 35780 Jind 4.89 6.59 6.62 6.79 6.03 5.32 7.04 9.26
Jind 34083 33900 54029 43846 52060 46255 62663 64555 Kaithal 7.4 6.47 4.94 5.8 10.86 5.93 4.94 5.72
Kaithal 17666 32850 43545 28524 38593 32262 53829 57092 Karnal 6.18 6.9 3.72 5.21 2.96 5.1 4.67 4.23
Karnal 38506 24078 57615 56076 82606 76590 76993 86321 Kurukshetra 8.96 7.49 4.78 8.24 6.07 6.66 7.29 5.32
Kurukshetra 24286 25175 38611 30386 33658 36413 40827 44516 Narnul 7.32 4.19 4.8 8.36 8.29 5.19 6 6.1
Panchkula 7.67 7.31 7.36 9.07 8.3 7.19 6.44 7.97
Narnul 25846 23805 39344 35143 43387 47046 47170 37716
Panipat 5.66 5.01 6 7.14 7.66 7.28 4.35 5.44
Panchkula 44033 41606 58489 49939 53383 65844 76978 67120
Rewari 12.35 7.59 17.69 5.98 9.65 7.13 10.01 6.12
Panipat 33600 29336 39341 31183 38673 39191 42572 49835
Rohtak 8.25 8.14 8.27 7.61 7.66 7.73 4.76 5.01
Rewari 20256 47431 33935 29532 29135 39944 28578 57230 Sirsa 10.69 9.96 5.8 6.05 6.42 5.89 5.71 4.91
Rohtak 26520 37206 42802 39277 45281 52961 62694 63021 Sonepat 9.42 6.76 4.02 9.62 7.72 5.88 4.87 4.63
Sirsa 25485 26623 43729 31326 40033 48192 49936 47945 Yamunanagar 12.98 9.94 10.8 11.48 24.98 13.18 14.1 13.06
Sonepat 28409 34438 53818 19945 55910 51601 65505 66954 Total 7.8 6.63 7.22 6.94 7.21 6.21 6.9 5.97
Y’nagar 27207 35926 44313 33757 18356 21082 27654 26243
Figure 4.3 : Cost/OPD for the Month of August 2004
Figure 4.3 : Cost/OPD for the Month of August 2004

Above observation shows there is subsequent increase

in OPD figure in Government Health Institution after
implementation of Med Centre in Health Department.
Around 85% increase in OPD patient attendance.

Table 4.1 - District-wise OPD Figure in Haryana

Health Institution
0 – Non Repoirting District (Jhajjar in Jan-04)

Figure 4.2: District Wise Comparison of COST/OPD

Patients for the Year –2004

There is about 25% decrease in cost/OPD despite there

is increase in Patient attendance cost/OPD decreasing
from Rs 7.80 to Rs. 5.97

69
 Cost of
Number Of Medicine
Cost /
District OPD Consumed (
OPD
Patients Rs. in
Annexure-I
thousand)
Ambala 55924 466 8.33
Med-Centre- Domain View
Bhiwani 51629 218 4.23
Faridabad 82039 405 4.94
Fatehabad 52858 420 7.95
Gurgaon 92812 409 4.40
Hisar 73917 398 5.39
Jhajjar 35780 211 5.91
Jind 64555 598 9.26
Kaithal 57092 327 5.72
Karnal 86321 366 4.23
Kurukshetra 44516 237 5.32
Narnul 37716 230 6.10
Panchkula 67120 535 7.97
Panipat 49835 271 5.44
Drugs Procurement
Rewari 57230 350 6.12
Rohtak 63021 316 5.01
Sirsa 47945 235 4.91
Sonepat 66954 310 4.63
Yamunanagar 26243 343 13.06
Total 1113507 6645 5.97
Cost/OPD comparison of District with State Avg
Cost/OPD
(Note: Unlined Cost/OPD is above state Avg
Cost/OPD)

Table 4.3: Statement Showing District Wise Cost/OPD

for the month of August-2004

5. References:
Drugs Distribution
Procurement Manual (2001), Material Store
Department (MSD), Haryana.

Haryana At a Glance (2004), Directorate of Health

Services,Haryana

Niels Warrer ,The Escapism of Operations Research,

David Brown Book Co

David Bordwell ,Fundamentals of Operations

Management, McGraw-Hill College

Drugs Utilisation

70
 DATA MANAGEMENT SYSTEM FOR COMMUNITY-BASED
RANDOMIZED CONTROLLED TRIAL IN INDIA

Shivaprasad S Goudar,
Prof of Physiology; Research Coordinator, Global Network for Women’s and Children’s Health Research Site 8;
& Secretary, Dept. of Medical Education, J N Medical College, Belgaum 590010 Karnataka India sgoudar@jnmc.edu
Margo F Brinkley,
Manager, Research Computing Division,
RTI International, 3040 Cornwallis Road, P.O. Box 12194, Research Triangle Park, NC 27709-2194 USA mfb@rti.org
David G Forvendel,
Research Computing Division,
RTI International, 3040 Cornwallis Road, P.O. Box 12194, Research Triangle Park, NC 27709-2194 USA dgf@rti.org
Arthur Pradeep Singh,
IT Support Engineer, J.N.M.C - U.M.K.C Women’s and Children’s Health Research Unit,
J.N.Medical College, Belgaum 590010 Karnataka India arthur@jnmc.edu
Amit P Revankar,
Asst. to Research Coordinator, J.N.M.C - U.M.K.C Women’s and Children’s Health Research Unit,
J.N.Medical College, Belgaum 590010 Karnataka India amit@jnmc.edu
Bhalachandra S Kodkany,
Professor of Obstetrics and Gynecology, Chief Investigator, HRRC Center (ICMR)
J N Medical College, Belgaum, Karnataka, India drkodkany@jnmc.edu
Richard J Derman,
Associate Dean for Women’s Health; Schutte Chair in Women’s Health; Professor of Obstetrics and Gynecology;
University of Missouri at Kansas City, Kansas City, MO, USA dermanr@umkc.edu

Abstract Key Words

Electronic data management systems offer unique Data Management System, Edit checks, Double key
advantages over paper based data collection systems, verification, Audit trail, Randomized Controlled Trial
especially for implementing and monitoring
community based research projects. Until recently, data Introduction
collection in the health sector in rural India has Traditionally, data collections for community-
primarily been paper-based. Ensuring data quality is based research activities have been limited to paper-
critical for obtaining meaningful inferences from such based methodologies. The most common approach has
studies. An electronic Data Management System has been a centralized paper-based system wherein
been successfully developed and used for an ongoing completed data collection instruments are brought to a
Global Network for Women’s and Children’s Health central location and converted into electronic records
Research community-based randomized controlled trial using a variety of methods including imaging of paper-
in rural India. The system incorporates features such as based forms and data keying in a computer. [1]
edit and range checks, between and within form However, recently Data Management Systems (DMS)
consistency checks, double key verification and an have been employed not only for capturing information
audit trail log for capturing valid and reliable data. electronically but also for providing biostatistical, data
Reports are generated to monitor and evaluate progress processing, and administrative support for randomized
of the study and to provide feedback to field staff to clinical trials. With the advancement of
improve enrolment of study participants, ensure data telecommunication technologies and the availability of
quality and achieve compliance in data collection at personal computers at low prices, the distributed data
multiple time points prescribed in the study protocol. entry model has been widely adopted for data
Limited computing skills of study personnel and collection and management in large-scale multi-site
difficulties in executing a distributed data entry system clinical trials. [1]
were the major barriers in implementation. Erratic
Clinical trials may involve the administration of
power supply and poor quality telephone connectivity
new or experimental treatments to patients and the
necessitated a switch from distributed to centralized
subsequent observation of responses to these
data entry. Training of personnel and use of wireless
treatments over appropriate periods of time. During a
technology may overcome these barriers.
clinical trial, large volumes of data describing the

68
course of each patient must be gathered and analyzed. managing clinical studies have been attempted.
Until recently, computers have been used only for final However, these have been largely confined to hospital
statistical calculations after labour-intensive data based clinical studies. Several agencies, including
capture and tabulation. These methods are becoming those of the central and provincial governments, have
increasingly expensive, and problems with traditional conducted community-based research projects.
data management techniques in clinical research are
The Indian Council of Medical Research (ICMR),
compounded by trends including increasing numbers of
through its network of Human Reproduction Research
patients, increasing length of follow-up period,
Collaborating Centers (HRRCs), has been a pioneer in
increasing numbers of treatment and response
this area. The GN Site 8, a member of this network for
variables, and participation of geographically dispersed
the past two decades, has participated in the
research groups in 'multi-centre' trials. [2]
community-based research projects initiated by that
The authors report on the design and network. The centrally-developed research protocols
implementation of a DMS for an ongoing Global are implemented at the community level and data
Network for Women’s and Children’s Health Research collection activities are supervised by qualified field
(GN) community-based Randomized Controlled Trial supervisors. [7] The data forms are brought to the
(RCT) sponsored by the National Institute of Child HRRC centers and keyed using a data entry program
Health and Human Development, USA (NICHD).[3] developed for that specific project using EPInfo
The Global Network comprising 10 US academic software. The centers then transmit the data to the
institutions and their collaborating partners from the central coordinating office of the project for analysis
developing world was constituted to implement and reporting. In the HRRC system, the data entry
community-based research projects for reducing program generally does not have error checking
maternal and neonatal mortality and morbidity. GN capabilities and the centers do not usually generate
sites are located in South America, Africa and Asia and reports for monitoring protocol implementation and
J N Medical College, Belgaum, Karnataka, India reviewing the project’s progress. In studies requiring
(JNMC-GN Site 8) is one such site. The primary goal data collection at multiple points in time, the paper
of the GN Site 8 trial, which is conducted in four forms are retained by the field supervisors till all the
Primary Health Center (PHC) areas of Belgaum data for an eligible participant has been recorded. A
District, Karnataka, India, is to assess the effectiveness single form captures all of the information, from
of Misoprostol 600 mcg orally in reducing the screening through randomization to outcome
incidence of acute Post Partum Hemorrhage (PPH) i.e. measurement. Completed forms are then manually
blood loss ≥ 500 ml in women delivering at home or in scrutinized and keyed periodically. This precludes
neighboring sub-centers. timely error resolution and determination of
Table 1: Study Progress 9/11/2002 to 15/12/2004
compliance with study protocol.
An attempt has been made to develop and utilize a
Sample Screened Eligible Randomized Completed
DMS not only for data capturing but also for assuring
1600 3122 1938 1095 998 data quality and reporting and monitoring the progress
of the study. Many of the features recommended in the
Auxiliary Nurse Midwives (ANMs) screen,
industry guidance for computerized systems used in
randomize and follow-up study participants. Data is
clinical trials by the US Department of Health and
collected on paper-based forms and submitted to field
Human Services have been incorporated in the DMS.
supervisors on a weekly basis. After verification, the
[8] The DMS was developed by local experts in
data is double-keyed and transmitted to the Data
collaboration with their counterparts from RTI.
Coordinating Center at Research Triangle Institute,
North Carolina, USA (RTI). The features of the DMS
to ensure accuracy of the data, use of its capabilities for Data Management System
generating reports to monitor the progress of the study A computerized DMS was developed to
and challenges encountered in implementing a accomplish electronic data capture for participants
distributed data-keying program are also discussed. enrolled to a community-based RCT in rural India.
Initially an attempt was made to implement a
Background
Distributed Data Entry (DDE) system with a more
In recent years, India has attempted to harness versatile and sophisticated data entry program capable
innovative information technology solutions for a of checking out of range data and missing data at the
variety of applications in the health sector. However, time of entry in the field along with systems to transmit
for the most part, these efforts have been restricted to the data by telephone lines to the central facility at
urban tertiary care settings. The primary health care JNMC. The basic intent of DDE was the creation of a
set-up continues to mainly depend on paper-based computerized database. The data entry component had
record keeping although the Health Management the capability of recognizing and marking incorrect or
Information System has been developed and adopted inconsistent data. There was very little report
for capturing health sector related data from the generation capability in the system.
villages. [4, 5, 6] Computer based systems for

72
 Fig 1: Distributed Data Entry and Data Flow for performing study related tasks. Hence, the access
levels have been defined based on their needs.
A transition from a DDE system to a DMS Administrators, Data Entry Operators, Double Key
requires additional capabilities to generate the needed verification personnel, the Project Coordinator and the
Investigator have access to the system and utilize
specific features customized for their needs. The study
personnel using the DMS were trained initially at the
time of its commissioning and subsequently as the
system was upgraded with additional features.
Ensuring data quality is a fundamental requirement
for any research study. The commonly used manual
administrative and other types of reports based on the methods like vigilance and accuracy during data
data already available in the computer database. The collection and keying, primary "sense" checking of
DMS focuses on the processes involved in data records, visual verification of all records, and
collection, data quality assurance, creation of a independent resolution of all discrepancies were
computerized database, and generation of regular and supplemented by preprogrammed range checks as well
ad hoc reports from the database needed for as consistency checks between fields, both within and
implementing the study. Timely availability of reports between forms. The database includes extensive
is a factor that helps in improving the efficiency of the information on maternal medical conditions, labor and
conduct of the study. The DMS facilitates the efficient delivery events, and neonatal outcomes as well as
management of the study, through its capabilities for information on life-style and demographic factors.
generating routine and ad hoc reports and provides Periodically, other consistency check programs are also
evidence to take timely corrective actions as needed. In run as part of an ongoing data quality assurance
the absence of a computer based DMS, some of these initiative and have contributed to validity and
reports must be generated manually. A well designed reliability of the information gathered. The DMS is
DMS avoids the need for a majority of these manual designed to capture any changes effected in the
systems. database automatically and these are catalogued and
securely stored in a database. Data fields are locked
The DMS was developed using Microsoft® Visual with a complete audit trail of cell and configuration
Basic Version 6.0 ™ as the front end and Microsoft® changes for subsequent review. Additionally, batch
Access XP Professional Version as the back end tool. edits are also run periodically to verify the data quality.
The user interface includes controls such as list select
boxes, check boxes, and radio buttons. It provides for: The DMS was installed on computer terminals
data editing (validation) during entry; incorporating located at the PHCs. Data forms turned in by the
automated skip (or conditional fields) rules; the ANMs were keyed into the database and then
capability to flag fields or records with problems; the transmitted to a central server. The server, a Compaq
ability to annotate fields; comparing required and ProLiant ML350 server with redundant power supply
derived fields; and maintaining an audit log. For every and level 5 RAID array for fault tolerance, is centrally
field, descriptions of the variable and acceptable ranges located at JNMC. The computer terminals at the PHCs
have been specified and are displayed to the user are equipped with 56 kbps modems and connected to
during data entry. All the fields are connected by a land telephone lines and standard equipment for
unique seven digit identification incorporating check providing an uninterrupted power supply. The server
digit function assigned to every participant enrolled to has capability to receive the data from the PHC through
the study. This relational database enables capturing of an external RAS Unit. The data is then retransmitted
information at multiple points in time as the screening, through a Local Area Network to a local computer. The
enrolment and completion of the study protocol involve individual databases from the PHCs are merged and
data entry at periodic intervals. The user has the then the data is made available for editing. All the
freedom to access any of the data forms for a study forms are double keyed to verify the accuracy, the
participant, but the DMS alerts the user if a previously identifiers are stripped and the data is then transmitted
required record is not available in the database. To to RTI after encrypting it. The transmission is
verify the accuracy of the keyed data, a double keying accomplished using Blast Data Pump software. The
system with real-time comparison has been data is backed up on a daily basis on Digital Linear
incorporated. Tape (DLT) Drive using Veritas Backup System. The
entire network is secured behind a SOHO (Sonicwall)
Security of the DMS has been achieved through firewall with McAfee antivirus support.
user level authentication. The DMS is operated by
users with varied background and computing skills. Selected study-specific reports, programmed in
Further, the users have varying demands of the DMS visual basic, are run through a DMS system interface
and are available on Data Key computers with the

73
client (data entry) application. The reports generated on electronic data entry but one that has much wider
weekly basis include those for monitoring accrual to applications including data quality assurance, study
the study, tracking progress of recruited participants monitoring, analysis and reporting. The DDE
until completion of protocol, monitoring compliance in component conceived for this study could be further
submission of forms, tracking movement of developed by providing for reporting capabilities at
randomization envelopes, and of errors in keying the data entry terminals at the PHCs. Once reliable
data. Further, an automated e-mail system is alternatives to data transmission issues have been
incorporated for notifying Serious Adverse Events to worked out, the system can be adopted by the primary
the regulatory authorities. The weekly study status health care delivery system for collecting and
report, generated to monitor recruitment of study transmitting health sector data from remote rural sites
participants, enables the study coordinator to compare
to central locations. The databases developed will have
the expected and actual recruitment and provide
tremendous utility for administrative, planning and
feedback to the Field Research Officers (FRO) so they
research purposes.
can initiate measures to strengthen the activities of low
performing ANMs. A report of the Expected Date of
Challenges and Solutions
Delivery (EDD) for recruited study participants is
made available to the FROs and ANMs to ensure that The implementation of this project in an
participants potentially eligible for randomization are environment that was just beginning to be a part of the
not lost to the study. This has helped in improving the information technology revolution sweeping the
enrollment rate since in the initial months of the study country posed unique challenges. Upgrading the skills
it was noticed that the ANM could not be available for of the users who have been brought up on traditional
many of the deliveries due to her preoccupation with paper based record keeping systems was perhaps the
multiple responsibilities mandated by the ministry of least difficult barrier to overcome. Persuasion and
health. The EDD report has enabled the ANM to plan training enabled the research team to resolve this issue.
her work schedule in a more efficient way and be
Implementing the DDE plan at PHCs was more
available for a majority, if not all, of the deliveries
challenging. An erratic power supply situation and less
occurring at homes or when the women arrive at the
than desirable quality of the telephone connectivity
sub center for delivery.
compelled the team to switch over to centralized data
Similarly, a Follow-up Due Date (FDD) report is keying. Use of wireless technology to overcome
utilized to ensure adherence to the schedule of follow problems of connectivity through land telephone lines
up visits prescribed in the protocol for assessing the is an option that merits further exploration. The
secondary outcomes of the study. This has permitted an expanding wireless coverage could be harnessed to link
improvement in the compliance rate with more than remote villages with urban centers and establish a
98% of the follow up visits being conducted on the communication and data transmission channel.
prescribed days. DMS reports are used to track the
The DMS model employed for the present study
movement of randomization envelopes between the
may as well be adopted by the HRRC network as well
research unit and the ANMs as well as for checking
as other clinical research data collection efforts. The
their sequential assignment by the ANMs. A double
participating sites in multi-centre trials will have the
key verification report is generated to assess the
capability for timely error resolution and developing
frequency and type of errors in data. The errors in
good quality databases. Such electronic databases will
keying and inconsistencies in data that were ignored
provide insights for designing and implementing
during the manual verification of the forms and/or
research protocols that have the potential of making the
primary keying are compiled and are rectified after
greatest public health impact and at sites where it really
consulting the ANM and the FRO. If a pattern is
matters.
noticed in the type of errors committed, the same is
brought to the notice of the concerned study staff for
remediation. Thus the DMS reports are used for
Acknowledgements
This project was funded through the Global
managing the study, periodic retraining of the study
Network for Women’s and Children’s Health Research
staff and for detecting deviations from the protocol.
(NICHD grant # 1 U01 HD42372-01) and the Bill and
Many aspects of DMS development are continuing to Melinda Gates Foundation. Special thanks to: G.M.
evolve to cater to the demands of the study. As the Cressman, T.D. Hartwell, D.W. Jackson, C.E. Kelley,
study progresses to the data analysis phase, analytical Nora Kropp, and Vijaya Rao – from RTI International,
capabilities will be incorporated in the DMS. The goal RTP, NC,USA ; S.A. Edlavitch, from UMKC, Kansas
eventually is to develop a comprehensive DMS, in City, MO; and Susan Meikle, Nancy Moss and Linda
contrast to the basic data entry program used by the Wright – from NICHD, Bethesda, MD, USA
HRRC network, capable of not just providing for

74
References: for Computerized Databases in HMIS
http://www.hsph.harvard.edu/takemi/rp176.PDF
1. M. Abdellatif, & D.J. Reda A Paradox-based data
collection and management system for multi- 5. P. K. Baliar Singh, Health Information System:
center randomized clinical trials. Comput Methods India http://w3.whosea.org/ehp/his/india.htm
Programs Biomed. 73(2), 2004, 145-64. 6. Development of the Health System In: Country
2. M.N. Pollak. Computer-aided information Health Profile, India
management systems in clinical trials. A http://w3.whosea.org/cntryhealth/india/indsystem.
physician's perspective. Comput Programs htm
Biomed. 16(3), 1983, 243-51. 7. V.K. Tiwari, R. Dhingra, J.P. Shivdasani, R. Lal,
3. B.S. Kodkany, R.J. Derman, S.S. Goudar, S.E. R. Gandotra. Review Study on National Health
Geller, S.A. Edlavitch, V.A. Naik, A. Patel, M.B. Information System in two States of India
Bellad & S.S. Patted. Initiating a novel therapy in http://www.nihfw.org/material/Research/R197.doc
preventing postpartum hemorrhage in rural India: a
joint collaboration between the United States and
8. Computerized Systems Used in Clinical Trials.
U.S. Food and Drug Administration, Department
India. Int J Fertil Womens Med. 49(2), 2004, 91-6.
of Health and Human Services, April 1999.
4. B. Ranganayakulu, Evaluation of Health http://www.fda.gov/ora/compliance_ref/bimo/ffina
Management Information System in India Need lcct.htm

75
 DEVELOPING E- CARE PATHWAYS FOR THE 21ST CENTURY HEALTH
INFORMATION SYSTEMS
Sheenu Jhawar
Director,
ACE Vision Health Consultants Pvt. Ltd., India
sheenujhawar@yahoo.com, info@acehealthconsutants.com

Abstract critical examination of clinical practice should be an

integral part of patient care.[1] This care should ideally
Hospital notes present an age old problem: they are include the development and implementation of
often confusing, there is considerable duplication guidelines, together with continuous evaluation of
of information and large numbers of forms may clinical practice and outcomes.[2].
mean that key information could be lost. There is
an inability with existing systems to easily obtain Although many management practices and
quantitative and qualitative data. theorems identify the need for this practice, however
As a means to improve this situation, care few specify exactly how it can be done. For example,
pathways are a modern solution. clinical audit is recognized as a primary mechanism
through which research based guidelines can be
Care pathways are multidisciplinary process of introduced into routine practice and the patient benefits
patient focused care. They set out assessment, tasks subsequently measured [3]. However as is general
and interventions that should occur in a timely fashion, knowledge, audit on its own has hardly been included
to produce the best prescribed outcomes. in clinical practice.
In the same vein, development of this care With the dawn of modern technology,
pathway electronically, breaks new ground in contemporary data management and increased
providing clinicians and managers alike, in the ability globalization, certain practice models of the west have
to create assessment tools, like audit and risk found acceptance in the Indian system of management.
management- that reflect the most up-to date methods
and practices as well as tools that suit their own clinical One such system that integrates managerial and
practice. This functionality provides electronic decision clinical service amalgamation, and has its roots firmly
support system by in-built evidenced based medicine, embedded in the contemporary technology is the
reduces clinician time spent on documenting patient concept of ‘care pathways’.
examinations, reduces the need for patients to be asked Care pathways can be defined as problem-specific,
the same questions repeatedly by different clinicians or diagnosis specific patient treatment/ management
and provides greater access to up to the minute clinical plans that delineate key steps along an optimal timeline
data. to achieve a set of described intermediate and ultimate
This paper outlines the concept, and benefits of patient goals.[4]
care pathways, and their electronic transformation for Developed electronically, in modular structure, the
the very first time in India. pathway incorporates evidence based medicine, clinical
audit as well as risk management. This is so because
Key Words the pathway is based as per the evidenced based
E-care pathways, Evidenced based medicine, risk treatment guidelines for the specific clinical problem.
management, clinical audit. The pathway is divided into time intervals (usually
days) during which specific goals and expected
progress are defined, together with appropriate
1. Introduction investigations and treatment. Any variation from the
pathway gets pointed out due to its electronic nature
Hospital notes present an age old problem: they and this forms the ‘variance’ from norm. A later review
are often confusing, there is considerable duplication of of these variances forms the backbone of clinical audit.
information and large numbers of forms may mean that
key information could be lost. There is an inability Integrated care pathways provide guidelines and
with existing systems to easily obtain quantitative and patient management, which is unambiguous, practical,
qualitative data for the purpose of management and achievable and consistent. Effective e-care pathway
clinical information systems. The fact remains that the can provide substantial benefits to the patient and staff
alike.[5]

76
2. E- Care Pathways As suggested, ICPs essentially plot out the optimal
anticipated sequence and timing of interventions by
Electronic- care pathways are a latest advancement physicians, nurses, and other professionals for a
within quality improvement initiatives in the western, particular diagnosis or procedure. This anticipated
and some parts of the eastern world. sequence derives from: ‘evidence based practice’ or
By definition, an Integrated care pathway (ICP) is ‘clinical guidelines’ which can be described as: the
a multidisciplinary outline of anticipated care, placed ‘conscientious, explicit and judicious use of current
in an appropriate timeframe, to help a patient with best evidence in making decisions about the care of
specific condition or set of symptoms move individual patients.
progressively through a clinical experience to positive Pathways are derived from these guidelines, but
outcomes. are customised to the organisation in which they will
It is a document that charts the pathway of care for be used. They are multidisciplinary in nature,
a patient afflicted with a particular disease or set of combining both the clinical and non-clinical teams in
diseases thereby integrating the care to be received service delivery. Moreover they are timely sequences,
form all members of the clinical and non clinical team. thus they not only outline what should be done but
also-what will follow which in the continuation of care
In a way, it replaces the conventional medical and who will deliver what, starting from the admission
record. It is not only a form of documentation, but also process until discharge arrangements. Pathways also
serves as a tool to assist the quality of clinical derive inputs locally from the multidisciplinary teams
processes, by promoting evidence based medicine and in outlining the course of care. This involves varied
auditing its effectiveness, in a constantly changing issues like the socio-economic culture, patient
clinical environment. expectations, payment methods etc.
Care pathway techniques were first developed for The actual care provided, to the patient (as against
use in industry as a tool to identify and manage the rate the anticipated sequence) is documented electronically
limiting steps in production processes. [6] Also these by the patient bed-side. The documentation is process
were still kept as manual records, rather than the based – ‘tasks to be done’, and outcome based –
proposed electronic version of today. In industry, any ‘outcomes to be achieved’.
variation in production process is sub-optimal. Thus,
by defining the processes and timing of these While the ICP acts as the template of the care to be
processes, managers could target areas that were provided, it is not intended to compromise clinical
critical, measure variation, and try to make judgment. The members of the clinical team can
improvements. Once steps were taken to improve the deviate from the pathway if there is a valid reason for
process, there would be a re measurement. In time, doing so. The pathway asks each clinician to determine
variation would decrease, the time it took to complete whether each defined intervention is appropriate for a
the pathway would decrease, costs would decrease and given patient, thereby promoting clinical freedom
quality of production would improve. based on the needs of the individual.

When applied to health the technique of care Thus there might be several variations from the
pathways has obvious differences. First unlike in anticipated sequence of care. A subsequent analysis of
manufacturing, not all patient variances are negative. these variations provides invaluable input, not only for
This is also done to justify clinical freedom, as long as the prospective patient treatment -thus minimising
this freedom still falls within the limits of evidence clinical risks, but also if the pathway itself needs to be
based medicine. Individual patient factors may modified to incorporate the changes in the long run.
contribute to variation that cannot and should not be ICPs are dynamic documents and change is to be
controlled by the system. For example, if post- expected as new evidence, clinical guidelines and
operative ex-tubation occurred within a pre-specified treatment patterns emerge.
time period based on a pathway, there would be early Even though, steps to manage ‘common
ex-tubations with potential for harm. Also, unlike in variations’ are incorporated within the main ICP
manufacturing, in which products are standardized, document, yet this list cannot be exhaustive.
patients are different and may not fit within a pathway.
Second there exists concern that streamlining care may Sometimes it might be pertinent to deviate from
have a negative impact on patient outcomes. For one pathway of care and go to another in case of too
example, if a care pathway suggests a 2 day stay in the many variations owing to the patient suffering from a
cardiac care unit, a provider may alter care against his set of complications For example in the case of a
or her best judgment to stay within the plan. patient suffering form fracture neck of femur, as well
as a history of chronic heart problem there may be a
Despite these differences and the obvious need to consult the two different pathways for both
complexity of development of a care pathway, it has conditions. In this case, simple variance documentation
been embraced with open arm in the west. is inapt. Instead, by combining the different pathways,
and multidisciplinary teams, an ‘integrated care’

77
pathway is achieved. Having an electronic version of by having a care pathway, the cost per treatment is
the care pathway is a boon in such a scenario. By stated and therefore the budget can only cover a set
interlinking with appropriate modules of treatment, amount of treatments. It has
truly integrated care can be provided.
From a legal point, care pathways are useful because
Developing ICPs is a time consuming and a very with specified steps, doctors covering areas that are not
complex phenomenon. However the benefits of ICPs their speciality can follow it and there should be fewer
overshadow the difficulties. There are reports of a medical errors and negligence claims.
number of benefits associated with the introduction of
Currently these care pathways are in widespread use in
integrated care pathways into clinical practice. These
their manual form. However the agenda for converting
include a reduction in the length of stay in hospital,
them into electronic format has started as a huge step in
reduction of costs of patient care, improved patient
countries like the UK. The benefits for this initial
outcomes, improved quality of life, reduced
capital intensive act are multidimensional.
complications, increased patient satisfaction with the
service, improved communication between staff, and E care is much more than an IT implementation
reduction in the time health staff spent in carrying out project. It is a continuously evolving and gratifying
paperwork. implementation. The benefits that are recurrently
accrued can be life saving remedies for a patient.
Furthermore, Integrated care pathways-
Integrated care pathways demonstrate a sound
• Facilitate the amalgamation of local protocols
mechanism for quality improvement, and if developed,
with universal clinical guidelines in a easy to
can be an operational asset to the organisation.
follow format
• Facilitate shared practice with minimal 3. Conclusion
variations
• minimise delays, maximise quality of care Taking the initiative of electronic care pathways not
only requires capital, but also a fundamental change in
• provide co-ordinated and consistent care
culture and attitude. The culture should be one that is
- by providing explicit standards, help reduce innovative and challenging, incorporates technology
unnecessary variations in patient care, eg. delimiting and is sensitive to the needs of patients. It is universally
the diagnostic tools required. believed that the quality improvement agenda of the
21st century can result in demonstrable improvements
• provide a visual representation of the care in the quality of care given to patients by incorporating
plan . care pathways in their electronic format.
• improve communication between the clinical
team, as the holistic care given is documented
on a single record References:
• care pathway can be costed to achieve the
most cost effective way of treatment- in [1] McIntyre N. (1995) Evaluation in Clinical
conjunction with the best practice norms. Practice; problems, precedents and principles.
Journal of Evaluation in Clinical Practice 1, 5-
• clinical audit of the variances can be 13
undertaken to determine their impact on the
outcomes of care, and thus improvise the [2] Davies H.T.O. & Crombie I.K. (1995) Assessing
existent pathways the quality of care. British Medical Journal 311,
766
ICPs can also be invaluable in minimising clinical
risk and thereby litigation since they inherently [3] Harman D. & Martin G.(991) Medical Audit and
incorporate complete documentation. As this is done the Manager. Health Services Management
against the current best practice norms, and ‘accounted Centre, University of Birmingham.
for’ variations, the hospital and its staff base are well
[4] Strassner L. Critical Pathways: the next
protected too.
generation of outcomes tracking. Orthop Nurs.
Care pathways are also a proven tool for both cost 1997;16:56-61
savings (by identifying and eliminating practices which
[5] Grimshaw J. Eccles M. & Russell I. (1995)
are not evidence based and are streamlining the validity
Developing Clinically valid practice guidelines.
of diagnostic tests etc.) as well as improving the quality
Journal of Evaluation in Clinical Practice 1, 37-
of care. For the purpose of costing the hospital
48
services, each step of that pathway can be costed and
the whole treatment can then have an exact figure. [6] Buffa ES. Modern Production management 3rd
This is better than being handed a budget total and Ed. New York, NY: John Wiley & sons;1969.
being given a set number of treatments to aim for i.e.

78
 ONLINE SPEECH RECOGNITION FOR MEDICAL TRANSCRIPTION

K. Ramamohan Reddy, S. Prajit Nair, Y. Madhusudhana Reddy, and S. Dandapat

Department of Electronics and Communication Engineering
Indian Institute of Technology, Guwahati
India
Email: krr@iitg.ernet.in, ymreddy@iitg.ernet.in, and samaren@iitg.ernet.in

Abstract different templates e.g., name, date of birth, height,

weight, diagnosis, and treatment etc. Recognition of
This paper presents the implementation of an online words from different templates, as shown in Fig.1, can
speech recognition system for generating electronic be considered separately by implementing independent
health records (EHR) and its integration into Health codebooks for different templates. This system is
Care Information System (HCIS). The salient features named as Independent Component Codebook system.
of this system are online speech acquisition, real time The system asks for the medicine name first then
speech processing and speech to text conversion. In number of doses, age of patient etc. Age of patient is
speech recognition, the success rate depends on the assumed as a two digit number. It can be entered in two
number of words in the codebook. In this work, we steps, first the ten’s place digit then unit place digit.
have proposed an independent codebook approach The templates can be added to the system, in the order
for separate categories of entries in the health record that we like. The system also checks to see whether the
form e.g., medicine name, age, number of doses, etc. spoken word is in the codebook or not. If the word is
This method has improved the overall recognition rate not found in the codebook then that word is rejected
by 18%. A new cepstral weighting function named and the user is prompted to speak again.
mixed weighting function has been introduced for
improving speech recognition. A maximum of 12%
increase in recognition rate has been observed by using Template
mixed cepstral weighting. selection
control
Key Words
Electronic Health Record, Health Care Information
System, Medical Transcription, Speech Recognition, T1 C1
distance frame weighting.
word
Speech
analysis T2 C2
1. Introduction speech

Speech recognition by machines is an important

technology for the 21st century. Speech signals are
picked up by a microphone, digitized and processed by
Tn Cn
a computer. Speech recognition for medical
transcription generates medical reports [1]. In the
online speech recognition system, the clinician can T – Template, C – Codebook
interact with the speech recognition engine on a local
personal computer (PC). Once the report is completed
the clinician signs off on the report, and process is Fig. 1 Independent Component Codebook System
complete.
The whole process of medical transcription, as shown
The completed electronic medical reports are kept in Fig.2, is divided into two stages: training and
in the patient directory. A database of the medical recognition [3]. In the training stage, the speech
records of all the patients is maintained in website of samples are recorded using Goldwave software. The
the healthcare organisation. These reports can be sampling frequency is set at 8 kHz. Each sample is
accessed by the authorized parties (doctors and patients stored as 8 bit unsigned Microsoft’s “wav” format. In
etc,). They will be provided with a user ID and this work, ten medicine names are considered for
password to have a secure access to the data [2]. medicine codebook. They are ampicillin, avomin,
anacin, aspirin, crocin, cefotaxime, cosavil, dispirin,
2. Methodology paracetamol and quinine. A digit codebook of zero to
Our medical transcription system is designed in the nine is made. For each word twenty samples were
following manner. The health record card consists of taken from four speakers [4].

79
2.1 Online Implementation information bearing cepstral variabilities for reliable
discrimination of sound. Our proposed cepstral
The medical transcription system, shown in Fig.2, is weighting function named as mixed cepstral weighting
implemented in Microsoft’s Windows XP operating function, is defined by
system. Recognition is done online. The soundcard of
PC is programmed to sample the analog speech signal  i 
2

  4π 
at 8 kHz and 8-bit unsigned format. Two buffers, w(i ) =    exp  sin  − 1  + 0.5  , where 1 ≤ i ≤ p
 2  
which can hold 0.4 sec. of speech data, are queued to    ip  
the recording device. The system processes the data in
one buffer while the other buffer is being filled by the … (2)
device. The online system is implemented using where p is the cepstral order. The cepstral coefficients
multithreaded programming in Win32 API [5]. In this are weighted with this function as shown in Fig. 2.
system three threads are used: 1. acquires speech data
from the microphone, 2. calculates the linear predictive
cepstal coefficients [6] and finds start and end of the Speech
uttered word, 3. recognises and displays the word. The Signal
program has been run on Intel Pentium-4 processor Minimum distance
with a clock rate of 2.00 GHz and 256 MB of RAM. A Preemphasis from the test vector is
sample of the program output is shown below. & assigned as Spoken word
Boundary Detection
Utter Medicine Name: - Crocin
Utter Doses: - 2
Utter Age (One’s place): - 4
Utter Age (ten’s place): - 2 Linear Predictive Distance Frame
Do you want to continue if yes type Analysis Weighting
‘y’ else ‘n’: - n

Mixed Cepstral Distance

2.2 Proposed Cepstral Weighting
Weighting Measure
When comparing speech frames using cepstral
parameters, cn, a simple Euclidean distance measure is Recognition
Training
used. This measure reflects the uncorrelated nature of
cn. Since logarithmic power spectrum log s(w) of
speech is the Fourier transform of its cepstrum, CBSS
Parseval’s theorem can be used to equate Euclidean CBSS
cepstral distance to the RMS log-cepstral distance:

∞ π C1 C2 C3 C4
∑ ∫
2
d 2 cep = (rn − tn)2 = (2π ) −1 ln R ( w) − ln T ( w) dw
n =−∞ w =−π
… (1)

Where rn and tn are the cn for reference and test Cn: Codebook
utterances, respectively. Under certain regular CBSS: Codebook Select Signal
conditions [7], cn (except for energy c0) have zero
means and have variance that varies approximately as
the inverse of n2 (cn generally decays in amplitude with Fig. 2 Medical Transcription System
n and so the number of terms in the summation above
can be truncated, to about 10-30 [8]). Higher order
coefficients vary significantly with transmission factors 2.3 Distance Frame Weighting
(e.g. bandpass filtering in telephone channels) and
speaker characteristics which are irrelevant for This is one of the novel methods that has been tested
phonetic decisions, and because of the sensitivity of for online digit recognition. This method is applied to
lower order cepstral coefficients to overall spectral only digit recognition part and for those digits whose
slope and the sensitivity of the higher order cepstral recognition rate is comparably low with respect to
coefficients to noise, it has become a standard other digits. In this method weights are provided to
technique to weigh the cepstral coefficients by a those frames of digits, which showed poor recognition
tapered window so as to minimize these sensitivities with equal weighting. The weights and frames are
[3]. Therefore a cepstral weighting, or liftering chosen by experimentation. Since the duration of digit
procedure w[n] is designed to control the non- utterance normally lies within 0.6sec therefore every

80
utterance will have fifty-sixty frames (frame duration is recognition rate for digits “zero” to “nine” for cepstral
10msec). These fifty-six frames were divided into 7 order-12. For digit utterance “zero”, “one” and “four”
sets, each set having 8 consecutive frames. Distances recognition rate is improved slightly when mixed
were calculated for these seven sets from the reference
cepstral weighting is used. In case of digit utterance
code vectors. This results in a matrix of 10 x 7 in which
first row and first column signifies the cumulative “two”, mixed weighting performed well compared to
distance between first eight frames and so on. Then log weighted and equal cepstral weighting functions.
through analysis and experimentation by uttering the For digit “three” there is no improvement in
digit, which showed poor recognition, weights and the recognition rate.
frames to which weights should be applied are
determined. There is an improvement in recognition
rate after applying this technique. The recognition rate
100
of digit utterance “three” is improved greatly using this

Percent Success Rate

method. The weights for our database of order-12, is 80

given in Table 1. These weights, given in Table 1, 60

when tested for cepstral order-16, it has been found 40

that the recognition rate of other digits has been 20
decreased. To maintain the recognition rate different 0
weights are given for cepstral order-16, as shown in 0 1 2 3 4
Table 2. Digit Utterances

Equal weighing Log weighing

Table 1 Distance frame weighting for Cepstral order-12 Mixed weighing Mixed+Distance Frame weighing

set set set set set set set

Digit 1 2 3 4 5 6 7 Fig. 3 Recognition rate for cepstral order-12
0 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1
Fig. 4 shows recognition rate for digits utterances
2 1 1 1 1 1 1 1
“five” to “nine”. For digit utterance “five” there is no
3 0.9 0.9 0.9 0.9 1 1 1
substantial gain in recognition rate. In case of digit
4 1 1 1 1 1 1 1
utterance “nine” mixed cepstral weighting performed
5 1 1 1 1 1 1 1
well compared to log and equal weighting.
6 1 1 1 1 1 1 1
7 1 1 1 1 1 1 1
90
8 1 1 1 1 1 1 1
Percent Success Rate

80
70
9 1 1 1 1 1 1 1 60
50
40
30
Table 2 Distance frame weighting for Cepstral order-16 20
10
0
set set set set set set set 5 6 7 8 9
Digit 1 2 3 4 5 6 7 Digit Utterances

0 1 1 1 1 1 1 1 Equal weighing Log weighing

1 1 1 1 1 1 1 1 Mixed weighing Mixed+Distance Frame weighing

2 1 1 1 1 1 1 1
3 0.7 0.8 0.9 1 1 1 1 Fig. 4 Recognition rate for cepstral order-12
4 1 1 1 1 1 1 1
5 0.8 0.9 0.7 1 1 1 1 Fig. 5 and Fig. 6 show the recognition rate for cepstral
6 1 1 1 1 1 1 1 order-16. There is a substantial increase in recognition
7 1 1 1 1 1 1 1 rate for digit utterance “two” when mixed weighting is
8 0.9 1 0.8 1 1 1 1 used compared to equal weighting. There is an increase
9 1 1 1 1 1 1 1 of 50% in the recognition rate of digit utterance “three”
when distance frame weighting is used. There is slight
decrease in recognition rate of digit utterances “zero”,
3. Results and Discussion “one”, “two” and “four” when weights are applied to
digit utterance “three”. The same reduction in
The recognition rate for the utterances “zero” to “nine”
recognition rate is observed for the digits “five” to
and medicine names are plotted. Fig. 3 shows
“nine”.

81
 100
Fig. 7 and Fig. 8 show the recognition rate for digit and
medicine utterances with multiple codebook and
Percent Success Rate

80
independent component codebook. There is an increase
60
in recognition rate of almost all the utterances when
40
independent component codebook is used.
20

0 4. Conclusion
0 1 2 3 4
Digit Utterances
The concept of Independent Component Codebook for
Equal weighing Log weighing different categories of entries in the Health record card
Mixed weighing Mixed+Distance Frame weighing helped in an increase of 18% in recognition rate. A
maximum increase of 12% recognition rate is achieved
Fig. 5 Recognition rate for cepstral order-16 with mixed cepstral weighting. With distance frame
weighting method a maximum increase of 45%
recognition rate for digit utterance “three” is achieved.
90
80
This method has led to a slight reduction in recognition
Percent Success Rate

70 rate of other digits. The advantage of online medical

60
50 transcription system is immediacy; the clinician works
40 with the system in real time. Some of the many
30
20 applications of this technology include electronic
10
0
clinical charting and patient documentation.
5 6 7 8 9
Digit Utterances References:
Equal weighing Log weighing
Mixed weighing Mixed+Distance Frame weighing [1] S. Prajit Nair, Speech to Text based Medical
Transcription System, Master’s thesis, Dept.
ECE, Indian Institute of Technology Guwahati,
Fig. 6 Recognition rate for cepstral order-16 Guwahati, India, January 2002.
[2] Maisie Wang, Christopher Lau, Frederick A.
90
Masten, and Yongmin Kim, Personal Health
Information Management System and its
Percentage success rate

80
70
60
Applications in Referral Management, IEEE
50 Transactions on Information Technology in
40 Biomedicine, vol. 8, September 2004.
30
20 [3] L. R. Rabiner and R.W. Schafer, Fundamentals of
10
0
Speech Recognition (Prentice Hall: NJ, 1983).
0 1 2 3 4 5 6 7 8 9 [4] S. Itasi, Consideration on common speech
Digit Utterances
database, Recent Research Towards Advanced
Multiple code book Independent component codebook Man Machine Interface through Spoken
Language, pp. 503-514, Science University of
Tokyo, Tokyo: Elsevier Science, 1996.
Fig. 7 Recognition rate for digit utterances with
multiple and independent component codebook [5] Jim Beveridge and Robert Wiener,
Multithreading Applications in Win32 (Addison-
Wesley Developers Press, 1997).
Percentage success rate

100
90
[6] Atif Zafar, MD, J. Marc Overhage, MD and
80
70 Clement J. McDonald, MD, Continuous Speech
60
50 Recognition for Clinicians, Journal of The
40
30 American Medical Informatics Association, vol.
20
10
0
6, pp. 195-204, May/June 1999.
[7] K.K. Paliwal, On the performance of the
l

ne
e
in

n
in

ol
i
n

av

iri
ir i

ci
l li

m
ac

ni
ro

p
xi
sp

os
ci

frequency-weighted cepstral coefficients in vowel

vo

ta
An

ui
is
pi

ta
C
A

ce
D

Q
A
m

fo

ra
A

Ce

recognition, Speech Communication, vol. 1, pp.

Pa

Digit Utterances
151-154, May 1982.
Multiple code book Independent component codebook
[8] Douglas O’Shaughnessy, Speech
Communications: Human and Machine, IEEE
Fig. 8 Recognition rate for medicine utterances with Press and John Wiley and Sons Inc, 1999.
multiple and independent component codebook

82
 REFERRAL SYSTEM PLANNING USING GIS

Chandreyee Das and Jibanananda Roy

Hijli Inspiration
GC-1, Salt lake, Kolkata - 700106, India
{chandreyee, jroy}@inspiration-india.org

Abstract using GIS[5] to study the influence of locational

attributes on Health conditions and also to determine
Referral system, the basis for efficient public health
the nature of disease diffusion across geographic
care delivery mechanism has been designed and
regions. Some research has been attempted on creating
introduced in most of the districts of West Bengal with
Health database and using it as a support for Health
an aim to link up the multi-tier system of health centres
Facility Planning. The study by Mili Ghosh. Shantanu
and hospitals for management of patients according to
Lal and Dr. MS Nathawat[6] of BIT Meshra is on these
the degree of severity of illness. Complicated cases
lines and it provides a facility upgrading plan. So far no
beyond the scope of treatment in a particular facility is
attempt has however been made for Referral System
referred to a technically equipped higher tier hospital
Design in India as institutionalization of the referral
following a definite referral chain designed on the basis
mechanism is a relatively new management concept in
of distance and facility- availability criteria. The
Public Health Care Delivery System. A somewhat
outcome is mixed in terms of compliance and non-
similar study in identifying referral regions based on
compliance to the defined hierarchical pattern of
the service population and catchment area features has
referral due to a number of factors ranging from
been attempted by Dartmouth Atlas of Health Care in
awareness, spatial and supportive logistics and socio-
the United States.
economic conditions of the beneficiaries.
The referral mechanism has been introduced and
This is particularly relevant in areas with a difficult
made functional in most of the districts but are
terrain like the Sunderbans, Hilly Districts of
providing mixed results in terms of compliance and
Darjeeling or mixed terrain areas of the Duars. Thus
non-compliance to the defined hierarchical pattern of
planning for an adequate health system with an
referral. As identified, this has been due to a number of
efficient referral mechanism requires a combination of
factors ranging from inadequate knowledge of
facility and spatial analysis to derive an optimal service
providers and takers, lack of confidence of
delivery system. The present work uses GIS as a useful
beneficiaries on government services, inadequate
tool for decision support planning considering the
infrastructure and logistic facilities . But in regions like
incorporation of spatial and non-spatial data in a single
the Sunderbans barriers of problematic terrain vis-à-
reference frame.
vis distance, transport and conveyance facilities and
associated costs to reach the designated referral facility
Keywords: Referral System, Referral Chain, Spatial
emerge as the most important causal factor behind non-
Data, GIS.
compliance.
Introduction Health Referral System is regarded as a network of
healthcare facilities within a geographical area to
The use of GIS in health has been attempted by achieve effective and efficient use of the available
different agencies in India. Danida-assisted National resources.
Leprosy Eradication Programme [1] is one of the
foremost in introducing GIS in health in the country. The referral system planning involves two decisions
Apart from DANLEP, many development agencies and Deciding on the referral protocol
government institutions are exploring Health GIS in Deciding on the referral chain
India. Malaria Research Centre, New Delhi [2], Vector
control research center, Pondichherry, UNICEF, WHO Referral protocols contain administrative guidelines,
for leprosy, TB, malaria and Pulse Polio programmes, facility, equipment and service norms, referral chains,
HIV/AIDS programmes in TN, Orissa and MP are few at present, are designed based on primarily two criteria
recommended studies[4]. However all these studies
aim at developing Health / Disease Maps to aid in • Minimum distance
facility and preventive planning. An interesting work
has been carried out by LN Balaji of NATMO Calcutta

83
 • Availability of required service facility at the most often not following the referral chains, but
nearest point following the hierarchical following the principle of “facility available at the
pattern of health care delivery nearest government center/hospital”. This however
eases the referral procedure to some extent but
While lack of awareness, poverty etc. are basic to the compliance is still not to the desired extent as terrain
fact, inappropriate utilisation of the health care specific traversal is further dependent on seasonal
facilities which are vowed for providing minimal-cost conditions as well as availability of communication
preventive and curative services to the mass, also facilities. A significant percentage is also sent to
results in this unhealthy situation. A major component private service centres in order to avoid communication
of health sector planning basically involves the uncertainties
development of the health units by equipping them A study on a representative sample of
with adequate facilities and also making the system Gynecological and Obstetric Cases in three selected
beneficiary friendly by addressing the issues of care- blocks of South 24 Parganas Sunderbans revealed that
comfort –courtesy and required information support. in case of patients originating from hospitals, an
Government has launched different programmes to overwhelming majority still did not comply to the
make the health system efficient but these mostly relate referral chains and this is mainly due to lack of
to improvement in facilities, creating supportive communication logistics.
logistic and administrative back-up, awareness and
confidence generation among the providers and takers It is indeed difficult to address the problems of
etc. The introduction of the referral system is a major public health care utilisation in the Sunderbans in
step towards improved service delivery but the totality, as it is extremely dependent on geophysical
efficiency of the system still remains questionable and natural conditions. Study of the referral and
unless all factors deciding compliance to referral chains referred cases reveal the degree of compliance with and
are taken care of. Thus designing terrain and situation utilisation of the referral mechanism and resulting
specific referral chains, which has not been done, improvement in morbidity status. The realistic referral
would be attempted through this project. Integrating chains may also give indications for facility allocation
the referral chain design system into the holistic health among different units depending on the service load.
sector development planning is bound to deliver INSPIRATION had undertaken a programme on
effective results in terms of improved facility and Evaluation of Mobile Health Care Services undertaken
service utilisation. by NGO-s in the Sunderbans area of South 24
Parganas. While assessing the different components, it
Study Area revealed that any pre-defined norm based system of
health delivery like the referral system, which may be
Situational Overview of Sundarbans applicable to different districts of West Bengal may not
be applicable in the case of Sunderbans particularly
The Sundarbans in the eastern part of India and in because of its geographical features. Thus any
Southern West Bengal, with a population of more than programme planning for Sunderbans should necessarily
3.5 millions spreading over 19 blocks of both the incorporate the spatial features along with the other
districts of 24 Parganas, is one of the underdeveloped features to arrive at a practicable and appropriate
regions in the state with predominance of small and outcome. This led to the conception of the project in
marginal farmers. The 54 islands, interspersed with using GIS tools to design an optimal mechanism of
bodies of water, are covered with forests and swarms. service delivery through the multi-tier health care
Wide tidal rivers and estuaries and narrow tidal creeks system. Thus planning for an adequate health system
intersect them. Transport and communication networks with an efficient referral mechanism calls for the
are inadequate in this hostile geographical and design of an interactive and dynamic system for
topographical location. People have to travel in an optimal design of referral chains considering the spatial
assortment of improvised country boats, cycle- and non-spatial attribute. It requires a combination of
rickshaws and buses to reach their destination, which is facility analysis along with a spatial analysis to arrive
extremely time and cost inefficient. There are no major at an optimal service delivery system and GIS is the
hospitals in the region and travel time varies between 6 most useful tool for decision support planning
to 8 hours to reach Sub-divisional or District Hospitals considering the incorporation of spatial and non-spatial
from the core of Sunderbans. 11 RH-s, 8 BPHC-s and data in a single reference frame.
45 PHC-s are located in the region with 659 SCs. Most
of the BPHC-s, PHC-s and SC-s are situated in the
riverine area where as RH-s are located at the entry/exit Proposed Methodology
point of the mainland area of Sundarbans.
To design an optimal and effective referral system a set
The referral mechanism has been initiated in the of relevant data both spatial as well as non-spatial are
Sunderbans but not systemised. Tentative referral to be collected and analysed. GIS layers are to be
chains have been developed. Cases are being referred, created using one or more such data.

84
Data Collection • Study of non-compliance from survey data
• Demographic data analysis
While some data and maps are available from various • Disease profile with seasonal changes
organisations, survey must be conducted to generate • Transport Network with time of the day,
additional data. These tables are indicative but not fortnightly and seasonal factors
exhaustive set of required data. Depending on the • Terrain analysis
region, terrain additional data may be generated/
collected.
Non-spatial Conclusion
• Population features –district, block and village Serious attempts are being taken to tackle some of the
level problems by providing supportive logistic facilities but
• Ethno religious composition a planning process in-building the conditioning factors
• Literacy status is bound to strike the problem at the root and thus
• Occupational pattern create a situation for improved utilisation through
• Earning and per capita income appropriately designed referral chains. However, since
• Reproductive health the system would be designed considering only the
• Morbidity and disease profile governmental hospitals and health facilities, the
• Seasonal disease profile influence of alternative private health facilities would
• Service seeking behaviour possibly be ignored which may result in restricted
• Water and Sanitation status efficiency of the designed referral mechanism. Further,
• Hospital tier-specific manpower and facility status long-lasting impressions about certain health centres,
• Patient record negative or positive, may also lead to bypassing of a
• Referral tables few, although they feature as distinctive nodes in the
referral chain.
Spatial

• Administrative Map
References
• Terrain map [1] http://www.danlep.org/gismis.html
• Land use map
[2] Srivastava, Aruna and B.N. Nagpal. Mapping
• Transport network map with seasonal implications
malaria. GIS Dev., 4(6): 28-31.
• Maps indicating social infrastructure
• Map indicating health unit specific catchment area [3] Dhiman, R.C., R. Sudarshana, V.P. Sharma, M.K.
• Maps indicating referral chains Das and S.K. Bhan. Targetting mosquitogenic
• Social maps indicating region specific disease conditions with emphasis on Anopheles sundaicus
prevalence, seasonal variations. on Car Nicobar using remote sensing and
Geographic Information System techniques: A
Analysis pilot study. Asian-Pacific Remote Sensing and GIS
J., 13: 23-28.
The following analyses needed to be carried out using [4] http://www.gisdevelopment.net/application/health/
the collected data. From these, a set of attributes may overview/index.htm
be chosen with proper weightings for designing the
optimum and effective referral chain. [5] http://www.gisdevelopment.net/application/health/
overview/healtho0003.htm
• Generation of base map of the region along
with administrative boundary [6] Spatial Decision Support System Using GIS based
• Mapping of the Health facilities with attribute Infrastructure: Planning in Health education for
table indicating capacity and capability Ranchi District, Mili Ghosh,Shantanu Lal,Dr. M.
• Study of the current referral tables S. Nathawat, Map India 2002.

85
 KNOWLEDGE MANAGEMENT IN HEALTH SYSTEMS: THE EMERGING ROLE OF
CHIEF INFORMATION OFFICER (HEALTH) FROM DEPARTMENTAL LEVEL TO
HEALTH SYSTEM

Dr. Sanjay Bedi

M.B.B.S. DCh, M.D.Pathology
Associate Professor & Head,
Department of Pathology,
Sri Guru Ram Das Institute of Medical Sciences, Amritsar
Address 366 Green Avenue Amritsar
Tel No. 91-183-2503785 Email drsanjaybedi@gmail.com

Abstract Key Words

In this paper six Levels of knowledge management Knowledge Management , health systems, Chief
have been identified starting from departmental head Information Officer (Health)
level, medical superintendents, civil surgeon, state
headquarters, director general health services, and Chief Information officer will have a difficult job
finally World Health Organization. of running a function that uses a lot of resources but
offers little measurable evidence of its value .The Head
Knowledge management in health system is the of Department, Medical Superintendent Civil Surgeon
art of creating health system values, by leveraging the must assume more responsibility for planning, building
intangible assets of knowledge, which will become the and running information systems that effect their
next big competitive differentiator in improvement of operations.
health systems in the Information Century,
To respond to changes in the technological
Science based hospital are leading the new wave environment and availability of information, Chief
and survey results show that they have gained by Information officer must build relation ships with
knowledge Management initiatives which nurtures Information Technology industry as well as the
radical innovation. Advance Planning, change Medical community and assume new and more
Management, hospital culture & well-balanced strategic roles .The strategic role of CIO Health
approaches are essential part of knowledge becomes even more complex requiring an expansion of
management. It offers creation of values from better organizational and structural possibilities for filling
management of intangible assets. Benefits will surface that role. Also a formal medical information system
in measures such as cost reduction, cycle time planning needs to be adopted at Hospital, State and
reduction, better resource returns, higher patient National Levels. With increasing number of people
satisfaction index and increased doctors and reporting to the CIO at various levels of hierarchy
paramedical staff education levels. respectively.
Resignation and migration of doctors is forcing Thus a CIO Health is needed with right
periods of massive downsizing which means medical qualifications having an amalgamation of Medical and
superintendents are continuously losing the vast Information Technology backgrounds.
knowledge bases that they had accumulated over the
years, either due to loss of doctors of doctors or Strategic Management is an important step in
because of changing times. How many times have ensuring the long term viability of organizations .It
doctors in doctors wished, there was someone to tell involves the reading of signs portents of the future and
them what to do in a particular situation? Do not all of interpreting them in order to chose an appropriate
us, in the doctors world, wish we could learn from the direction for the future development of the
rich past experiences of the hospital and its men, in the organization. Information technology gives strategic
least time, so that we could obviate passing through the advantage and improves the productivity of the
same muck and concentrate on doing our jobs better? organizations. A Medical Informatics department is
And with the information load on individuals suddenly needed at all Medical Institutions.
becoming too big; there is no understanding of how to
The chief Information Officer Health can be a
extract value of it. Effective usage of available
powerful change agent by suggesting new methods of
knowledge resources is the key rather than making a
of Information Processing and coordinating both
duplicate copy of whatever knowledge resources are
development of technology and planned changes in the
available.
organization.

86
 interdepartmental communication electronically. He
will ensure proper functioning of the Hospital
Ironically while the power of IT in the information
Information Systems. He will be needing a Computer
age is readily apparent in the growing use of such
Network in the hospital and a programmer in his staff
terms such as cyberspace, paperless society the power
along with data entry operators as per need. Also ICD
of Information officers in Health has not been
Coding of the Case sheets, MIS reports, census data,
highlighted. Yet these are traditional entities. The Chief
formats become additional duties. He may be a initially
Information Officer Health has to move the
drawn from within the Hospital Doctors and later the
organization from an era of when controlled
position may be taken by a full time Health Informatics
information resources to an era where users are more
specialist as and when available. How ever since one of
involved in designing the Information systems.
his duties will be close liaison with the staff of the
hospital.
The need for CIO Health has been driven by two
organizational needs. First, the accountability is
increased making a single executive responsible for
organization’s information processing needs. Second,
3. District Level
creation of CIO position facilitates the closing of the At the district level he/she will maintain the
gap between and IT strategies which has long been the electronic mailing list of all the doctors and
cited as primary business concern. disseminate information about the diseases prevalent in
the area electronically. He will collect information
The Chief information officer in Indian System is from the health facilities in that area and provide the
needed at epidemic alerts on a real time basis using a
1) Departmental Level Geographical Information System if possible. He will
maintain the disease registries in the area and maintain
2) Medical Superintendent level
close liaison with the state and National Level
3) Civil Surgeon Level counterparts so as to be a part of an effective state and
4) State Health Officer Level National Level Health Information system. One of his
5) Director General of Health Level important duties will be strategic Information system
planning in Health Departments. He will ensure
6) World Health Organisation Level
computer literacy levels in the staff of the District and
will play the role of Chief architect, change leader,
product developer ,technology provocateur , coach and
1. Department Level
chief operating strategist in collaboration with the
district Health authorities . He will be needing a wide
At departmental level he may be chosen from one
array of staff data entry operators, epidemiologist,
person in the staff who may be given specialized
social investigator, Network operators, Programmers in
training in Information technology and later he may
his staff.
take up additional duty with some incentive to work as
CIO Health for the department. He will ensure basic
training level for rest of the staff including doctors, 4. State Level
Nurses and paramedical staff. He will ensure that data
of the department is converted to electronic format and At the state level he will get the data from various
also made available to users in required format. He may districts in the state electronically and collate it and
add some data to disease registries as and when send real time information preferably on the same day
required. He will make available the necessary to the districts back so as take effective preventive
information from the net as and when required. He will health measures in time. He will be involved in setting
be needing a computer and a data entry operator as his a state wide Virtual Private Network for Health
staff. Institutions/Hospitals/Clinics in collaboration with the
Information Technology community and will decide
2. Medical Superintendent Level the most cost effective solutions as per the needs of the
Health Community.
At the Medical Superintendent Level he will ensure
that the data coming from various departments is
5. National Level
converted into useful form so as to make meaning
decisions by the Medical Superintendent. Also he will At the National Level he/she will get the data from
maintain close contact with the IT industry and various states gets merged into National Data
Medical Informatics world and make necessary repositories which gets processed quickly and the
recommendations. He will make sure that economically results are translated into policy guidelines and
viable decisions are taken for the welfare of the corrective actions based on hard data. He will be
hospital. He may be the moderator of a mailing list of assisted by a team of Data analysts drawn from various
all the departments in the Hospital so as to facilitate sections of health system including statisticians,

87
epidemiologists, as well as Information Technology
specialists.
Knowledge management techniques combined with
6. World Health Organisation level
cultural and process changes, with enabling technology
to realise the bottom line result can drastically improve
It should be a coordinating body where
patient care and satisfaction levels. The primary
knowledge generated at various parts of the globe
tools applied in the practice of knowledge management
be available for dissemination and global strategies
are
formulated for disease control thus improving the
system continuously. - Hospital work culture
- Process Engineering
At all levels it is necessary to practice a knowledge - Technology
sharing culture that will marshal resources to determine This three work in harmony to streamline and
what people need to know, organize and promote a enhance the capture and flow of hospital’s information
knowledge sharing culture. Knowledge teams may be & delivers it to the teams, engaged in accomplishing
needed for this purpose. specific tasks.
For this purpose efficient knowledge management Conclusion
practices are needed.
A Chief Information officer Health is needed at levels
Knowledge Management in the context of
in the Health System to implement Knowledge
Hospitals is “ a patient care activity with two Management initiatives in the System at strategic
primary aspects: Treating the knowledge component of operational and management levels. This is the need of
patient care activities as an explicit concern of patient
the day.
care reflected in Strategy, policy and practice at all
levels of the hospital; and making a direct connect
between hospital’s intellectual assets and improved References
patient care Strategic management of IS/IT functions: the role of
the CIO in Norwegian organisations
“Knowledge management is a discipline that Petter Gottschalk* International Journal of Information
promotes a collaborative and integrated approach to Management 19 (1999) 389}399
the creation, capture, hospital, access and use of an
health system’s information assets. This includes Bedi Sanjay et al Stages opf Information
database, patient records and, most importantly, the Technology Usage In a State Health Systems
un-captured, tacit expertise and experience of Corporation.
individual workers”. Pubished in Information Technology For
Innovation edited by Dr. S.P.Mudur Publisher Tata
Knowledge management is first and foremost a Mcgraw Hill
management discipline that treats intellectual capital as Bedi, Sanjay “District Health Information Systems”
a managed asset. Many innovative hospitals are Published in CSI Communication, April, 1997.
examining how they can better their intellectual capital
through the emerging field of knowledge management. Bedi, Sanjay “Computers and Rural Health Systems”
Published in CSI Communications, Sept. 1996.

88
 COMPUTERIZED PHYSICIAN ORDER ENTRY: WAY FORWARD

Dr Pankaj Gupta
BDS, PGDCA, Dip. Bioinfo.
Ex-Project Manager,
FCG Software Services (India) Pvt Ltd.
HM Towers, Brigade Road, Bangalore, India
dr_pankajgupta@yahoo.com

Abstract It has been found that lot of deaths happen due to

human error at the physician’s order entry stage itself.
CPOE is seen by most renowned organizations CPOE is aimed at alerting the physician about
worldwide as the answer to reducing medical errors by potentially dangerous/erroneous orders before the
bringing in a major breakthrough in the area of patient orders are really executed, thereby facilitating a
safety. CPOE standards came into being primarily as a solution for this long-standing issue in medical set up.
result of the Institute of Medicine’s 1999 report on The testimony to the benefits of CPOE is that more
medical errors and the subsequent coming together of than 10% of U.S. hospitals now have CPOE.
Fortune 500 employers to form the ‘‘The Leapfrog
Group’’. The Leapfrog Group preferentially directs the The CPOE concept has existed for a long time and
employees’ healthcare to those institutions that install now The Leapfrog Group is laying down the standards
clinical systems compliant to CPOE standards. to give it a proper direction. Large Healthcare-IT
Adoption of CPOE has resulted in benefits to some; vendors are already putting efforts to make their
however there is considerable skepticism in the market clinical systems CPOE compliant. However the
against CPOE. This paper discusses the potential obstacle is that many clinical systems from different
benefits of CPOE in a clinical system. At the same vendors still do not talk to each other. The Computer
time, it shows that the computer systems in hospital systems in hospital environment have to evolve a lot
environment need to evolve to enable CPOE deliver on before the true benefits of CPOE can be realized. Large
its promise. Healthcare-IT vendors have to play a responsible role
in integrating the Health delivery industry, and thereby
Key Words move towards a greater patient safety.

CPOE, The Leapfrog Group, Institute of Medicine IOM study and CPOE
(IOM)
The starting point of thought provoking
discussions on patient safety has been the Institute of
Introduction Medicine’s (IOM’s) report ‘To Err is Human, Building
Safer Health Systems’1, 2. The IOM pointed out that
44,000 - 98,000 people die in the USA hospitals 44,000 to 98,000 deaths happen every year in the USA
each year as a result of medical errors that could have due to preventable medical errors. Errors are costly in
been prevented.1 These facts were cited in the Institute terms of psychological discomfort, diminished
of Medicine’s report of 1999, which was based on satisfaction and loss of trust in the healthcare system by
estimates from two major studies in the USA. Even patients, and loss of morale and frustration in health
using the lower estimate, preventable medical errors in professionals. More US citizens die in a given year as
hospitals exceed attributable deaths to such feared a result of medical errors than from motor vehicle
threats as motor-vehicle wrecks, breast cancer, and accidents, breast cancer, or AIDS. Beyond their cost in
AIDS. As a result of the IOM report, Fortune 500 human lives, preventable medical errors have been
employers in the USA found that though a lot of estimated to cost (including the expenses of additional
healthcare standards were being put in place, but care necessitated by errors, lost income and household
patient safety still remained an area where focus was productivity, and disability) between $17-29 billion per
needed to reduce errors and thereby control the year in hospitals in USA. 1
spiralling medical costs and premiums. This led to the
formation of The Leapfrog Group with the aim of Shortly after this report was made public, the
improving patient safety. Developing Computerized Business Round Table founded The Leapfrog Group, a
Physician Order Entry (CPOE) standards is one of the national association of Fortune 500 chief executive
many initiatives from The Leapfrog Group to improve officers (CEOs). The Leapfrog Group focuses on 3
patient safety. main areas of patient safety — CPOE, ICU Staffing

89
and Evidence based hospital referrals. The Leapfrog individual recklessness or the actions of a particular
Group was created to ‘‘help save lives and reduce group – this is not a “bad apple” problem. More
preventable medical mistakes by mobilizing employer commonly, errors are caused by faulty systems,
purchasing power to initiate breakthrough processes, and conditions that led people to make
improvements in the safety of health care and by giving mistakes or fail to prevent them. 1
consumers information to make more informed
hospital choices.’’ 2 Errors happening due to illegibility of Physician
notes in the clinical setting have been a long-standing
The intent of The Leapfrog Group is to and ever present complaint from the paramedical staff
preferentially direct their corporate members’ health (pharmacists, nurses & other ancillary staff). Electronic
care to those organizations that adhere to patient safety Medical Records (EMR) solves the illegible physician
standards specified in the guideline documents based notes issue to a large extent. However, EMR leaves
primarily on the IOM report2. scope for human error at the stage of physician order
entry itself. This is where CPOE standards fill the gap
The IOM study has had its fair share of criticism to prevent erroneous or dangerous orders to get past the
especially from the physicians. However it has been the Physician stage.
principal argument for the push for adoption of patient
safety standards like CPOE. Mistakes can be best prevented by designing the
healthcare systems in a way that makes it more difficult
What is CPOE? for people to do something wrong and easier for them
to do it right. Medication process provides an example
Computer physician order entry (CPOE) systems
where implementing better systems will yield better
are electronic prescribing systems that intercept errors
human performance and reduce errors.
when they most commonly occur — at the time
medications are ordered. With CPOE, physicians enter
orders into a computer rather than on paper; these
orders are then integrated with patient information,
CPOE Evolution
including laboratory and prescription data. The order
sets are automatically checked for inappropriate/ Decentralized and fragmented nature of healthcare
dangerous orders before they are executed. Over a delivery system has been the oft-cited problem that has
period of time many clinical decision support systems contributed to medical errors. When patients see
have come out with CPOE concepts. However, The multiple providers in different settings, none of whom
Leapfrog Group has now laid down a set of standards have access to complete information, it becomes easier
for the computer programs for alerting health care for things to go wrong.1 Fully integrated Hospital IT
providers to potentially harmful therapeutic decisions infrastructure, EMR, and computer based data capture
before orders are processed. and data storage are pre-requisites for institutionalizing
CPOE standards.
The Leapfrog Group includes the following
language in their CPOE Fact sheet. “In order to fully Computer systems have evolved in the hospital
meet Leapfrog’s CPOE Standard, hospitals must: environment over a period of time. The early computer
systems in the hospitals were essentially stand-alone
1. Assure that physicians enter at least 75% of islands of patient data that could not communicate with
medication orders via a computer system that other systems in the same department, let alone the
includes prescribing-error prevention software; other systems in the hospital.
2. Demonstrate that their in-patient CPOE system can
Nowadays, most hospitals are investing time and
alert physicians of at least 50% of common,
effort for integrating various stand-alone systems
serious prescribing errors, using a testing protocol
across departments to reduce errors during double entry
now under development by First Consulting Group
of patient data and enable physician order execution in
and the Institute for Safe Medication Practices;
near real-time.
3. Require that physicians electronically document a
reason for overriding an interception prior to doing Fig. 1 below shows that the computer systems in
so.’’3 hospitals are evolving from stand-alone data collection
mode to an integrated healthcare enterprise (IHE).
Why CPOE? Once the systems are integrated the stage is set to
institutionalize CRM (consumer relationship
One of the IOM report’s main conclusions is that management), evidence based medicine and the
the majority of medical errors do not result from topmost layer of patient safety-- CPOE.

90
 from 10.7 to 4.86 events per 1000 patient days.
Preventable ADEs declined 17% from 4.69 to 3.88 per
1000 patient days, while non-intercepted potential
ADEs declined 84% from 5.99 to 0.98 per 1000 patient
days. The prevention of errors was attributed to the
CPOE system’s structured orders and medication
checks.3,4

CPOE has paid other dividends. Length of stay at

Wishard Memorial Hospital in Indianapolis fell by 0.9
days, and hospital charges fell by 13% after
implementation of CPOE. A recent study at Ohio State
University also identified substantial reductions in
pharmacy, radiology, and laboratory turn-around times,
and there was a reduction in length of stay in one of the
Fig 1: Evolution of CPOE two hospitals studied.3

Today all of the large clinical systems vendors are Some of the benefits of CPOE include:
making sincere efforts to make their own products
CPOE compliant. However the reality is that the • Prompts that warn against the possibility of
hospital can have islands of computer systems drug interaction, allergy, overdose etc.
installed/ built at various periods of time by different • Accurate, current information that helps
product vendors. Large Healthcare-IT vendors have to physicians prescribe the new drugs as they are
come out of business silos and move to open standards introduced into the market
to help Health delivery industry for integrating the • Drug-specific information that eliminates
stand-alone systems and tiding over the chasm. Only confusion among drug names that sound alike
then, true CPOE can be implemented enterprise wide. • Improved communication between physicians
and pharmacists
• Reduced healthcare costs and hospital stay
CPOE Alerts due to improved efficiencies.

CPOE standards recommend that alerts be given to CPOE Evaluation application by FCG
the Physician for basic to expert level warnings. The
range of alerts varies from the drug allergy and drug First Consulting Group (FCG) had developed the
overdose (basic alert) to contraindication based on methodology to help hospitals evaluate whether their
individual’s laboratory studies (expert alert). CPOE systems meet the Leapfrog CPOE standards.
Now FCG has also developed the CPOE evaluation
Basic level alerts are simple alerts for allergy to application for The Leapfrog Group. This web-based
penicillin and overdose of antihistamine. Whereas application is expected to be used by 5000 hospitals
alerts for unusual drop in blood clotting time and across US to test their CPOE compliance.
prothrombin laboratory values in patient’s charts due to
increasing dose of anti-coagulant like warfarin is an Flip Side of CPOE
expert level alert.
Physicians and medical staff need real time access
Alerts need a huge enterprise-wide knowledge to data that is relevant to the case at hand. They need to
base to operate in the backend. Some of the medical be able to record a maximum amount of information in
knowledge is readily available whereas some of it is a minimum amount of time and in such a way that it is
still state, region and hospital specific. Expert/ most useful to other health care professionals involved
Advance level alerts e.g. drug-lab-document alert will in the handling of this patient. It is totally unacceptable
need data from across different hospital systems. These if the alerts do not appear real time and increases the
systems need to be integrated to yield full benefits of physician’s time per patient.
CPOE.
Decision support systems also suffer from the
Potential Benefits of CPOE problem of an overdose of reminders, alerts, or
warning messages. This delay can be dangerous in
CPOE systems can be remarkably effective in emergency situations. CPOE compliant systems are
reducing the rate of serious medication errors. A study infamous for “…causing cognitive overload by
led by David Bates MD, Chief of General Medicine at overemphasizing structured and ‘Complete’
Boston’s Brigham and Women’s Hospital, information entry” 5.
demonstrated that CPOE reduced error rates by 55% —

91
 There is a rather large grey zone of informal To completely replace legacy clinical systems with
management, which can be entirely rational given the a single-vendor, monolithic solution would be
everyday organization and exigencies of health care expensive and cumbersome. As an alternative, taking
work. In emergency and some other special situations, the application integration approach to meet CPOE
orders may be entered after the order execution. For requirements will typically cost less in terms of time
example, while transferring a patient between the and material.
emergency department and ward, orders could not be
transferred or new orders could not be entered in the Large Healthcare-IT vendors should focus on
system because the patient was not yet ‘‘in the larger benefits by integrating the health delivery
system”. industry rather than competing with each other for the
same piece of the pie.
In the case of urgent medication orders, nurses can
give a medication before the physician formally Acknowledgements
activates the order. During nightly routine medication
administration, nurses can initiate distribution without I wish to thank FCG for giving me the opportunity
waking up the junior doctor who is formally to design and develop the CPOE evaluation
responsible for signing the order. Within this same grey application. I am thankful to my wife Dr Savita for
zone, there could lay many practices that would putting on the editor’s cap and editing the content of
contribute to unsafe medication routines such as this paper. Dan Coate and Vaishali Bhinde gave me
doctors actively discouraging nurses to call them for valuable inputs to improve the presentation of the
medication requests or nurses taking too many liberties content in the paper. Like always, Vishal Kirplani
with dosing.5 All of these practices exist within the helped me in putting the images and formatting as per
current paper medication systems, but many CPOE specifications.
systems do not leave room for such practices.
References:
Inexperienced computer users can face issues like
a slip of the mouse on a data entry form leading to an
order for the right medication for the wrong patient. [1] LT Kohn, JM Corrigan, MS Donaldson, To err is
Such errors due to inexperience lead to arguments that human: building a safer health system (Committee
pen and paper are simpler and better. However expert on Quality of Health Care in America, Institute of
level alerts in CPOE systems are expected to take care Medicine Washington DC, USA: National
of such issues to some extent. Academy Press, 1999).
[2] Robert G. Berger, JP Kichak, Computerized
Conclusion Physician order Entry: Helpful or Harmful?,
Journal of American Medical Informatics
The Institute of Medicine’s report has had its Association, 11, 2004, 100-103.
desired effect. Formation of The Leapfrog Group and
coming out with the CPOE standards is a right step [3] Fact Sheet: Computerized Physician Order Entry
towards patient safety. (CPOE), The Leapfrog Group for Patient Safety,
Available at:
CPOE systems can reduce unnecessary repetitive http://www.leapfroggroup.org/media/file/Leapfrog-
orders and also significantly cut down the delays Computer_Physician_Order_Entry_Fact_Sheet.pd
between writing and completing orders. They can also f. Accessed at 14 Jan 05.
cut staff costs directly by reducing the time spent by
nursing, pharmacy, and other ancillary services on [4] DW Bates, LL Leape, DJ Cullen, N Laird, et al.
callbacks to clarify orders and by eliminating the Effect of computerized physician order entry and a
personnel time of transcribing orders. So, health care team intervention on prevention of serious
institutions have much to gain in efficiency and cost medication errors, Journal of American Medical
savings from CPOE systems. Association, 280, 1998, 1311-1316.
[5] JS Ash, M Berger, E Coiera. Some Unintended
In the late 1980s and 1990s, some people criticized Consequences of Information Technology in
that no one else used or ever would use CPOE. Healthcare: The Nature of Patient Care
Whereas more than 13% of U.S. hospitals have CPOE Information Systems-related Errors, Journal of
today.6 American Medical Informatics Association, 11,
2004, 104-112.
To derive the true benefits of CPOE the challenge
is to create user-friendly, seamless systems that [6] CJ McDonald, JM Overhage, BW Mamlin, PD
integrate all critical disparate systems throughout the Dexter, WM Tierney. Physicians, Information
enterprise- including patient records, order entry, Technology and Healthcare Systems: A Journey,
pharmacy, radiology and Lab. Not a Destination, Journal of American Medical
Informatics Association, 11, 2004, 121-124.

92
Telemedicine Systems & Hospital Information Systems
 SECURED NETWORK SOLUTION FOR HEALTH CARE INDUSTRY
WITH HIPAA COMPLIANCE
Surid Kumar Das, A.Vadivel
Computer and Informatics Center
Indian Institute of Technology, Kharagpur
Email: - {surid,vadi}cc.iitkgp.ernet.in.

Abstract access of the health care information. In section 3, we

discuss HIPPA proposed security standard. In section
Health care industry is a multi disciplinary in 4, we propose a solution for implementing a secured
nature with different entities such as doctors of network for health care industry and in the final section
different specialization, nurses and pharmacists at we conclude the paper.
different location. Patient’s information needs to be
shared with privacy and confidentiality by these 2. Need of Secured Access of Health Information
entities for faster diagnosis using a secured network
infrastructure with proper encryption mechanism using The different entities of the health care industry
foolproof technology. Compromises to data integrity, such as doctors, nurses, hospital personnel, insurance
availability, and confidentiality often take days of agents [6], and others can provide more efficient care
network downtime to repair, can elicit expensive civil to patients by accessing health information and
lawsuits from patients, and can greatly decrease ensuring that the data is correct and perfect.
employee efficiency and patents satisfaction. We Pharmacists can easily avoid prescribing conflicting
propose a secured network solution using HIPPA medications and doctors can make faster diagnoses or
standard for the health care industry. prescribe minor treatment from a remote location. But
medical records often contain information that
Key Words individuals would like to keep secret from others and it
is simply patient preference. In the Internet perspective,
HIPPA, secured network, Firewall, IDS, Virtual
an unsecured data server that stores patient’s
Private Network
information data and data transmission between two
health care industries without proper encryption
1. Introduction attracts the attention of hackers who can modify the
The health care industry is a documentation- information. So it is imperative that a well secured and
intensive in nature and has faced significant challenges proper encryption mechanism is required to safeguard
in migrating to the near paperless environments. The the patient’s data.
volume of patient sensitive data has grown
exponentially over the year. The evolution of computer 3. HIPAA Proposed Security Standard
network has enabled to provide enhanced services,
greater access to information and higher levels of Health Insurance Portability and Accountability Act
availability resulting in increased customer satisfaction. (HIPAA) proposed national standards regarding patient
The health care industry, dealing with such sensitive data privacy and security. HIPPA says that the
patients data has raised an important questions organizations that maintain or transmit health
regarding the privacy [1,3,4,10], confidentiality and information electronically must establish and maintain
security [3,9] risks of patient sensitive data. The cost reasonable and appropriate administrative, technical
reduction on the products of computer and and physical safeguards to ensure the integrity,
communication devices, non-availability of specialists confidentiality and availability of the information.
at their location and reusability of the existing Recognizing that no single standard exists that
resources forced health care industry to automate their integrates all the components of security such as
work. HIPPA [5] proposed standards regarding patient administrative procedures, physical safeguards and
data, privacy and security with a goal to reduce the technical security services, the legislators who
administrative overhead. We will identify and discuss developed HIPAA proposed a new security standard to
some of the issues associated with the HIPAA define the security requirements. HIPAA states that the
requirements and how technologies available today that final health care security standard should be comprised
can be part of a comprehensive network design and with the following three conceptual guidelines as a
security policy to ensure safe and secure basis [12]
communications between the various entities and Scalable— All sizes of health care entities should be
components in the Healthcare Industry. able to comply with the standard. Comprehensive—
The rest of the paper is organized as follows. In the The required security solutions should act as a unified
next section of the paper we discuss need of the secure system, not a series of piecemeal products that do not

89
communicate with each other. Technology-neutral—It of network resources. These security solutions address
should not reference or advocate specific security all compliance [7] outlined by the HIPAA requirements
technology, it should not dictating specific system in section 3. The blueprint describing how to design a
architectures and technologies and have the flexibility practical security infrastructure to protect all areas of a
to choose the best solutions to fit their particular network, enable organizations to create scalable,
environments. manageable, and reliable security infrastructures that
should meet the most stringent security regulations that
HIPAA recommends several requirements that come down the road. We are providing a
should be included in the final health care security comprehensive suite of technical security solutions for
standard. Based on the proposed rule, there are five building secure network infrastructures for health care
components to the security regulations as stated below. industry with HIPPA compliance. Our solutions
Administrative procedures-This is to manage the addresses secure authentication, perimeter protection,
development and execution of security measures of an intrusion detection, encryption, and network
industry. Documented formal practices should be made monitoring and management.
available to manage and execution of security measures
to protect patents data. The procedures would include 4.2 Access Control Server based security
items such as formal termination procedures, security mechanism:
incident procedures, and security training. Physical
safeguards -It is related to the protection of physical In an access control server based security mechanism
computer and communication systems, buildings, as shown in figure 1, it validates users’ identities, and
equipment from fire, environmental hazards and determines which information the users can access
physical intrusion. It also covers the use of locks, keys, based on stored user profiles. The Secure Access
and administrative measures that control access to Control Server (ACS) is a high-performance, highly
computer systems and facilities. Technical security scalable, centralized user access control framework.
services-Security services include the processes to ACS offers centralized command and control for all
protect, control, and monitor information access, such user authentication, authorization, and accounting
as access control and data authentication to limit data (AAA) and distributes those controls to hundreds or
access based on the particular role of health care thousands of access points in a network. With ACS,
workers, as well as audit controls to record and review network managers can control and administer user
system activity by analyzing log files of the systems. access for specialized security software solutions for
Technical security services must be performed with AAA, which are Authentication, Authorization, and
user passwords, biometrics, identity cards, or other Accountability. Authentication makes sure that the
mechanisms. Technical security mechanism-This is to right people are the only ones who can gain access to a
prevent unauthorized access to data transmitted over a network, system. Authorization makes certain that once
communications network. This requires access control, users are inside the system, they can only access the
system alarms, and processes for reporting security information and resources that are appropriate for them
breaches. In addition, data encryption must be and can be performed in a variety of ways like passed,
performed when information is transmitted over the smart card, digital certificate and by using biometric.
Internet or other public networks. Electronic signature Accountability holds people responsible for their
standards-This regulation is intended to provide non- actions by keeping detailed records or audit logs. By
refutable proof of data integrity and authentication, tracking which users are making requests, which
which allows recipients of health care documents to resources they are requesting. AAA security can ensure
confirm that the information has not been altered and that the only people who gain access to individuals’
that it originated from the claimed sender [9]. Digital health information are the right and appropriate people.
signatures are the only existing technology that can
satisfy these criteria therefore, if electronic signatures
are to be required by the final HIPAA security rule.

4. Solution for Secure Health Care

Network
4.1 Goal of designing and Implementation of
secured Network: The principle goal of design and
implementation of secured network is to provide best
practice information. This type of design focuses on the
expected threats and their methods of mitigation, rather
than just arbitrarily specifying the location of firewall Figure 1. A typical secured network with Access
and intrusion detection system. This strategy results in Control Server.
a layered approach to security where the failure of one
security system is not likely to lead to the compromise

94
4.3 Firewall based security mechanism:
In a firewall based security mechanism as shown in
figure 2, it keep a network secure [9] depends on the
configuration level. Firewalls provide barriers to traffic
crossing network and permit only authorized traffic to
pass, according to predefined security policies.
Firewalls create protective layers between networks
and the outside world. They also can log attempted
intrusions and report them to network administrators.
Firewalls are also important since they can provide a
single choke point where security and audit can be
imposed. Firewalls provide an important logging and
auditing function. However, Firewalls can't protect
against attacks that don't go through the firewall and
cannot really protect you against intruder inside your
network.

Figure 3. A typical secured network with IDS

An intrusion detection system must run continuously

without human supervision, impose minimal overhead
on the system, easily tailored to the system in question,
observe deviations from normal behavior, cope with
changing system behavior over time as new
applications are being added, survive a system crash
regardless of mechanism it is based on and must be
difficult to fool.

4.5 Network Scanning based security mechanism:

This mechanism comes under technical security
services of the HIPPA recommendation. Network
scanners conduct detailed analyses of networked
systems to compile electronic inventories of the assets
and detect vulnerabilities that could result in security
compromises. This technology allows network
Figure 2. A typical secured network with Firewalls managers to identify and fix security weaknesses
before intruders can exploit them. Vulnerability
scanning software is run from a PC inside or outside
4.4 Intrusion Detection based security mechanism: your organization and is used to find vulnerabilities on
Security threats and level of security needs of health your servers, PCs, routers, firewalls, and even printers.
care organization is increasing. We don not have only
one technology that addresses all of the problems 4.6.Encryption And Virtual Private Networks: This
rather, a layered defense approach provides the best mechanism comes under Electronic signature standards
results. Organizations should implement and encryption of HIPPA recommendations. Health
complementary security technologies to counter risks care companies must protect confidential patient
and vulnerabilities that one mechanism alone cannot information from eavesdropping or tampering during
address. A network-based intrusion detection system transmission. By implementing Virtual Private
(IDS) as shown in figure 3, provides around-the-clock Networks (VPNs) as shown in figure 4, we can
network surveillance. Unlike firewalls, an IDS establish private, secure communications across on a
monitors and analyzes packet data streams within a public network usually the Internet and extend their
network, searching for unauthorized activity, such as corporate networks to remote offices, mobile users,
attacks by hackers, and enabling users to respond to telecommuters, and partners. Encryption technology
security breaches before systems are compromised. ensures that messages traveling across a VPN cannot
When unauthorized activity is detected, the IDS sends be intercepted by anyone other than the authorized
alarms to a management console with details of the recipient by using advanced mathematical algorithms
activity and often orders other systems, such as routers, to scramble messages and their attachments. In VPN
to cut off the unauthorized sessions. networks they are emulated using data encapsulation
and tunneling that the data is wrapped with a header

95
that provides necessary routing information. In order to government mandates can clearly be overwhelming.
enhance data confidence and integrity, packets being However, the benefits of a secure network extend far
sent may be encrypted prior to entering the tunnel and beyond government compliance and the avoidance of
even if intercepted, they will remain indecipherable government imposed penalties. Properly secured SAFE
without encryption keys. networks help health care organizations to avoid
network attacks and breaches in privacy that is time
consuming and costly repercussions. Regardless of
legislation, avoiding privacy and security breaches is
pivotal to the success of a health care organization.
Therefore, designing network security architecture
SAFE [9] with adequate security mechanisms will
enable organizations to run efficient, cost-effective,
and competitive e-health operations.

References:
[1] Coalition of Voluntary Mental Health Agencies.
Proposed modifications to HIPAA privacy rule
(March 2002)
http://www.cvmha.org/policy/2002/HIPAAmodifi
cations.pdf.
Figure 4. A Logical View of Virtual Private
Network (VPN) [2] DHHS. Office for civil rights homepage.
http://www.hhs.gov/ocr/
5. Cultural Issues and Organizational [3] DHHS. Protecting the privacy of patients’ health
information: Summary of the final regulation.
Security Policy Elements Associated with
http://aspe.hhs.gov/admnsimp/final/pvcfact1.htm.
Network Security
[4] DHHS. Standards for privacy of individually
Although cutting-edge network technology is identifiable health information, Proposed rule
available to make networks secure, technology is only modification.http://www.hhs.gov/news/press/2002
an enabler. Organizational and cultural [2] issues are pres/20020321.html.
paramount in making the technology fulfill its [5] M.A.Graber, Surfs up Protecting the privacy of
potential. Indeed, as sophisticated as network-security health information on the Internet: We need new
hardware and software is today, technology is the easy privacy laws and better encryption of information,
part compared to the task of instilling a culture of Western Journal of Medicine, 176(2), 2002, 79-81.
security in an organization. Technology must defer to
culture if a security program is to succeed. [6] Health Insurance Portability and Accountability
Organizations may follow five key steps in developing Act of 1996 (HIPAA) Page
their security polices http://www.jcfa.gov/hipaa/hipaahm.htm
[7] HIPAAdvisory. Status of HIPAA Regulations:
1. Understand the drivers behind security and assess Compliance Calendar
the risk to the organization. Avoid security that hinges http://www.hipaadvisory.com/regs/compliancecal.
on end-user participation. 2. Develop policies and htm.
standards with policies in short, succinct, and
enforceable. Connect policy violations to specific [8] NEMA. Security and privacy: An introduction to
threats and make it real. 3. Define security architecture HIPAA.
and the processes to support that architecture. 4. http://www.nema.org/index_nema.cfm/704/.
Develop an education process to facilitate awareness [9] Network security solutions for health care making
and ongoing training. Listen from employees about on HIPPA safe,
the corporate grapevine. 5. Monitor the security http://www.cbeyond.net/_docs/hipaa_wp.pdf,
program through audits and other tools to evaluate white paper, Cisco systems Inc, and pp.1-7.
effectiveness. 6. Review and advise as needed and give [10] Office for the Advancement of Telehealth.
rewards for finding and reporting breaches. Telehealth Update: Final HIPAA Privacy Rules.
http://telehealth.hrsa.gov/pubs/hipaatxt.htm
6. Conclusion [11] J. Pritts, J.Goldman, Z.Hudson, A.Berenson, and
E.Hadley, The state of health privacy: An uneven
HIPAA arguably impacts the health care industry
terrain. Institute for Health Care Research and
more than any other recent legislation, causing major
Policy at Georgetown University, 1999.
organizational and financial disruptions for many
health care entities. The prospect of large new [12] G.Wachter, HIPAA’s privacy rule summarized:
technology implementations or overhauling an entire What does it mean for telemedicine?
network infrastructure to accommodate and anticipate http://tie.telemed.org/legal/issues/HIPAA2001.asp

96
 ROLE OF ONLINE THROMBOLYSIS IN CORONARY HEART DISEASE
(CHD)
Dr. Atanu Ghosh
G. B. Pant Hospital, Agartala

Dr. Tapas Dutta, Dr. Devasis Pal, Dr. Avijit Dutta

Tripura sundari District Hospital, Udaipur
Address: Ramnagar Rd. 3, Agartala 799002, Tripura.
E mail: ghosh_adr@hotmail.com.

Abstract
than in developed countries. [4]. Ever increasing young
India is heading towards a Coronary Heart victims of CHD leave behind their young Widows and
Disease (CHD) epidemic. Early thrombolysis remains dependent children to the mercy of the already
the mainstay of treatment of acute heart attack. stretched society.
Mortality reduction of 47% is observed when
streptokinase (SK) is administered within 1 hour Treatment scenario
against 11% when thrombolysis is done between 6 to 9
The mainstay of treatment of acute heart attack is early
hours. Early prehospital thrombolysis is practiced in
Thrombolysis. The administration of products like
western countries. But this is logistically not possible
Streptokinase (SK) or tissue plasminogen activator
in India. Infrastructure for thrombolysis is available in complex (tPA) intravenously opens up occluded
urban India. Vast population of rural India still does arteries, which is termed Thrombolysis. The sooner the
not have any access to this facility. Supervised treatment is initiated the better is the outcome.
thrombolysis from specialty hospital can be extended Mortality reduction of 47% is observed when SK is
to rural hospitals by utilizing Information Technology, administered within 1 hour against 11% when
which is termed “Online Thrombolysis”. One of the thrombolysis is done between 6 to 9 hours. [5].
earliest trials in online thrombolysis is started in Understandably transfer to referral hospital of acute MI
Tripura Sundari District Hospital, since August 2001. patients is not logistic.
By utilizing electronic signal medical images and
patient data can be sent to specialist centers. Some 50 to 65 % of deaths from Heart attack occur
Average time of transmission of such data takes within an hour of onset of symptoms.[6].Substantial
8.9 minutes. Complications of online thrombolysis are reduction of death is possible if treatment is initiated at
1% as against standard rate of 0.5 to 0.7%. Mortality of the onset of symptoms before transfer to referral
0.5 % is recorded in online thrombolysis as against hospital, which is termed prehospital care. For
0.4% as standard in superspeciality hospital. Online prehospital care prerequisites like regional Emergency
Medical Service (EMS), toll free three digit telephone
thrombolysis promises to bring down morbidity and
access like 911, state of the art patient transfer system
mortality in rural India at affordable cost.
and trained Emergency Medical Attendants (EMA),
who can defibrillate and thrombolyse are essential.
Key words Major trials like EMIP, GREAT have proved beyond
Thrombolysis, prehospital, online, streptokinase. doubt that prehospital care of heart attack called
“prehospital thrombolysis” can reduce mortality
substantially and has since been included as a preferred
Coronary Heart Disease (CHD) epidemic in India practice in developed countries. [7, 8].

The hard end point of CHD is Myocardial Infarction In India state of the art critical care transport system
(MI) also known as heart attack. CHD with trained paramedics for prehospital thrombolysis is
practically not feasible in near future. Semblance of
has assumed epidemic proportions in India. Prevalence facilities like thrombolysis are available in the urban
rate has increased from 2% in 1960 to 10.5% in 2000 hospitals. But the rural population does not have any
in urban India. It has increased from 2% in 1970 to access to this facility. At present there are 15.7 Million
4.5% in 2000 in rural India. [1, 2].In terms of absolute CHD patients in rural areas. Disability adjusted life
numbers, there are at present 29.8 Million CHD years (DALY) lost due to CHD in rural India stood at
patients in India. [3]. 7.69 Million for 2000, as per Global Burden of Disease
(GBD) analysis made by World Health Organization.
The course and outcome of CHD in India is more [9].
malignant and occurs 10 years prematurely

97
Solution iii) ISDN failure
iv) Microwave failure
Application of Telemedicine to provide acute
coronary care locally within golden hour is possible, v) Expensive Lease line
which is called “Online Thrombolysis”. It involves vi) High cost of digital medical equipments.
setting up of facilities with telemedicine link and vii) Failed thrombolysis.
training of local Doctors, nurses and paramedics who
can thrombolyse under supervision of Specialists viii) Nonavailabilty of specialists in specialist end
located at a distant hospital. The facility thus made is round the clock.
integrated to the local health care delivery system for
smooth running under public private partnership (PPP) Alternative cost effective approaches
program. This is called Integrated Telecardiology and
Telehealth project (ITTP). One of the earliest ITTP in i) V Sat Link
India was initiated in August 2001 at Tripura Sundari ii) Simple digitizing procedures using digital
District Hospital (TSDH) under joint collaboration of camera, Scanner
Asia Heart Foundation (AHF) of Dr Devi Shetty and iii) E mail
Government of Tripura. The project has a 6-bedded
iv) Telephone and FAX
CCU with local hospital Doctors and Nurses assisted
by paramedics provided by AHF. Telemedicine link is v) Locating specialist end of Telemedicine room
established through 2 ISDN lines and V Sat link with adjacent to CCU for round the clock availability
Narayan Hrudyalaya, Bangalore and Rabindra Nath of specialists.
Tagore International Institute of Cardiac Sciences,
Kolkata. The project covers a rural population of more Superiority of online thrombolysis over
than 8 lakh people. Prior to August 2001, they had no prehospital thrombolysis
access to thrombolysis facility.
5 – 10 % chest pain patients require thrombolysis. [11].
Early thrombolysis opens up occluded coronary Correct selection of patients becomes difficult on the
arteries by dissolving the thrombus thereby restoring part of the Emergency Medical Attendant in
reperfusion and minimizing myocyte loss. 3 months prehospital thrombolysis. In online thrombolysis
patency of such reperfused coronary artery is 70 %. specialists make screening and accurate diagnosis
[10]. That gives the patient 3 months time to undergo online and EMAs receive appropriate advice.
further tests, like Echocardiography, TMT etc in the Decreasing cost of IT will make online thrombolysis
regional referral hospital of the state. An assessment is further cost effective.
then made on the basis of the test reports and the
specialist at the state referral hospital decides further
course of management. Those requiring medical Which Lytic agent to be used in online
management are sent back home with appropriate thrombolysis
advice. A small number may require further tests like
coronary artery angiogram and advised to attend Aim: Early complete patency of Infarct related artery
tertiary super specialty hospital. This stepwise and to maintain it. TIMI 3 flow or complete flow rate
screening prevents unnecessary referral outside state. for Streptokinase (SK) is 35 % against 54 % for tissue
plasminogen activator complex (tPA). But there is
Why online thrombolysis makes sense marginal mortality benefit between the two. Moreover
complication and cost is more in case of tPA than SK.
i) It reduces mortality and morbidity as patient [12, 13]. As such SK is preferred lytic for online
gets prompt thrombolysis locally. thrombolysis.
ii) It saves time and thereby life of the patient. Evidence of benefit of Tripura ITTP project
iii) Post thrombolysed patients after stabilization (Between Aug 2001 to Aug 2004) (source: TSDH
has enough time to travel to regional referral &AHF)
hospital for further investigation.
i) No. of OPD patients treated 1373
iv) Referral hospital specialists will decide upon the ii) No. of IPD patients treated 393
next course suitable for him. iii) No. of Death 37
v) Unnecessary referral is thus avoided. iv) No. of patients thrombolysed 51
Average image transmission time: 8.9 Minutes
Barriers of online thrombolysis
i) Limited OFC route Admission during golden period Dec 2001: 20 %
ii) Limited OCB exchanges

98
 Dec 2002: 80%, this
is an indicator of increasing awareness among local
community in favor of online thrombolysis.
Mortality pattern 2. Jajoo UN, Kalantri SP, Gupta OP, Jain AP, Gupta
K. The prevalence of Coronary heart disease in the
Online SK : 0.5 %
rural population from central India. Assoc Phys Ind
Hospital SK : 0.4 %
36 : 1988; 689-693
3. Gupta Rajeev, Rastogi Priyanka. Burden of
Is online thrombolysis safe? Coronary Heart Disease in India .Cardiology
Update (Cardio logical Society of India 2003).
Complication: 1.0 % which includes Hypotension,
Minor Hemorrhage and Major arrhythmia .The 4. Enas A Enas.Arresting and reversing the epidemic
standard complication is 0.5% for SK, 0.7% for SK + of CHD among Indians. Cardiology Update (CSI
Heparin and 0.9 % for tPA. 2000).
5. GUSTO (Global Utilization of streptokinase to
open occluded arteries) Investigators.
Conclusion Streptokinase, tPA or both on coronary artery
patency. N England J Med. 329: 1993: 1615-1622.
Thrombolysis is a time tested lifesaving treatment of
acute heart attack, which is being practiced for more 6. Hurst’s, The Heart (McGraw-Hill 10th
than two decades now. Its application is ever increasing Ed.2001):1291
despite introduction of primary per cuteneous
intervention and coronary bypass surgery in acute MI. 7. GREAT Group. Feasibility, safety, efficacy of
To avoid delay in initiating therapy, it is now prehospital thrombolysis: Grampian Region Early
recommended to start prehospital paramedic Anistreplase Trial. BMJ :305, 1992, 548
thrombolysis. Prehospital thrombolysis is not feasible 8. Koren G, Weiss A, Hasin Y et al. Prevention of
in India because of lack of Ambulance and trained myocardial damage in AMI by early treatment
paramedics. The other simple and cost effective with intravenous streptokinase. N England J Med
alternative is online thrombolysis. Universal 313: 1985; 1384-89.
thrombolysis in rural India by using telemedicine can
substantially reduce mortality and morbidity as a result 9. Anonymous. Global Burden of Disease. Institute
of CHD. of International Health. Available
atwww.iih.org/about/burden.html.
10. GUSTO Investigators. Streptokinase or tPA or
Acknowledgement both on coronary artery patency, ventricular
function and survival after AMI. N England J
The authors are grateful to authorities of Asia Heart
Med.329:1993; 1615-1622.
Foundation and Tripura Sundari District Hospital,
Udaipur, Tripura for the data used in the paper. 11. Hurst’s. The Heart (McGraw-Hill 10th Ed.
2001):1294.
12. ISIS-2 (Second International Study of Infarct
Reference Survival) Collaborative Group. Randomized trial
of intravenous streptokinase among 17,187 cases
1. Gupta R,Gupta VP,Sarana M , et al. Prevalence of
of AMI: ISIS-2. Lancet.2; 1988; 349.
Coronary heart disease and risk factors in an urban
Indian population. Indian Heart Journal 13. GISSI: Gruppo Italia no per lo studio della
54:2002:59-66. streptochinasi nell’Inferto Miocardio.
Effectiveness of intravenous thrombolysis in AMI
Lancet .1; 1986; 397.

99
 DATA CONFERENCING OVER A LOW-BANDWIDTH
COMMUNICATION LINK
Anunay Nayak *, V.Pallavi
Department of Computer Science & Engineering
Indian Institute of Technology, Kharagpur, West Bengal, India.
anunay_n@yahoo.com, pallavi@cse.iitkgp.ernet.in

Abstract telemedicine is a great help to the common people. At

the Indian Institute of Technology, Kharagpur, we have
Telemedicine can be defined as the use of developed a telemedicine system (Telemedik) for the
telecommunication technologies to provide medical treatment of various categories of diseases [7].
information and services. It involves the practice of Telemedik is being used in the Government hospitals
delivering health care over a distance using of West Bengal. Development of the Telemedik
telecommunication equipments as simple as telephones System [8] has great challenges, as there are several
or as complex as PCs and full-motion interactive perspectives to be considered. The System has to run
multimedia. Telemedicine involves the electronic efficiently on POTS and even in high bandwidth
conveyance of medical information for the purpose of channels like ISDN. This causes problems on the
diagnosis and treatment of patients using PC, online conferencing session .So the transfer of images
telecommunications links, specialized video, audio, during the online session was discouraged in the
and imaging equipment. Telecommunication links Telemedik System and an interactive protocol for
include modems connected to analog telephone lines sending multi-resolution images has been designed.
operating at speeds up to 56Kbps, ISDN lines of up to
128Kbps, and satellite links of 100 Mbps. This paper
gives a description of different features and 2. Telemedik System
functionalities of data conferencing tools developed at
The Telemedik system supports medical
IIT Kharagpur for the diagnosis and treatment of
patients from a remote place. consultation between distant hospitals/health centers,
which are classified as Nodal Centers and Referral
Centers. At the nodal center, a patient reports for
Key Words teleconsultation and gets registered. Subsequently the
Telemedicine, Data Conferencing, Communication, patient’s personal and medical information are entered
Online Consultation. into the system in the form of text, hypertext, graphs,
contours, digital images, video, audio etc. The patient
record is stored in a local database and is also sent to
1. Introduction the referral center. Specialist doctors are available at
referral centers. On receiving the data at the referral
There has always been a paucity of health-care
center, specialist doctors browse through the patient’s
providers, especially specialists, in rural areas.
records and send their advice. This is the offline mode
Telemedicine provides a means for specialists to
of consultation supported by the system. Online (live
electronically travel to underserved areas, thereby
consultation) session is also supported by the system in
increasing the size of their region of coverage without
which the doctors can discuss with each other.
physically traveling to these areas. Telemedicine can
thus provide direct patient care by which a doctor and a Figure 1 shows the communication link established
patient are geographically apart but by means of between a Nodal Center and a Referral Center.
communication devices, the patient can obtain the
doctor’s diagnosis and treatment. Telemedicine can
also provide teleconsultation through communication
among doctors or between doctors and other related
personnel by which the knowledge of experts can be
transmitted to nonspecialists to support medical care.
Telemedicine consists of both real-time interactive
consultancy as well as batch processing of patient data
and information for diagnosis at a later time. Different
telemedicine systems are installed not only in advanced
countries [1], [2], [3], but also in developing countries
like China [4], Thailand [5] and even in India [6]. In a
country like India, where basic health care facilities are
Figure 1: Telemedik System
poor and people have to travel a lot for obtaining
expert opinion and the treatment for a disease,

100
2.1. Online Graphics Communicator
Online Graphics Communicator is one of the most
important software modules of the Telemedik System.
This Online Communicator is a Graphics, Image as
well as a Textual Communicator. Doctors from the
Referral as well as Nodal Centers can have an online
session between themselves if they feel its necessity. A
Nodal Center can connect with any referral center with
which it is attached. Similarly a Referral Center can
also give online consultation to any Nodal center with
which it is attached. At a scheduled time doctors at
both end consult with each other in front of the data
conference console using the graphics communicator.
Initially they have to select the patient for which they
Figure 2: Online Graphics Communicator
want to have an online session. After the connection is
established, the doctors can do textual chatting with
each other using the Textual Communicator present 2.2. Sending Referral Doctors Remarks With
with the Online Graphics Communicator. The Graphics Annotations Over Images
Communicator provides a common canvas for
conferencing with graphic primitives (i.e. by drawing, Sending Referral doctors’ remarks with
annotating etc.) with the patient data. As the same annotations over images to the Nodal Center is another
patient data is available at the both end, the actual important feature of Telemedik System. This is an
bandwidth requirement is very small. This kind of Image Marking Utility by which the Referral Center
conferencing is possible with the various backgrounds doctors can view any type of bmp or jpeg images of a
such as white background or white board, a medical patient. The doctors can also do markings and text
test image, special images such as human profiles etc. annotations on these images. This is the offline mode
The module supports conferencing with multiple of teleconsultation provided in the Telemedik System.
canvases (presently four in our implementation). After viewing the patient’s images and giving their
Operations such as drawing lines, contours, circles, advice, the doctors can also save these images in bmp
ellipses, text annotations etc. are provided. Some image or jpeg format. These marked images are then sent to
processing operations such as zooming of patient’s the Nodal center. The doctors at the Nodal Center can
images or a portion of image etc are also supported. then view the images marked by the Referral Center’s
The colors of the graphics primitives identify the doctors as well as their diagnosis and later give a
centers communicating with each other. Figure 2 proper advice to the patients. This is a very cost-
shows a patient’s image at the Nodal Center marked by effective tool. Here the communication charges
a Referral Center’s doctor during an online session incurred is very less as the doctors give their advice in
using the Online Graphics Communicator. The Online an offline mode and a connection with the remote end
Graphics Communicator also has an Image is established only while sending the patients data.
communicator designed for sending a new image from Figure 3 shows an image of a patient being marked by
the nodal center to the referral center. As the transfer of a doctor.
image data in low speed communication link will take
considerable time, an interactive protocol for sending
multi-resolution images has been designed. The Online
Communicator is designed in such a way, that it can
function even at a very low bit rate data
communication channel (telephone line). Instead of
transferring the images, only the coordinate points of
the markings are sent across the communication
channel, which makes the online session more real and
interactive. In fact, few trial runs were conducted with
the System between Bhubaneshwar and Kharagpur and
it was found that this Online Session could be done at a
speed as low as 5 Kbps during the peak hour of the Figure: 3 Sending Referral Doctor’s Remarks
day.

101
2.3. DICOM Compatibility Burdwan Medical College Hospital and N.R.S Medical
College Hospital were the Referral Centers. In our
The Digital Imaging and Communications in
experimentations, we used both the ISDN links and
Medicine (DICOM) standard was created by the
telephone lines even in the busy hour (between 11 AM
National Electrical Manufacturers Association [9]
to 2 PM) of a working day. Also, in-house testing over
(NEMA) to aid the distribution and viewing of medical
LAN has been performed. In the experimentations, a
images, such as CT scans, MRIs, and ultrasound. The
set of patient data, test reports and images like blood
goal of the DICOM Standard is to achieve
slides, X-Rays were captured at the Nodal Centers. The
compatibility and improve workflow efficiency
data was organized, distributed and sent to the Referral
between imaging systems and other information
Centers. At the Referral Center, the images were
systems in healthcare environments worldwide. A
browsed. Annotations and markings were made on the
single DICOM file contains both a header (which
images and were again sent back to the Nodal Centers.
stores information about the patient's name, the type of
Online telemedicine sessions were also carried out.
scan, image dimensions, etc), as well as all of the
Images were sent in the online session from Nodal
image data (which can contain information in three
Centers to the Referral Centers. Textual
dimensions). DICOM is the most common standard for
communication was performed which was followed by
receiving scans from a hospital. The Telemedik System
Graphics Communication in which the images opened
has an integrated Dicom Viewer by which the doctors
in the Nodal Centers’ canvas were displayed in the
can view the DICOM images of patients. The formats
Referral Centers canvas too. ). Operations such as
of the DICOM images supported by the system are:
drawing lines, contours, circles, ellipses, and text were
• Monochrome 1 native images. also performed. It has been observed that the data
• Monochrome 2 native images. conferencing could be done even at a low speed
The doctors can also mark these DICOM images and bandwidth of 5 Kbps.
save the images again in DICOM format. Thus the
doctors in the Telemedik System can view the DICOM
images obtained from different medical instruments 4. Conclusion
supplied by different vendors. Figure 4 shows a
DICOM image of a patient being marked and This paper has described different features and
annotated by a doctor using the Dicom Viewer of the functionalities of the Telemedik System developed at
Telemedik System. the Indian Institute of Technology, Kharagpur. This
system is running successfully in different Government
hospitals of West Bengal like Purulia District Hospital,
Suri District Hospital, Midnapur Sadar Hospital,
Burdwan Medical College Hospital and N.R.S Medical
College Hospital. One of the most important and used
feature of the Telemedik System is its Online
Communicator, which can be used even in a low
bandwidth communication link. The doctors of these
hospitals find these tools very useful to provide best
teleconsultion.

5. Acknowledgements
This work is sponsored by the Ministry of
Information Technology, Govt. of India. We also
acknowledge Prof A.K.Majumdar and Prof. Jayanta
Figure: 4 DICOM Viewer Mukopadhyay from Computer Science and
Engineering Department, IIT Kharagpur who are the PI
and Co-PI of the project under whose guidance the
Telemedik System could be developed.
3. Experiments & Tests

Several experimental trials had been carried out

6. References
between Kharagpur and Calcutta, separated at a
distance of 116 Km. Besides, many experimental trials [1] Robert Sh. H. Istepanian. Telemedicine in the
have also been conducted between different United Kingdom: Current status and future prospects.
Government hospitals of West Bengal like Purulia IEEE Transactions on Information Technology in
District Hospital, Suri District Hospital, Midnapur Biomedicine, 3(2): 158 -- 159, 1999.
Sadar Hospital which were the Nodal Centers while

102
[2] T. Takahashi. The present and future of [6] B. Jankharia. Current status and history of
telemedicine in Japan. International Journal of teleradiology in india. International Journal of
Medical Informatics, 61: 131--137, 2001. Medical Informatics, 61: 163--166, 2001.

[3] H.S. Chen, F.R. Guo, C.Y. Chen, J.H. Chen, and [7] Mukherjee, Jayanta; Majumdar, A.K.; Banerjee,
T.S. Kuo. Review of telemedicine projects in A.; Acharya, B.; Nayak, A. and Reddy, U.V
Taiwan. International Journal of Medical Telemedicine for Leprosy. IETE Technical
Informatics, 61: 117--129, 2001. Review, Vol. 18, no. 4, pp.243-252, 2001.

[4] R.K.C. Hsieh, N.M. Hjelm, J.C.K. Lee, and J.W. [8] http://www.telemedik.iitkgp.ernet.in
Aldis. Telemedicine in China. International
Journal of Medical Informatics, 61:139--146, [9] http://www.medical.nema.org
2001.
* Mr Anunay Nayak was associated with the
[5] N. Kasitipradith. The ministry of public health Telemedicine Project at Computer Science and
telemedcine network of thailand. International Engineering Department of IIT Kharagpur from
Journal of Medical Informatics, 61:113--116, 2001 November 2001 till April 2004.

103
 STRATEGIES IN ASSESSING TELEMEDICINE QUALITY

M.PalaniNathaRaja§ Subash Wadhwa£ S.G.Deshmukh£

§
Research Scholar
Department of Mechanical Engineering, IIT Delhi 110 016
Email: pnatharaja@tce.edu
£
Professor
Department of Mechanical Engineering, Indian Institute of Technology, Delhi 110 016

Abstract professional. Information Technology plays a major

role in storing and retrieving the records at the
Health care is becoming a business giant to human convenience of doctor to decide on the issue.
beings in all parts of the world. Information technology Technological advancement in information systems
is a buzz word in the modern context, since it interestingly it is more sophisticated and influenced.
influences the every aspect of the common man. It Interestingly, as a social scenario the support of the
plays an important role in delivering the modern health patient and its co-operation is not recognized in
care service too. It has been practiced in the developed telemedicine.
countries and there are numerous exploring focuses
elsewhere. Information and communication technology Challenges of Telemedicine
has the prospective to supplement the activities
involved in health care service. Telemedicine is an India is a diversified country in cultures and the
information based business, which provides basis for people need to be trained in using the Information and
the treatment at high speed and accurately. It is well Communication Technologies (ICT) for the health care
known that the doctor’s availability is minimal in rural service. And on the other hand, health care professional
areas to the common public and hospitals do not have has to provide a fullest support to implement the
proper infrastructure to cater the health care service. system of telemedicine. It is very difficult for them to
Literatures have been surveyed to identify the various spare the time for this activity by the doctors initially.
factors associated with quality of health care. Other The system may incorporate the paramedical staff to
online assessments of quality have been consulted and initiate this activity. In due course, the patient will
are reported. This paper is an attempt to chart out the acclimatize to the system since the patient might be
possibility of establishing criteria in assessing the seen as a “problem to be solved”. Off course one has to
quality on telemedicine system. agree that much of the diseases like psoriasis, mental
health conditions may not be managed effectively
Key words through telemedicine. However, the relationship will
Telemedicine, Healthcare, Service Quality, E- be maintained by the system between patients and
quality, Standards and assessment health care consultant to raise the issues to move
ahead. For a stranger, any doctor provides a treatment
physically, because the consultant may get the
Information Technology information through the patient. In health care service,
the value of the first visit relies on the information
Rose et al.(2004) stated the healthcare is about provided by the patient, in terms of expressing
meeting the physical, psychological and social needs of themselves well, clearly and accurately. Sometimes to
a person who seeks care. There are certain confusion the extent the patient express the importance of his pain
exist in telemedicine, teleconferencing, health and related issues. For an effective and crucial decision
informatics commented by Curry et al.(2002). on the treatment, it is clear that the knowledge and the
However, information technology mediates the role information available to the specialist needs to be
between human or patient and health care professional, accurate and detailed one. Informally it helps the
but not necessarily a doctor. It started interfacing the patient to know his health, history and experiences.
subjects of health care on mental health, diabetes, The benefit of ICT is reaped through the interaction
retinopathy, foetal monitoring and accident and between patient(s) and healthcare professional(s) or
emergencies care. Doctors are not embracing the between health care professional in the workplace. The
technology with the same enthusiasm as other written record about the patient and the information
professionals in the usage of computers. Health care provided by the patient may not be potential in
service improvement needs two basic objectives to be information-sharing between those working in primary
fulfilled. One, it may provide more information to the and secondary health care. ICT helps in making the
patient about the health care. Two, it tries to use ICT crucial decisions by the decision making utility
effectively to support the interaction with health care available with the telemedicine system. Although, the

104
system cannot equate to the senses like, taste, smell and Telemedicine requirements
touch, but it provide light and sound that might help to
understand the context. Conventional service quality • High definition images – Scanners/Digital camera
dimensions cannot be instantly applied to determine the • Shared access to data applications –
e-services, since interpersonal relation is a typical Networking/Satellite
feature. Infrastructure of the Data, storage space and
speed are the forth coming technical challenges in the • Computers and Peripherals – Security and storage
telemedicine. Integrity of the data and its security is devices
another concern, since the data in the global • Medical systems that might be considered useful
environment has to exchange patient information and in presenting clinical data - databases and
the vulnerability of the information. Hospital data is algorithms
seldom standardized in the health care providers and it
influences the quality of the service. • Medically trained person to mediate
specialist/consultant – Training , Time

Images Computers
through and
Scanner/ Peripherals
Camera

Networking

Databases Medicallytr
And ained
Algortihms person

Figure 1 Telemedicine Components

Issues and problems affecting the quality of tends to increase the interaction between patients and
telemedicine system doctor. But lot of training is desired to improve and
access to treatment and understanding. It either
Nature of the information is highly enriched to the enhances or reduces the cost of medication. It
doctor/healthcare professional and this aid in their accelerates the adoption of new information and new
diagnosis and subsequent care of their patients. treatment available in the world through effective
Practically, updating and maintaining the patient consultations. The decision will be made on site; hence
information are unwieldy. This results in large gap it enhances the psychological confidence to the patient
between patients and doctors. Also, the benefits are about the telemedicine system. The quicker diagnosis
mutual to the patients and the doctors. It provides an helps in saving the time and it helps in achieving the
opportunity to the patients with flexibility of treating at routine daily life. The initial contact will play an
the time, place and pace, they can interact with health important role in the future treatment. The physical
care professionals. It helps in improving the quality of trauma associated with the subsequent visit to the
diagnosis and treatment, because the online treatment hospital is alleviated.

105
 There are many empirical studies carried out in healthcare. Appropriateness, Effectiveness, Efficacy,
the recent past, show that there are number of common Timeliness, Availability, Respect and caring,
problems that telemedicine patients have encountered, Continuity, Tangible, Safety and Efficiency are the
which include dimensions responsible for the Quality performance of
a telemedicine system. Initial e-quality dimensions
• Technical problems, including bandwidth could be classified as reliability, responsiveness
limitations, browser problems etc., customization , assurance/trust and user interface. The
• Inaccessible doctors problem of integrating in the implementation of a
• Lack of depth in treatment. picture archiving and communication systems in health
care is reported by Fu.H et al.(2003). The top level
• Lack of intimacy associated with traditional
hospitals need more customized special requirements
environment.
such as integrating of the IT systems and picture
• Limited interaction with doctor archiving. Further experience leads the e-quality
• Delays in receiving clarification dimensions as reliability, responsiveness, access and
Application of Information Technologies without flexibility , ease of navigation, efficiency,
appropriate treatment. assurance/trust, security/privacy, price knowledge, site
aesthetics and customization/personalization.
Customization and Personalisation is equivalent to
empathy in interpersonal services. Personalisation
Need for assessment in telemedicine could also mean customizing the service to individual
preferences. Health care rules are of limited use; the
Even in conventional healthcare services patients value provide an organization framework that
“physicians are unresponsive to patient complaints” is permits them to respond to the patient in a variety of
the common phenomena. The primary objective of any situations. Duty hours of a doctor are viewed as a
modern health care institution is to ensure that in the complicated problem requiring direction of particular
care of patients all required information for medical rules. In order to ascertain the quality healthcare there
diagnosis is both accurate and available to health care should be strict professional standards, high empathy
professionals. The ultimate objective of integrated with customers and long training or capital intensive
healthcare system is to implement not only the Rogelio Oliva(2002). The issues addressed above need
operational management system, but also the strategic an understanding on the system and measurement alone
quality assurance system. Accountability in the health can ensure the system that it can operate in an e-
care service is arguable. Health care service is environment.
intangible in nature and in many decades experts were
of the opinion that evaluating the intangible service
may not be prompt. But due to the compulsion, many Standards and criteria for quality of
intangible services (eg. Education) are in the way to
assess their service quality. It gives the correct path to
telemedicine system
improve the existing system for future sustenance.
Measurement is a process and it requires a lot of
“Service” is the newer conceptualizing research and
spade work in establishing standards and criteria. There
many findings are conflicting in nature. Quality is the
have been a variety of perspectives about telemedicine
term which explained more in business and what should be the effective criteria for measuring
competitiveness and its relevance to the cost of the the quality. Quality in telemedicine comprises three
business. On the other hand, the quality is of meeting crucial elements: Quality as value for money, quality as
the customer expectations by encountering the meeting the expectations of the patient, Quality as
constraints fixed by the customers. Quality has been delivering the treatment and monitoring the patient.
explained in different contexts and it is of process Operating standards for assessing the quality of
specific. Telemedicine is a process centric and telemedicine have been proposed, which include:
associated with high labour intensive and more of
sophisticated infrastructure necessary for the system to • Patient success rate
be implemented. Health care service cannot be equated • Good knowledge rating
with other service process since it is of life oriented • Skill rating in treatment
and beyond anything else. The recent research model
based on IT techniques is as IT Infrastructure,
There are many discussions and studies used by
Healthcare management, Quality performance. The IT
the consumers to judge the quality of e-environment
infrastructure will be assessed on the standardization, are: (Dobbs, 2000)
integration and sophistication of networks. Scientific,
Patient care, customer satisfaction and administrative • Ease of navigation ie. Free movement with in
are the four dimensions in the management of the website

106
 • Scope of the site coverage. - Contents
• Details available on the topic – Details No. Components Assessment factors
1. Images through Reliability, Tangibility
• Correctness of information provided on the Scanner/Camera
site – Trust 2. Computers and Navigation and
• Completeness represented at each site – peripherals Flexibility
Knowledge 3. Networking Access and Security
• Interest inspired of satisfied by the site- 4. Data bases and Privacy and Efficiency
Education Algorithms
• Site aesthetics. - Attraction 5. Medically trained Responsiveness,
person Assurance and Trust
• Utility of the site for any other purpose –
Value Table 1.0.Telemedicine Components and
• Linkage of the site , directly or indirectly to Assessment factors
other sites – Help
Reliability is the equipment operating condition,
Rajiv Gandhi National Quality Award is aiming which support the capturing images without failure.
towards excellence model with added element for Tangibility is the physical evidence of an object and
better business proposition. This award fit very well how conveniently it happens to the patients. The
into a TQM approach and the award provides a base images captured through the digital devices should be
visualized with more navigational ability and flexibility
for commitment to quality and for continuous quality
to create in the doctors’ mind of patient images.
improvement.
Networking of the system has to provide uninterrupted
The examination points for Rajiv Gandhi National access and to maintain medical confidentiality in a
Quality Award are Leadership, Policies and Strategies, more secured environment. Also, the data gathered
during the diagnosis may be individualized/privatized
Human resource Management, Resources, Processes,
and its efficiency shall be improved. The system is in
Customer Satisfaction, Employee Satisfaction, Impact need of medically trained person and he/she should
on Society and Business results. The main thrust is on possess the quality of replying the queries/clarifications
leadership, process management and customer to the patients. Through competencies, the assurance
satisfaction. Detailed guidelines are provided to shall be given to patients with more trust. Managing
enterprise to interpret correctly the scoring system. It telework environments is a typical one and the
categorizes the award for different industrial sectors; following recommendations are aid to improve the
also the mentioned factors shall be re oriented for management of the process Kurland et. al(2000). They
health care sector, further leads to the telemedicine are: Guidelines, Infrastructure, Availability,
environment. Communication, Scheduling, Trips to the main office,
and Performance measurement. Interoperability is yet
Similarly, the European Foundation for Quality another issue need to be addressed. Similarly,
Management (EFQM) excellence model has been used scalability is required to suit the dynamic changes of
by several organizations and an International EFQM the system rapidly. In sum, effective preparation of
Health sector group was set up to share experiences assessing criteria will facilitate a growing telemedicine
relative to it. However, this model is meant for system. Finally, the criticality of assessment is how the
individual skill sets meet the requirements of
delivering awards and to involve in self-assessment. It
telemedicine environment.
is felt that the quality management models have been
redefined to suit the telemedicine system.
Conclusion
Parasuraman et al.(1988) formulated a measure of
service quality derived from data on a number of
Various research journals have been consulted in
services. The initial findings yielded ten dimensions of identifying the assessment criteria for the health care
service quality that included Tangibles, Reliability, services on e-environment and are reported. It is the
Responsiveness, Competence, Courtesy, Credibility, high time in establishing the criteria for assessing the
Security, Access, Communication and Understanding quality of a telemedicine system, since the system itself
the customer. Numerous studies have correlated the is in a budding stage. The further study on these factors
above dimensions for a e-environment. In order to suit may conclude the precise factors and its analysis leads
the modern context of evaluation and from the study, it to know the accurate weightages for the assessment.
is concluded that tabulated factors are possible This type of quality management and monitoring
assessment factors, which are responsible for essentially gives a better quality performance on the
measuring the quality of the telemedicine system. system, which ultimately benefit to the patients in
particular and society in general.

107
References with IHE”, International Congress Series 1256,
(2003), 915-923.
1. Rose Raduan Che, Uli Jegak, Maohamed Abdul
and Kim Looi Ng, “ Hospital Service Quality: a 5. Kurland Nancy B and Bailey Diane E, “ Telework:
managerial challenge”, International Journal of The advantages and challenges of working Here,
Health Care Quality Assurance”, Vol. 17. No.3, Ther, Anywhere, and Anytime, IEEE Engineering
2004, pg146-159. Management Review, 2nd Quarter, (2000), pp.
49-59.
2. Curry A, and Sinclair E, Assessing the quality of
physiotherapy services using SERVQUAL, 6. Parasuraman A,Zeithaml V.A and Berry L.L,”
International Journal of Health Care Quality SERVQUAL: A Multiple Item Scale for
Assurance, 15/5, (2002), 197-205. Measuring Consumer Perceptions of Service
Quality”, Journal of Retailing, Vol.64, No.1,
3. Dobbs K. "What the online world needs now”, Spring (1988), pp. 12-40.
Training, (2000), Vol.37, No.9, pp. 84-87.
7. Rogelio Oliva, “ Tradeoffs in responses to work
4. Fu H , Jin Zhengyu, Dai Jianping, …., “Picture Pressure in the Service industry”, IEEE
Archiving and Communication system in China: Engineering Mnagament, 1st Quarter, (2002), pp.
the development, problem, and integrating strategy 53-62.

108
 TELEMEDICINE SYSTEMS - AN OVERVIEW
P. Hari Krishna Prasad
Lecturer, Dept. of ECE, National Institute of Technology, Warangal-506004, India.
phkp@nitw.ernet.in

Abstract iii. a processing unit: Basically the processing unit

This paper presents a brief overview of the acquires the data from the biosignal acquisition
essential elements of a general telemedicine system, module and performs necessary operations to
the evolution of telemedicine technology, the make the data collected suitable for the next stage
classification of telemedicine systems followed by operation.
latest developments and future trends in telemedicine iv. a communication module: This includes a
systems. modem that is appropriate for GSM or Satellite
Telemedicine building blocks classification or POTS.
future trends
In addition to the biosignals the Telemedicine
Introduction Unit transmits information concerning the monitor,
Telemedicine [1] is defined as the delivery of such as the alarms or the monitor status.
health care and sharing of medical knowledge over a
distance using telecommunication means. The
availability of prompt and expert medical care can Base Unit
meaningfully improve health care services at The base unit, located at the place where the
understaffed rural remote areas and also proves to be signals and images of the patient are received, consists
very important in several emergency cases. Continuous of a dedicated PC equipped with a modem
monitoring of vital bio-signals of a patient, either in (responsible for the data interchange) suited for
Intensive Care Unit (ICU) or in homes, and communication in GSM, Satellite or POTS mode and
simultaneously displaying all telemetry information to capable of receiving or sending data from/to the
the competent doctor anywhere, anytime ensures telemedicine station.
proper monitoring of the health condition of the The biosignals and other relevant information
patient. To design and develop a telemedicine system are transmitted, either in real mode or store-forward
of practical utility, an overview of the features of the mode, after subjecting the data to compression,
existing systems is a prerequisite. In this context, this encryption and using a suitable network protocol to
paper presents some of the important basic information reduce the bandwidth requirements, to ensure the
secrecy of transmission and nearly error free
needed for better appreciation of telemedicine concept.
transmission.

Basic Elements Of A Telemedicine System

Evolution Of Telemedicine Systems
A typical telemedicine system [2] consists of
Telemedicine unit, Base unit or Doctor’s unit which The evolution of telemedicine technology[3] can
can be portable or not portable depending on the be broadly classified as
application for which the system is designed. The Generation I-
design exercise is preceded by a detailed user
requirements analysis and the corresponding system Stage1: Early development,pre-1970-Primarily Video
functional specifications. based and cable television technologies.
Stage2: The 1970’s- Use of a number of
Telemedicine Unit communication technologies –microwave,
The telemedicine unit, usually located at the broadband television, picture phone using
patient’s site, mainly consists of four modules, namely, switched networks and cable television. The
i. biosignal acquisition module: This part is applications addressed are providing specialty
responsible for biosignal’s acquisition. This consultation, medical undergraduate
module will be designed to operate with some of education, continuing education to isolated
the most common portable biosignal monitors rural physicians, hospital administrator
used in emergency cases or in intensive care conferences and patient education.
units. Stage3: The early to mid 1980’s-The use of telephone
ii. a digital camera: This unit captures the images and black-and-white slow scan television have
and make them available for the processing unit.

109
 been considered as the most inexpensive patients. Systems used for synchronous
telemedicine technologies, communications may include specialized telemedicine
roll about interactive video units, interactive video
Generation II- The early to mid 1990’s- Use of digital
communications, T1 or multiplexed ISDN or room systems, computer- based desktop video
ATM networks, synchronous video conferencing units, videophones, portable or brief case
conferencing through compressed laptop videoconferencing. These may be supplemented
technologies, electronic peripherals such as a with peripherals including electronic stethoscopes,
document camera, electronic stethoscope, X- other electronic scopes, view boxes, combinations of
ray scanner, 3-chip camera and the emphasis telemonitoring devices, graphics stands, hand-held
being shifted to the implementation and mobile and wireless systems used for records,
maintenance of telemedicine services that are prescriptions and orders.
less expensive followed by the prominent use Asynchronous systems are used in connection
of internet. with non real time data monitoring which generally
Generation III- Decline of interactive video and the consists of still images, e-mail and possibly video clips.
rise of store-and –forward technologies Asynchronous telehealth is more often used for
(considered as more cost effective, more teleradiology or telepathology where the patient does
universally available and result in less time not need to be present for interactive communication
spent per consult when compared to and for independent continuing education. Store and
interactive video). forward technologies used in asynchronous services are
mostly PC based while some are software only.
Asynchronous systems include film scanners, DICOM
Classification of Telemedicine Systems compliant teleradiology systems, still image
The telemedicine systems can be classified into management systems, PACS, Video microphones,
remote and regional telemedicine systems based on the Dermascopes and more.
distance of coverage [4] over which the service is to be
operated. The aim of the services is to provide medical Cable modems and xDSL [3] would offer
care to the distant, underserved locations. bandwidth at or above ISDN and at lower cost. But the
difficulties in deploying the ISDN have limited its
The other classification is based on sophistication growth. Difficulties in installing satellite based systems
of the services involved. Based on this criterion, the may limit use in any but remote areas where there are
services are classified into startup level, medium level few terrestrial options. It is very likely that internet will
and high level. The startup level system involves remain the least costly and most commonly available
providing medical care to a patient in a small hospital universal network for some time. Due to bandwidth
with diagnostic equipment/ a super specialty hospital limitations and reliability, however, it remains unlikely
through videoconferencing. The medium level that interactive video through the internet will replace
telemedicine system involves the patient in a small room videoconferencing systems in the foreseeable
hospital / a super specialty hospital with only one future. Internet – based video remains too cumbersome
telemedicine center for teleconferencing and equipped and erratic and the resolution is not high enough for
with its own diagnostic equipment for referrals to other most synchronous telemedicine applications. On the
consultants through videoconferencing. The high level other hand, asynchronous multimedia telemedicine
telemedicine system involves applications continue to grow.
i. interconnection of each department of the super
specialty hospital through LAN for telemedicine Typical Telemedicine System
ii. capture and transfer of patient data using high
speed network The features of a typical telemedicine system [5]
iii. store the patient data of all the departments at one include programmability, high performance, flexibility
place and retrieve the data from multiple locations and upgradability. For instance, the tele-medic system
of the network. [6], developed by HW Communications Ltd in
collaboration with Simoco Digital Systems limited and
iv. Each specialist doctor can access the patient data Langdale /Ambleside Mountain Rescue Team
and treat from his place. (LAMRT), gives the mountain rescuers the capability
v. super specialty hospital provides consultancy to to send GPS positioning information, text and status
smaller hospitals through videoconferencing. messages, digital images and medical information such
as ECG, pulse rate, blood pressure and oxygen
Based on the nature of data transfer [3] the saturation level from the casualty site up on the
telemedicine systems can be classified as synchronous mountain to the mountain rescue base. The position
and asynchronous systems. Synchronous systems are information is useful for effectively managing search
used in connection with real time transfer of audio, and rescue operations from the base. The digital images
interactive full motion video and still images of

110
of injured and medical data transmitted over TETRA References
(Terrestrial Trunked Radio) from the casualty site will
give a better indication of the condition of the casualty 1. http://www.pubmedcentral.nih.gov/articlerender.
and a doctor can remotely give immediate advice to the fcgi?artid=153497
rescuers in the field on administering advanced
treatment. 2. http://www.google.co.in/search?q=cache:VA8lp9-
TBR8J:www.bel-
Latest Developments and Future Trends in india.com/Website/StaticAsp/Syst_telemedicine.as
Telemedicine Systems p+telemedicine++systems%2Bdesign+&hl=en&ie
=UTF-8
Hearing and sight are the only senses used in
telemedicine today. Engineers at the Pacific Northwest 3. http://www.google.co.in/search?q=cache:
Laboratory have developed Telesmell, a prototypical q2XpmnIqyZMJ:www.hci.uu.se/courses/1md100/
system that would use an electronic nose to capture the vt01/material/telemedicine/telmed-
essence of odours, encode and transmit the data to the evolution.pdf+basic+telemedicine+equipment%2B
telemedicine site, and then use a decoder to reconstruct hardware%2Bsoftware%2Bhistory&hl=en&ie=UT
the odour for the consulting expert to smell. F-8
Alternatively, a neural network might be introduced to
analyze the odour for the operator, negating the need to (Mary Moore, ‘The evolution of telemedicine’,
reconstruct it remotely. An ultrasound scanner that fits mmoore@choctaw.astate.edu
in the palm of a medic’s hand is being proposed. For
Later published in Future Generation Computer
rapid diagnosis of the illness of the space shuttle crew
System 15(2), 1999, ( Elsevier Science
members on board,NASA has developed a suitcase –
Publishers,1999) Special issue on IT IS-an
sized package, the telemedicine instrumentation
international telemedicine information Society,
package, that contains an endoscope, ophthalmoscope,
245-254
dermatology macroimaging lens, ECG, automatic
blood pressure sensor, electronic stethoscope, pulse 4. http://www.google.co.in/search?q=cache:hGaz
oximeter and a computer with two –way voice and 7JiTXYoJ:icsl.ee.washington.edu/projects/gsp9/sp
video control. ie95/ms5000/+telemedicine+system%2Bgeneral+b
lock+diagram&hl=en&ie=UTF-8
Conclusions 5. http://www.google.co.in/search?q=cache:hGaz
This paper presented some of the 7JiTXYoJ:icsl.ee.washington.edu/projects/gsp9/sp
prerequisites needed to understand telemedicine ie95/ms5000/+telemedicine+system%2Bgeneral+b
systems along with a few latest developments. lock+diagram&hl=en&ie=UTF-8
6. http://www.google.co.in/search?q=cache:
Acknowledgements EKQ_yzLGJNIJ:www.hwcomms.com/EMM.htm+
medical+devices+with+communication+capability
The author is grateful to Dr. G. Vijaya, Assistant &hl=en&ie=UTF-8
Professor, Dept. of ECE, NIT, Warangal for his
constant encouragement. The author is thankful to the
authorities for providing the facilities to carryout this
work.

111
 THE DESIGN OF A WEB BASED TELEMEDICINE SYSTEM

Amiya Kr. Maji, Jayanta Mukhopadhyay.

Department of Computer Science & Engineering,
Indian Institute of Technology, Kharagpur.
Kharagpur – 721302. India.
amiyamaji@gmail.com, jay@cse.iitkgp.ernet.in

Abstract videoconferencing, the internet, store-and-forward

imaging, streaming media etc. over the last few
Evolution of technologies such as interactive decades have been significant steps towards achieving
videoconferencing, the internet, store-and-forward that goal. This paper highlights the design of a web
imaging, streaming media etc. over the last few based telemedicine system, which can be used by
decades have broadened the horizon for several patients, doctors as well as organizations such as
research areas in the healthcare industry. Telemedicine hospitals for availing themselves of healthcare services
is one of these areas where we use computers and over the Internet. The telemedicine system described in
telecommunication technologies to provide medical this paper is a modification of TeleMedik, a software
information and services when distance separates the developed by researchers of Telemedicine Laboratory
participants. It is a tool for unification of healthcare at Indian Institute of Technology, Kharagpur. Though
services of different locations irrespective of their web-based telemedicine systems have been attempted
distances for better sharing of resources and expertise. by many researchers, most approaches are
In developing countries like India telemedicine can be experimented with the use of videoconferencing for
extensively used to extend healthcare services to the remote consultation [3]. Our approach differs in the
rural and urban underserved people. Several fact that this system at present uses store and forward
researchers across the world have tried to develop technology for consultation between patients and
telemedicine systems to meet the needs of doctors and doctor with extensive support for medical multimedia.
patients. Most of these systems are limited to specific This paper also outlines some areas of research for the
areas of medical sciences like teleradiology, improvement of the existing system.
teleoncology, orthopedics, videoconferencing for
remote consultation etc. In this paper we present the
design of a web based telemedicine system which can
2. TeleMedik
be used by patients, doctors as well as organizations
such as hospitals for availing themselves of healthcare
A brief account of the TeleMedik software
services over the internet. This paper also outlines
mentioned in section 1 is given below.
some areas of research for the improvement of the
existing system.
The TeleMedik software is a point-to-point
telemedicine system where different hospitals across
Key Words varied geographic locations serve as the service access
Telemedicine, healthcare, medical, store-and-forward, points. It broadly classifies different hospitals into two
videoconferencing. categories: 1) Nodal Hospital and 2) Referral Hospital.
The nodal hospitals are those, which require healthcare
services (suggestion, diagnosis etc.) from a specialist
1. Introduction doctor located at a remote place. The referral hospitals
are those which offer healthcare services to nodal
Telemedicine has been defined as the delivery of hospitals. At first a patient registers at a nodal hospital.
healthcare and the exchange of health information The data entry operator at the nodal hospital collects
across distances, including all medical activities: relevant documents from the patient and stores it in a
making diagnosis, treatment, prevention, education and local database. Doctors at the nodal hospital then verify
research [1]. Telemedicine is also defined as the use of the patient data and refer him/her to a referral hospital
telecommunication technologies to provide medical where a specialist doctor views the patient documents
information and services [2]. It is a tool for unification and sends back necessary suggestions and
of healthcare services of different locations irrespective prescriptions. This system also supports real-time
of their distances for better sharing of resources and communication between doctors at nodal center and
expertise. The ultimate goal of telemedicine is to create referral center.
a world, where no matter who you are or where you are
you get the health care you need, when you need it. Each of the nodal and referral hospitals uses a
Evolution of technologies such as interactive local database. When a nodal center refers some

112
patients to a referral center, the relevant data are copied and use the telemedicine facilities without going
to the other database. Though this system is very useful through personal verification phase.
for low bandwidth communication networks the main
disadvantage of this scheme is that it leads to
duplication of a large amount of data at different
centers. The web based telemedicine server removes
this disadvantage as it uses a centralized database
server and a different administrative schema.

3. Major features of the web based system

The major features of the web based telemedicine
system are as follows:
• Four major entities namely, Patient, Doctor,
Hospital and Kiosk build the framework for this
web based healthcare system. A description of
these entities is given in section 4.
• The patients and doctors can communicate with
each other conveniently through this system. It
also allows patients to send their medical
data/image/video/audio through the Internet.
• The system uses SSL for secured transmission of
data over the Internet, thus satisfying data
confidentiality, data integrity and authentication Figure 1: The signup procedure
requirements.
• This system allows patients to be registered over a 4.2 Referencing Module
period of time thereby giving doctors the added
facility of viewing the patient’s past ailments over This module is the most significant part of the
that period. telemedicine system, which handles the
• Provides an administrative hierarchy for the communications between the patients and the doctors
registered hospitals to help manage a large number as well as data uploading and viewing. The different
of patients at a time. entities that build up the referencing module are:
• While reporting his/her disease a patient can also
view the profiles of various doctors and patients • Patient: This system provides a patient with a
registered in the system. Thus it also acts as an varied range of options for obtaining telemedicine
information base for available healthcare facilities. services. A patient can register directly to the
website and access the e-healthcare facilities. If he
4. The design of the system does not have easy access to Internet he can still
enjoy telemedicine facilities by registering through
This system runs on a Windows 2000 Server a local hospital or a kiosk. The website provides a
platform. It uses a centralized database server for user-friendly interface for uploading and viewing
storing relevant data. The telemedicine website acts as of text/multimedia documents (Refer figures 3, 4).
the interface between the user and the database. The
system can be broadly classified into three modules: • Doctor: Under this web based telemedicine
Signup Module, Referencing Module and Security system a doctor can either operate as an
Module. A brief description about these three modules independent practitioner or he/she can be affiliated
is given below. to a hospital. Apart from viewing the current
documents of a patient, a doctor can also browse
4.1 Signup Module through his/her past ailments, which may help in
better diagnosis in some cases. After viewing the
This system supports a partly manual registration documents a doctor can issue prescription or he
procedure for different hospitals and doctors. When a may also set an appointment with the patient if he
person registers to the telemedicine website he/she is finds that the submitted data is insufficient.
also required to send some personal verification
documents by mail. Only after verification of these • Hospital: In this system a hospital can register for
documents the system administrator activates the obtaining or providing telemedicine services. Here
account. This manual verification procedure protects a hospital entity consists of several Departments.
the system from fake doctor/hospital registration. The departments consist of the doctors and the
However patients can register directly to the website patients registered under them. The system
provides each of the hospitals with one or more

113
 Hospital Administrator Accounts. Similarly each SQL Server. It also uses SSL for secured transmission
of the departments is given one or more of data. The website is hosted on a Windows 2000
Department Administrator Accounts. The main job Server platform. A few snapshots of the working
of these administrator accounts is to take care of system is given below.
respective patient queues by referring them to
specialist doctors.

• Kiosk: The kiosks act as an intermediary between

a patient and the telemedicine system. If a patient
requiring telemedicine services is not familiar with
computers and Internet or does not have easy
access to Internet, he/she may seek help from a
kiosk where the patient is registered and is referred
to a specialist doctor by the Kiosk Administrator.
When a prescription for that patient arrives it is
dispatched to the patient.

Other than these entities the Referencing Module

also consists of user disk space management sub- Figure 3: Data uploading
module. The following diagram outlines the
administrative hierarchy for this system.

Figure 4: Data viewing

6. Future areas of work

In this section we elaborate on the future areas of work
involving construction of the web based telemedicine
server.
• Distributed Database: At present our
telemedicine system has a centralized database
server. But storing all the data at a single location
Figure 2: The administrative hierarchy may lead to network congestion when a large
number of users try to access the database at the
same time. Therefore a distributed architecture is
4.3 Security Module necessary for better stability and performance [5].
It is needed to design a distributed database which
The security module takes care of the patient will be transparent to other modules of the website
privacy as well as the network security needs. It i.e. other modules will view the data to be located
prevents unauthorized viewing or modifying of patient at a single location.
data, which is a standard requirement in all • Concurrency Control: Since a large number of
telemedicine systems [4]. The telemedicine website patients and doctors access the database at the
uses SSL for secured transmission of data over the same time the data needs to be consistent.
Internet, thus satisfying data confidentiality, data Additional concurrency control measures must be
integrity and authentication requirements. taken for insert/update/deletion operations.

5. Implementations • Videoconferencing: The system currently uses a

store and forward technique for providing
The website has been built using ASP as the telemedicine services. However it has the
scripting language. HTML and JavaScript have also provisions for adding videoconferencing modules
been used for developing the system. The database for for real-time communication between patients and
the telemedicine server has been built with Microsoft doctors.

114
7. Conclusion References
The advent of modern telecommunication [1] O. S. Adewale, An internet-based telemedicine
equipments and information technology over the last system in Nigeria. International Journal of
few decades has brought about revolutionary changes
Information Management, 24 (3), 2004, 221-234.
in all spheres of life. Several researchers have tried to
build telemedicine systems using these communication [2] D. Perednia, A. Allen, Telemedicine technology
technologies in an effort to improve healthcare. In this and clinical applications. Journal of the American
paper we have described the design of an Internet
Medical Association, 273(6), 1995, 483–488.
based telemedicine system, which allows patients,
doctors and hospitals to participate in remote medical [3] X. Yu, Telemedicine and Its New Trends.
consultations. At present the system uses store and
Proceeding of the Pan Pacific Area VIP
forward technique for telemedicine services. It gives
the patients a varied range of options for using the Workshop,Sydney, Australia, 1997.
system (refer section 4.2), which is very useful in
[4] P. White, Privacy and Security Issues in
developing countries like India. This paper also
outlines possible areas of improvement for the system. Teleradiology. Seminars in Ultrasound, CT, and
The web based telemedicine system described in this MRI, 25(5), 2004, 391-395.
paper also does not have any special hardware
[5] Y. Xiang, Q. Gu, Z. Li, A Distributed Framework
requirement on the client side. However it requires
high-speed Internet connection for fast uploading and of Web-Based Telemedicine System. 16th IEEE
viewing of medical multimedia. With the advancement Symposium on Computer-Based Medical
of communication technologies such high-speed Systems, 2003, 108–113.
connections are gradually becoming cheaper and
affordable. Hence this requirement does not pose a big
problem for usage of this system. It has the potential to
extend healthcare services to the rural and urban
underserved people by better sharing of resources and
expertise among patients, generalists and specialists.

115
 THE SCOPE OF UTILISING TELEMEDICINE FOR IMPROVING THE
FUNCTIONING OF THE INDIAN PUBLIC HEALTH SYSTEM IN THE
PREVALENT SOCIO-ECONOMIC AND CULTURAL SCENARIO

Prasanta Pathak and Sanjeev Bakshi

Faculty Member, Population Studies Unit, Indian Statistical Institute, Kolkata.
# Research Scholar, Population Studies Unit, Indian Statistical Institute, Kolkata.

Health Scenario etc. was also significant contributor to the

improvement of the indicators. The morbidity and
The National Health Policy (NHP), 1983 was the mortality levels, however, are still
formulated with a goal of achiving health for all unacceptably high. Among the communicable
by 2000AD. The noteworthy initiatives under the diseases, incidence of Malaria has staged a
policy were (1) a phased and time-bound resurgence. An increasing level of insecticide-
programme for setting up a well-dispersed resistance has developed in the malarial vectors in
network of primary health care services, linked many parts of India and the incidence of the more
with extension and health education, designated in deadly Plasmodium Falciparum Malaria has risen
the context of the ground reality that elementary to about 50 per cent in the whole country. There is
health problems can be resolved by the people also no significant decline in the pool of infection
themselves, (2) inter-mediation through ‘health out of TB amongst the community and there has
volunteers’ having appropriate knowledge, simple been a distressing trend in the increase of drug
skills and requisite technologies, (3) establishment resistance to the type of infection prevailing in the
of a well-worked out referral system to ensure that country. Emergence of a new and extremely
patient load at the higher levels of the health virulent communicable disease called HIV/AIDS
system hierarchy is not needlessly burdened by since the declaration of the NHP-1983 has posed a
those who can be treated at peripheral levels and serious threat to the public health efforts as it has
(4) an integrated network of evenly spread got serious implications on economic
speciality and super-speciality services; development of the country.
encouragement of such facilities through private
investments for patients who can pay, so that the Common water borne infections like
draw on the government facilities is limited to Gastroenteritis, Cholera and some forms of Hepatitis
those entitled to free use (Ministry of Health and continue to contribute significantly to the current high
level of morbidity even though the mortality rates for
Family Welfare, 2002a). The noteworthy
these diseases have been moderate. Acute Respiratory
successes out of these initiatives over time are Infections and Asthma are also significant contributors
eradication of Small Pox and Guinea Worm. Polio to high morbidity level. With these, certain ‘life-style’
is now on the verge of eradication. Kala Azar, diseases like diabetes, cancer and cardiovascular
Leprosy and Filariasis are expected to be diseases have increased mortality. With expansion of
eliminated in the foreseeable future. the public health facilities to serve the increasing
population size, institutionalisation of antenetal, natal
In 1981, the life expectancy, the crude birth and post-natalcare of mothers and their babies has not
rate, the crude death rate and the infant mortality been satisfactory. As per the findings of the National
rate have been about 54, 34, 12 and 110 Family Health Survey (IIPS, 2000) only about 42 per
cent births have been attended by health professionals.
respectively. By 2000, these have changed to
Medically disapproved means of delivering births,
about 65, 26, 9 and 70 respectively. While the abortions and management of various sexual health
public health initiatives over the years contributed problems have resulted in high mortality and morbidity
significantly to the improvement of the health among mothers and other women. While high neo-
indicators, the outcome of several complementary natal mortality, post-neonatal mortality, infant
initiatives under the umbrella of the mortality and child mortality need serious attention for
developmental sectors, covering rural reduction, and demand for appropriate geriatric care is
development, agriculture, food production, on the rise with increasing life expectancy. Also,
sanitation, supply of drinking water, education increasing urbanisation of the country has resulted in

116
increasing burden of trauma cases due to accidents and 137, 88, 60 and 52 in Uttar Pradesh, Madhya
violent incidents. Pradesh, Jammu and Kashmir, Rajasthan, Bihar,
Punjab and Gujrat respectively. The shortfall at all
Sufficiency of Economic Input for Improving India level has been 2310. Large number of PHCs
Health Status have been lacking doctors, laboratory technicians
and pharamacists. The shortfalls of surgeons,
While with increasing population over time, obstetricians and gynaecologists, physicians and
more and more people are falling at risk of paediatricians at the CHCs have also been quite
developing some health problems, the government high (see Ministry of Health and Family Welfare,
investment or public health has not been 2002b). The shortfall of specialists at the CHCs at
increasing similarly. In fact, the investment as a all India level has been 7459. The CHCs have
percentage of Gross Domestic Product (GDP) has been lacking significantly the radiographers, the
declined from 1.3 in 1990 to 0.9 in 1999. While pharmacists and the laboratory technicians as well
the aggregate expenditure in the health sector as a and the shortfalls have been 1352, 6678 and
percentage of GDP is only 5.2 in 1999, most of 12,661 respectively at all India level. With respect
this expenditure (about 83 per cent) is out-of- to the norm of one nurse / mid-wife in a PHC and
pocket expenditure of people. Health being a State 7 nurse / mid-wife in a CHC, the shortfall of this
subject, the central budget allocation for health category of staff at all India level has been as high
over the said period has remained stagnant at 1.3 as 20,842.
per cent of total budget (Ministry of Health and
Family Welfare (MoHFW), 2002a). On the other Inadequacy in available health facilities is
hand, the financial resources being almost also reflected from the average rural population
inelastic in most of the States, the percentage of covered by a CHC and a PHC. In 2001, the former
total State Budget, which has been alLocated for has been 2 lakhs or more in 12 out of 28 States.
health, has declined from 7.0 to 5.5 on the The latter has been 30,000 or more in 15 out of
average. The result is the sad state of rural-urban the same number of States. Average radial
differential and state to state differential in health distance covered by a CHC has been more than 10
status as shown in Table 1. This wider regional km. in 19 out of 26 States before reorganisation of
disparity implies that economically backward Bihar and Madhya Pradesh. While average
sections of society in several States have number of villages covered by a CHC has been
unacceptably low access to public health services more than 100 in 20 out of 26 States, average
and hence their standard of health is grossly number of villages covered by a PHC has been
inadequate. The findings in Table 2 show more than 30 in 7 out of 26 States.
additionally that the differential in health status
exists also by different social groups. In fact, it has been acknowledged during the
formulation of the NHP-2002 (Ministry of Health
Though NHP-1983 has a thrust on making and Family Welfare, 2002a)that the existing
good the unmet need of public health services by public health infrastructure is far from
establishing public health institutions at a satisfactory. For the outdoor medical facilities in
decentralised level, a large gap in facilities existence, funding has been generally found
persists even today. Application of pre-2000 insufficient; the presence of medical and para-
government norms for provision of public health medical personnel has been found often much less
facilities to the projected population in 2000 than what has been required by prescribed norms;
shows that the estimated shortfall in the number of the availability of consumables has been found
sub-centres (SC) / Primary Health Centres (PHC) negligible frequently; the equipment in many
/ Community Health Centres (CHC) has been of public hopitals has been often obsolescent and
the order of 16 per cent. The shortage has been as unusable; and, the buildings are in a dilapidated
high as 58 per cent when disaggregated for CHCs state. In the indoor treatment facilities also, the
only. The prominent States with significant equipment has been often obsolescent; the
shortfall in infrastructure have been Uttar Pradesh, availability of the essential drugs has been
Bihar, West Bengal, Madhya Pradesh, minimal; the capacity of the facilities has been
Maharashtra, Andhra Pradesh and Orissa grossly inadequate, which leads to overcrowding
(Ministry of Health and Family Welfare 2002b). and consequently to a steep deterioriation in the
The shortfall of doctors at the PHCs, in the top 7 quality of the services.
States have been of the order of 1545, 221, 179,

117
Economic Aspects of People’s Health Seeking George, 1997) and it is true for both rural and
Behaviour urban areas. Direct costs (costs of consultation,
hospitalisation, medicines and diagnostic tests) in
It has been established through the results of totality has been found to be in the range of 69 to
NSS 52nd Round (National Sample Survey 93 per cent of the total costs (Duggal and Amin,
Organisation (NSSO), 1998) that financial 1989; Sundar, 1992 and 1994; George 1997;
problem has been the second most important Yesudian, 1990; Sujata Rao et al., 1997; NSSO,
reason for avoiding medical treatment for 24 per 1998) and of that, the share of medicines and
cent rural and 21 per cent urban patients. The consultation fees has been major. Except the
main reason behind avoiding medical treatment NSSO study, all other studies, mentioned in this
has been that about 52 per cent rural and 60 per context, tend to show that the cost of treatment per
cent urban persons with ailments have not found illness episode is higher in rural areas (Duggal and
their health problems serious. It has also been Amin, 1989; Sundar, 1992 and 1994; George,
shown that the average medical and other related 1997). NSSO study (NSSO, 1998) has established
non-medical expenditure per treated ailment has that the non-medical or indirect cost of treatment
been cheaper when availed from public medical per illness episode in rural areas is somewhat
centres (Rs. 146 for public medical centre and Rs. higher in non-government sector than public
185 for private and other centres). Average total sector. The study by Hotchkiss et al. (2000) has
expenditure per hospitalisation has also been less shown that the costs incurred on transport / food /
when availed from public hospitals (Rs. 2080 for lodging has been not more than 20 per cent of the
public hospitals and Rs. 4300 for private and other total cost of treatment to have reproductive and
hospitals). In spite of this low-priced public health child health care services. This has been
services, 81 per cent rural and 80 per cent urban substantiated by the NSSO studies also for any
patients have been going for non-government non- sort of treatment.
hospitalised treatments. Non-government
hospitalisation services have also been more
popular; for every 1000 rural and urban patients, Scope of Telemedicine in Improving the
562 rural and 569 urban patients have preferred Public Health System
non-government hospital facilities.
In last more than 20 years, the country has
In has been established through the National developed a huge public health infrastructure. The
Family Health Survey (NFHS) (2000) that questions are now being raised on its adequacy
majority of the households (about 65 per cent) and service quality. NHP-2002 attempts at solving
seek care for illnesses from private hospitals or this problem. Shortage of key health staff,
clinics. Only about 29 per cent households obsolete and inadequate equipment, minimal
normally seek care from institutions in the public supply of medicines and other supplies,
medical sector. The rating on quality of services, dilapidated or inadequate buildings and associated
measured in terms of time devoted to care, infrastructure explain partially the reasons behind
waiting time, behaviour of service providers, people’s predominant dependence on private or
maintenance of privacy and cleanliness have been non-government sector. The part that remains
found to be lower for public sector facilities both unexplained constitutes of the factors that
in the rural and the urban areas. The situation has contribute to unacceptable quality of public health
been better for the private sector facilities. services. Poor or inadequate input factors do
Inadequate public health infrastructure and poor contribute to fall in efficiency and quality, but that
quality of services at the public health facilities do not explain fully the low demand for public
are important causes behind patients’ high health care services though these are low priced
dependence on the private sector. The results of due to government undertaking of the running
NSS 42nd Round (NSSO, 1992) also have shown costs of the facilities. To have full explanation,
that private health facilities have been the sources one must first of all assess how well the current
of care for 76 per cent urban and 81 per cent rural public health system is functioning.
outpatients. The studies carried out at both
national and regional levels have indicated that Unfortunately, no comprehensive health
the proportion of patients who pay for health care system information is available at present. In
services can be quite high, ranging from 64 to 90 absence of this, it is difficult to answer questions
per cent (Sundar, 1992; Duggal and Amin, 1989; such as (1) how rationally and optimally the

118
health service facilities are located, (2) how well can be made to serve the health service seekers
the infrastructure of the facilities match with the through use of telemedicine facilities by
community needs, (3) how rationally and sufficiently trained and computer savvy junior
optimally the different categories of available staff doctor or senior nurse. Again, a public health
have been posted and utilised at health facilities of centre without adequate X-ray facility or a well-
different types, (4) how rationally and equiped laboratory facility can have tele-
economically the purchase of equipment, networking with a nearby private or non-
medicines, etc. are made and how scientifically government diagnostic centre with telemedicine
they are distributed, maintained or stored and facilities under certain agreement which would
made use of, (5) how scientifically and how help the seekers of public health services to get
patient-friendly manner the health service the tests done at subsidised rates. Also, tele-
facilities are managed, (6) how skilled and how networking with the government, the non-
accountable the health staff are so as to provide government and the private or non-government
the best possible services, and so on. hospitals at secondary and tertiary levels could be
Telemedicine has much to do in all these areas. quite economic and time-saving for those health
Development of a comprehensive information service seekers who require to consult with the
base with continuity over time and space, giving specialists at higher levels for realising the true
detailed input, process and output information by need of seeking care at those higher levels. This
each facility, is what is immediately required. can reduce to a great extent the problem arising
This will help grading performance of each due to over crowding at the secondary and the
facility on the basis of data-based monitoring and tertiary health service facilities. Also, the direct
supervision, taking prompt steps for improvement and the indirect costs of treatment will be reduced
of efficiency and service quality, facilitate better in this process. However, as mentioned earlier the
utilisation of limited available resources through current less-informed and less-knowledgable
sharing of information and available resources in health service seekers have to be assisted by
more than one facility and appropriate referral of properly trained and computer savvy junior
patients at different levels of the health system doctors or senior nurses. Also, it needs to be
hierarchy to stop overcrowding at tertiary-level ensured that the doctors and all other health staff
facilities . Tele- medicine could be useful even for are available at the public health facilities
effective public-private cooperation through throughout their duty hours. Telemedicine could
appropriate networking. It could be made use of be useful for disease surveillance also. Besides
for skill upgradation of health staff at different these, such facilities may also help the
levels by arranging appropriate computer-based underprivileged health service seekers to have
training programmes. acess to approved medical stores which would
make the prescribed medicines available at
Simultaneously with above-mentioned government-subsidised rates, lower than the
intervention on the supply side of health services, market prices.
attempts are to be made for increasing access.
With current literacy level of population (male Basically, computerised tele-transactions of all
literacy 76 per cent and female literacy 54 per types and at every level of health services will
cent as per Census 2001, and about 25 per cent bring about transperency in all functions, help
persons having education level matriculation and monitoring the flow of all kinds of input, process
above as per Census, 1991) and mental barriers in and output, increase accountability of the service
them, developed out of socio-cultural background providers at all levels and allow timely
and belief system, the seekers of health care modernisation at all levels whenever necessary.
services cannot immediately go for spontaneous Given that many public health facilities at
and direct use of the telemedicine facilities. They secondary level and below are lacking tele-
can at the best consult a remotely located communication advantages even today and
experienced doctor through telemedicine facility current teledensity in the country is 7.4 per 100
by taking help of a similarly trained person, persons (ref. http:\\www.indiastat.com), with
available beside him or her. The strategies for Chhatisgarh, Bihar, Jharkhand and West Bengal
popularising telemedicine should be through such having teledensities as low as 1.64, 1.75, 2.02 and
arrangements that minimise both direct and 2.29 respectively, the tasks ahead for successful
indirect costs. Thus, health facilities without implementation of telemedicine are substantial.
doctors or with inadequate staffing with doctors Additionally, large scale training programme is to

119
be conducted to make the health staff comfortable 8. National Sample Survey Organisation (NSSO).
with the use of computer and telemedicine-related 1992. Sarvekshana, NSS-42nd Round, 1986-
facilities and accessories. 87, Dept. of Statistics, Ministry of Planning
and Programme Implementation, Government
References of India, New Delhi.
9. _________________. 1998. Morbidity and
1. Duggal, R. and Amin, S. 1989. Cost of Health Tneatment of Ailments, NSS-52nd Round,
Care : A Household Survey in an Indian 1995-96, Dept. of Statistics, Ministry of
District, Foundation of Research in Planning and Programme Implementation,
Community Health, Bombay, India. Government of India, New Delhi.
2. George, A. (ed.). 1997. Household 10. Registrar General of India (RGI). 1999.
Health Expenditure in Two States : A Bulletin of Sample Registration System, Government
Comparative Study of Districts in of India, New Delhi.
Maharashtra and Madhya Pradesh,
Foundation for Research in Community 11. Sujata Rao, K., Ramana, G.N.V. and Murthy,
Health, Pune, India. H.V.V. 1997. Financing of Primary Health
Care in Andhra Pradesh : A Policy
3. Hotchkiss, D.R., Kanjilal, B., Sharma, Perspective, Centre for Social Services,
Suneeta, Sodani, P.R., Chakraborty, G. 2000. Administrative Staff College of India,
Household Expenditure on Reproductive and Hyderabad, India.
Child Health Care Services in Udaipur,
Rajasthan, in Financing Reproductive and 12. Sundar, R. 1992. Household Survey of
Child Health Care in Rajasthan, Indian Medical Care, Margin, June-March.
Institute of Health Management Research, 13. _____________. 1994. Household Survey of
Jaipur and Policy Project, the Futures Group Health Care Utilisation and Expenditure,
International, New Delhi.
Margin, July-September.
4. International Institute for Population Sciences
(IIPS), 2000. National Family Health Survey 14. Yesudian, C.A.K. 1990. A Study on Health
(NFHS-2) 1998-99, Mumbai, India. Services Utilisation and Expenditure,
Hinduja-Harvard Program for Community
5. Ministry of Health and Family Health in “Organising and Financing Health
Welfare(MoHFW), 2002a. National Health
Care Services for the Middle Income Group in
Policy (NHP)-2002, Government of India,
New Delhi. Bombay”, 15-16th January, 1990, Hinduja
National Hospital, Bombay.
6. ________________ 2002b. Bulletin of Rural
Health Statistics in India, March, 2002,
Government of India, New Delhi.
7. ________________. 2001. Health Information
of India 2000-01, Government of India, New
Delhi.

120
Table 1 : Rural-Urban Differential and State-to-State Differential in Health Status
Area Population IMR per Mortality of Maternal Malaria Leprosy
Below Poverty 1000 Live Children Aged Mortality Per Positive Cases per
Line (%) Births Below 5 Years Lakh Married Cases in the 10,000
(MoHFW, (RGI, 1999) (Per 1000) Women, Year 2000 Population
2001) (NFHS – 2) (RGI, 2000) (in thousand)
(1) (2) (3) (4) (5) (6) (7)
India 26.1 70 94.9 408 2200 3.7
Rural 27.09 75 103.7 -- -- --
Urban 23.62 44 63.1 -- -- --
Better Performing States
Kerala 12.72 14 18.8 87 5.1 0.9
Maharashtra 25.02 48 58.1 135 138 3.1
Tamil Nadu 21.12 52 63.3 79 56 4.1
Low Performing States
Orissa 47.15 97 104.4 498 483 7.0
Bihar 42.60 63 105.1 707 132 18.8
Rajasthan 15.28 81 114.9 607 53 0.8
Uttar
31.15 84 122.5 707 99 4.3
Pradesh
Madhya
37.43 90 137.6 498 528 3.8
Pradesh

N.B. From NHP-2002

Table 2 : Differentials in Health Status Among social Groups

Group Infant Mortality Under 5 Mortality Per cent Children Under-weight
Scheduled Castes 83 119 54
Scheduled Tribes 84 126 56
Other Disadvantaged 76 103 47
Others 62 83 41
Over all 70 95 47

N.B. From NHP-2002

121
 WEB BASED PATIENT RECORD SYSTEM
Y. Madhusudhana Reddy, Dandapat S., and Harshal B. Nemade.
Department of Electronics & Communication Engineering, Indian Institute of Technology, Guwahati.
North Guwahati-781039, Assam, India.
ymreddy@iitg.ernet.in, samaren@iitg.ernet.in, and harshal@iitg.ernet.in

Abstract patient is unable to respond to the questions, it will be

very helpful to have reliable information in the form of
In this paper, we have presented an information system electronic patient records. A web based electronic
for health care services. This system, named as Health patient record system can be tailored to be used as a
Care Information System (HCIS), is an electronic referral system. The information provided by the
referral system. HCIS can be used by healthcare patient consists of a brief problem description and
providers to collect and manage their patients’ health contact information. If the first contacted specialist is
information as well as contact information. Patients can unable to provide treatment, then the same information
also collect and manage their consultation history. A needs to be supplied again to another specialist. Patient
messaging system is included as a communication tool often need to obtain a copy of all the prior medical
between various users. A mathematical model has been information from his primary healthcare provider. So it
presented here to analyze the referral traffic, which in is easy to forward the patient’s electronic records to
turn optimize the performance of the overall system. another healthcare provider, rather than moving the
patient physically.
Key Words
We have developed an electronic referral
Web based patient record system, referral management system, called the Health Care Information System
system, e-medicine, e-health, Health Care Information (HCIS). In this system, users are categorized as
System (HCIS), medical informatics. doctors, patients, and staff. User authentication rights
are restricted depending on the category of the user.
Doctors have more authentication rights compared to
1. Introduction
other users. Using this information system, patients can
request self referrals for their on going problem by
Nowadays, internet is growing as a major and
providing the detailed description the problem. HCIS
cheap communication tool. Compared to other
can be used by the healthcare providers to collect and
communication tools, internet has more features in the
manage their patients’ health information and contact
form of data storage and data retrieval. Any time, one
information. Patients can also manage their
can access the web-information from any where around
consultation history. This system consists of a
the globe. These features will be useful in the field of
messaging system that facilitates the referral process
health care [1-2], where patients face lack of access to
between patients and doctors. This messaging system
their complete health information. Now patients are
can be used as general communication tool to provide
managing brief and update versions of their health
healthcare related information.
information in the form of paper records, which do not
describe detailed medical history of the patient. A physician referral is a request initiated by
Patients’ past medical records are important input in patients’ healthcare providers to see a specialist for a
decision making during the treatment. Without particular type of problem. In this context, physician
sufficient medical information, there will be a delay in referral is usually for diagnosis and treatment of a
diagnosis process, and there is a risk of inefficient particular type of physical problem. Generally,
treatment. Internet can help educate patients about seasonal effects of nature result in different health
healthcare and available treatment options. problems. It is expected that a particular type of
Wang et al [3] have developed a web based general health problem may arise during a season in a
Personnel Health Record (PHR) system. This system is year. In that period more users will request for similar
developed to integrate patients’ health information as referral services, and emergency cases may not be
well as contact information in the form of electronic adequately addressed during that time. Reliability of
data records. PHR includes patients’ medical history in HCIS will be affected due to increase in referral
detail, consultation history, past surgeries, and contact requests. A mathematical model [4-5] has been
information. Electronic patient records can play a key presented here to analyze and predict the referral
role at the time of emergency compared to the requests traffic, which in turn help optimize the system
conventional paper records, which fail to describe performance.
detail medical history. In emergency cases, when the

122
2. Methodology information about patient transfer protocols (PTP),
OPD timings, and doctor’s details etc.
Health care information system (HCIS) is a web based
self referral system used for collecting and managing • Mathematical Model: Generally, seasonal effects
patients’ medical information. Patients first need to of nature will bring special health problems. During
create a user account in HCIS to utilize the resources of that period more users will request for similar type of
the HCIS. Each user must be authenticated by user referral service.
name and password. The main features of HCIS are,
doctors do not require any software knowledge to Fig. 3 describes the number of patients visiting for
manage HCIS, and doctors can update their patients’ healthcare per month. This data is collected from IIT
records from their client system only. There is no need Guwahati dispensary. This data shows the number of
to walk to the server.
Use
• System Architecture: r’s Ac
req HTT IP Yes ces
Client uest P Per- s
for mitte
refe
de
Client Web MS rral No nie
Server ASP. Access serv d
(IIS) NET data ice
Client base Authen No
icated

Client Yes IIS

Fig. 1 System Architecture

Launch
ASP.NE
The architecture of HCIS is shown in Fig. 1. It is a
web based application running on Microsoft’s
Windows XP operating system, Internet Information
Req
Server (IIS), Active Server Pages (ASP.NET) HTTP
uest
technology, and MS Access database. ASP.NET is the St No Ser Yes Scri
No
ed ver pts
Microsoft’s latest technology in server side scripting refe ati Scr exec
environment that has been used to create interactive rral
c ipts uted
web pages and build powerful web applications. pa
serv
ASP.NET is chosen for its inbuilt security features and ice/
rich web application framework. Whenever IIS Alte Yes
receives a web page request from client, it will execute rna HTT P
the server side scripts, and it will send only the script tive Scrip MS Access
result. User cannot view the scripts. Web pages are trea t (Electronic
composed of HTML, ASP.NET and VB script tme result records)
programming languages. nt
Fig. 2 describes the User-Server Interface. All the ASP.NET
users are authenticated by user name and password.
Doctors have more authentication rights to view/edit
the documents related to health information. Patients’ Fig. 2 User-Server Interface
authentication rights are restricted to edit their own
health records. All the users authentication rights are patients visiting the dispensary between the months of
integrated with their user account type, for example, June-2003 to December-2003. It is clear that more
patients will not get any option to update their lab patients visited from August-03 to October-03. Peaks
records. are appeals occurrence of some seasonal diseases.
This system includes a feature which will provide Referral requests traffic is similar to web traffic. A
necessary information for first-aid. A first-aid database mathematical model has been presented here to predict
is created for generally/most frequently occurred the referral traffic level. For the analysis, we have
diseases and their prevention methods. This database is adapted a model proposed by Mark et al [4]. This paper
kept for general viewing. Patients do not need to login presents the web traffic patterns from Europe and Asia
to avail this facility, so referral requests traffic will be at a time scale of 300 seconds. These represent the user
somewhat reduced. This system also includes general request process submitted to the official web site for
the 1998 Olympic Games in Nagano, Japan. The web

123
traffic for HCIS is expected to be similar to this web 1
− lo g F ( x ) : (lo g x − µ ) 2 + lo g (lo g x − µ )
traffic. The peaks will represent sudden increase in 2σ 2
.
load due to the seasonal diseases.
That is the log tail distribution behaves as a quadratic
A stationary time series { Z n } [4] is said to satisfy
function of lo g x , and in this sense, X follows a heavy
an order (p, q) autoregressive moving average model,
tailed distribution.
denoted by ARMA (p, q), if it can be represented
We next consider the dependence structure of the
as
Z n − ∅ 1 Z n −1 − .... − ∅ p Z n − p =∈n −θ1 ∈n −1 − .... − θ q ∈n − q sequence { X n } . Let
lo g X n = µ + Y n
where p and q are respectively the orders of Auto
Regressive (AR) and Moving Average (MA) processes. for each n , with {Yn } represented by the AR(1) process
Y n = ∈ n + ∅ 1Y n − 1 ,

where ∅ 1 is a real parameter and { ∈ 1 , ∈ 2 . . . . } is a

sequence of independent and identically distributed
normal random variables with mean 0 variance σ ∈ .
2

Since { X } is assumed to be stationary sequence,

n

{ Y } must also be stationary sequence. Hence, for all

n

n, Y n is also a random variable with mean zero and

σ2
variance y that must satisfy
σ y2 = σ x2 + ∅ 12 σ y2
,
And we have
2
X n = e ( u +∈1 +∅ 1∈n −1 +∅1 ∈n − 2 + ...)
.
The k moment of X is given by
th
Fig. 3. Referral requests traffic n

k u + k 2σ 2
/2
E[X k
] = E [ e k u + k Yn ] = e y
n ,
Let { X n } denote the patient request process submitted And we can then derive the autocorrelation function for
{ X n}
to the web server, with X n denote the number of to be
1
th
requests of the patients that arrive in the n time ρ : σ 2
∅ k

, as k → ∞ .
k σ 2 y 1
e y
− 1
period. We suppose { X n } is a stationary sequence, and Hence, { X }
continues to be a process with short
n

denoted by X = X n the generic random variable that range dependence. We can therefore take the log of the
follows the marginal distribution F(.). data and apply standard time series analysis to obtain
the parameters for the model, using higher order
ARMA process if the data possesses more complex
Let us consider the traffic as bursts traffic resulting dependent structures.
from seasonal health effects. Let X follow a lognormal
th
. The n
µ +σ Z
distribution, i.e., X = e moment of X is Following the above approach for the traffic, we
given by have to find the parameters µ and ∅ 1 to find out the
2 2 model [4], from this model we can construct a sample
E ( X n ) = enµ E (enσ Z ) = enµ + n σ /2
trace by first generating a sample path from ARMA
and in particular, series. Note that the generated sample path contains the
2 2
some of the key characteristics of the web traffic, we
E ( X ) = e u +σ σ
, Var ( X ) = E ( X )( e − 1) .
/2
have considered. So we can predict the referral traffic
Let φ ( x ) and ∅ ( x ) respectively denote the of HCIS.
distributive function and the density function of Z.
3 Results and Discussion
Define φ ( x ) : = 1 − φ ( x ) , we then have
F ( x) = P( X ≥ x) = P(µ + σ Z ≥ log x) = φ ( zx ) , We have implemented some of these ideas in an
initial version of HCIS. Web pages of HCIS are
where designed using HTML programming, ASP.NET, and
z x := (log x − µ ) / σ . VB script programming. Home page of HCIS is shown
in Fig. 4. Each user is authenticated by user name and
It is easy to verify that φ ( z ) : ∅ ( z ) / z when z → ∞ .
password.

Hence, when x → ∞ , we have z x → ∞ and Patient needs to login to avail the facilities and
services of HCIS. Patients can request for prescription
from their login web page, as shown in Fig. 5. The

124
menu along the left side of the screen is used to Before and/or during the time of visit, doctors can read
navigate to the various categories which patients can the documentation provided by the patient and will
access. They can, request for appointment, find out quickly understand the need for treatment as well as
roster duties, find out doctors’ details, and can request patient’s health condition, and can search for the
for referral service. Patients’ can access their lab tests possible treatment options.
results. Patients can document their problem
description in the message box provided. The
documentation of the medical problem consists of a
brief description of the problem, how and when it
started, whether it is going better or worse, and the side
effects associated with the problem. Once the problem
has been documented, patients can choose to submit a
referral service.
Doctors can view the patients’ details from their
login. The web page relating to referral process is
shown in Fig. 6. The menu along the left side of the
screen is used to navigate to the various categories
which doctors can access. They can, view referral
requests, post messages
Fig. 6 HCIS View patient details

The database is developed with MS Access. Later

this can be upgraded to SQL server database, for
storing high volume of data like X-ray images etc.
Compression techniques will be applied for efficient
storage of images. A messaging system is included as a
communication tool between doctors to make them
aware of patient’s health information. Online medical
transcription can be implemented with HCIS to make
the system more user friendly.

4. Conclusion
In this work, we present the design and
preliminary implementation of an information system
for health care services. More features will be added to
Fig. 4 Home page of HCIS this system by incorporating online medical
transcription, compression techniques and security
relating to roster duties, and appointments. They can features etc. A mathematical model has been proposed
update first-aid database, and can prescribe medicines. to predict the referral traffic.
Doctors view patient’s details in the order they
requested.
References:
[1] Aziz Boxwala, Omolola Ogunyemi, and Qing
Zeng, Medical Computing (HST.950J), Harvard-
MIT Division of Health Sciences and
Technology, http://ocw.mit.edu/OcwWeb/Health-
Sciences-and-Technology/HST-950JMedical-
ComputingSpring2003/CourseHome/index.htm.
[2] Steven Locke, Bryan Bergeron, Jeffrey Blander,
and Daniel Sands, Health Sciences and
Technology (HST.921), Information Technology
in the Health Care System of the Future, Open
Courseware, MIT,
http://ocw.mit.edu/OcwWeb/Health-Sciences-
and-Technology/HST-921Information-
Technology-in-the-Health-Care-System-of-the-
FutureSpring2003/CourseHome/index.htm.
Fig. 5 HCIS request for prescription web page

125
[3] Maisie Wang, Christopher Lau, Frederick A. [5] Hung-ying Tyan, Ching-Fong Su, Guanghui He,
Matsen, and Yongmin Kim, Personal Health and Xun Su, Application-Driven Internet Traffic
Information Management System and its Analysis, IEEE Conference on
Telecommunications, GLOBECOM,
Application in Referral Management, IEEE
Volume: 7, Dec. 2003.
Transactions on Information Technology in
Biomedicine, vol. 8, September 2004. [6] Thomas Karagiannis, Mart Molle, and Michalis
Faloutsos, Long-Range Dependence: Ten Years
[4] Mark S. Squillante, David D. Yao, and Li Zhang, of Internet Traffic Modelling, IEEE Transactions
Web Traffic Modeling and Web Server on Internet Computing, vol. 8, September 2004.
Performance Analysis, Proc. of the 38th [7] D. J. Poulish, Methods and metrics for developing
Conference on Decision & Control high quality patient monitoring system software,
Phoenix, Arizona, USA, December 1999. Proceedings of 3rd annual Symposium on
Computer- Based Medical Systems, June 1990.

126
Medical Imaging & Computer Aided Diagnosis
 A DIAGNOSTIC TOOL FOR EARLY PREDICTIONS OF DEMYELINATION
USING M R IMAGES

Mehta S. B.,1 Chaudhury Santanu,2 Bhattacharyya A.,2 Jena A.3

1
Indian Institute of Technology Hauz khas, Delhi – 110 016.
2
Delhi College of Engineering Bawana road, Delhi – 110 042.
3
MRI Center, Rajiv Gandhi Cancer Hospital, Rohini, Delhi 110085.

Abstract changes, etc.) at an early stage in the MRI is an acute

problem. Most of the existing approaches for
diagnostic tools use learning or prior subjective
This paper proposes a diagnostic tool for early knowledge obtained from experts. Moreno et al. [1]
prediction of demyelination (white matter changes) proposed to solve the brain maturation problem by
using a temporal sequence of MR images. Exploiting using several neural classifiers. Udupa et al. [2]
the pattern of changes in gray level distribution in the proposed a method using dual echo and fast spin echo
MR images, the proposed diagnostic tool would try to MR imagery for quantification of multiple sclerosis
indicate early warnings. Appearance of white matter (white matter changes due to treatment protocol). The
changes due to demyelination in brain tumor cases for method uses fuzzy connectedness principle to estimate
patients, given radiation therapy, vary from case to the volume of diseased tissue. Wong et al. [3] proposed
case. Further, there exist inherent impreciseness in the learning of fuzzy rules by genetic algorithms. In
white matter change patterns. These characteristics addition, several standalone commercially developed
make use of fuzzy features well suited for describing tools are also used for offline analysis and visualization
image based temporal changes. Correlation between to improve diagnosis. In this paper, we have proposed
these temporal patterns and actual onset of use of genetic algorithms for supervised learning of
demyelination can be captured by fuzzy rules because diagnostic knowledge in the form of fuzzy rules using
of the inherent uncertainty associated with changes in labeled data patterns. Unlike other approaches, this
gray level pattern in the image and occurrence of the technique uses multiple temporally ordered MR images
disease. The tool uses genetic algorithms for learning and a set of novel features to address different disease
fuzzy rules using examples provided by medical scenario.
professionals. The genetic algorithm based machine-
learning tool generates an optimized rule set to indicate
positive, negative or doubtful cases of demyelination. 2. Overview
Experimental results have established effectiveness of We have used a fuzzy rule based classifier as the
the proposed tool. basic diagnostic tool.. Appearance of white matter
changes due to radiation varies from patient to patient.
1. Introduction Further, there exist inherent impreciseness in the white
matter change patterns. These characteristics make use
White matter changes are harmful side effect of of fuzzy features well suited for describing image
radiation therapy given for the treatment of brain tumor based temporal patterns. Correlation between these
cases. This causes permanent changes in the white temporal patterns and actual onset of white matter
matter tissues resulting in memory loss, loss of change can be captured by fuzzy rules because of the
balance, lack of co-ordination etc. These changes can inherent uncertainty associated with changes in gray
be controlled, provided these are diagnosed at an early level pattern in the image and occurrence of the
stage. The white matter changes can be due to other disease. We use genetic algorithms for learning fuzzy
reasons also. In this paper we have proposed a rules using examples provided by medical
diagnostic tool for prediction of white matter changes professionals. The genetic algorithm based machine-
due to radiation therapy. The tool uses a sequence of learning tool generates an optimized rule set to indicate
temporally ordered MR images. It is based on soft- positive, negative or doubtful cases of white matter
computing techniques and attempts early detection of changes.
white matter changes before they are distinctly visible
in the MR images. The relative efficacy of the tool The training data is a collection of m tuples,
appears to be acceptable for the test cases considered. X A = ( x A1 , x A2 , .....x An ), where A = 1, 2, ....., m
The MRI has a unique advantage of providing
multi spectral images of tissues with a variety of and n is the number of attributes in the feature space.
contrasts based on parameters like T1, T2, PD, etc. But Each training sample is associated with one of the
detecting diseases (like tumor, infection, white matter following output conditions: positive, negative, and

127
doubtful. Here, in the diagnostic tool, seven attributes MR image data [4]. These partitions have been
( n = 7 ) have been considered. These features have obtained based upon frequency distribution as well
been extracted from the sequence of MR images. slope of the histogram. These clusters can provide
Antecedent part of the fuzzy rules makes use of these natural partitions of the pixels corresponding to
attributes. Each of these input attributes correspond to different regions in the MR image, although these
three fuzzy sets: low, medium, high We have used regions may not strictly correspond to logical
triangular membership functions for these fuzzy sets. segments. Typically images are over-segmented. The
Using the labeled tuples and these fuzzy sets we have variation in the number of clusters, number of pixels in
used genetic algorithm based supervised learning for each cluster and the position of the corresponding
generating fuzzy rules. In the next section, we describe prototype on histogram of MR images obtained before
the features used. and after radiation can be used as attributes in the
diagnostic tool.
3. Features from Temporally Ordered Based on the clusters obtained, we use a fuzzy
MR images rule based system to segment each image into three
segments white matter, gray matter and CSF [7].
The image data set used pertains to the condition Although these segments for post-radiated cases are not
before and after radiation. One of the biggest problems strictly correct, we also use change in the size of these
in diseased cases is that due to the tumor a large segments as input attributes. In the case of radiation,
portion of the brain is affected. Moreover sometimes the affected WM pixels have a shift in their intensity
surgery is also performed on these cases, which makes value. Such pixels tend to get classified as GM or CSF.
analysis of images more difficult. Taking the full brain
image for white matter tissue analysis is not 4. Generation of Fuzzy Rules
appropriate in these cases. Here a small portion of the
brain image containing 1875 pixels (25 X 75) not Use of genetic algorithm for rule generator is
affected by tumor or surgery is manually cut from the discussed [5,6]. The same method is used here except
2D MR axial image for analysis. This part of the MR for the change in fuzzy rule definition and training data
image mainly comprises of White Matter (WM), Gray set. In the present case, there are seven inputs for each
Matter (GM) and CSF tissue. The same axial slice rule and one output. The output is crisp and it indicates
position is selected for both before radiation and after change in white matter with three possible values:
radiation images. This small image is analyzed to Positive (P), Negative (N), or Doubtful (D). The inputs
extract features based on histogram and gray value are Proto_No, Bin3, Bin4, Bin8, Bin9, Pix_WM,
distribution pattern in the image for diagnostic Pix_RM. These inputs measure the change in the
prediction. corresponding parameters. Changes in each input is
fuzzified using three triangular functions (Left triangle,
We have considered MR images with 255 gray Rt. Triangular, and triangular) into three fuzzy sets
levels. Histograms have been obtained using 255 bins. Low, Mid, High. Training data set was obtained from
We have sub-sampled the histogram to obtain 16 experts’ feedback. Table 1 shows some of the rules
equally spaced sample. Each sample value is obtained generated with the help of genetic learning. These rules
by applying a simple averaging filter. We refer to these are applied for the early detection of white matter
samples as bin values. Figure 1 shows images and changes. We infer the diagnosis corresponding to the
histogram before radiations and after radiations. The highest membership value.
bin value in the lower intensity region usually
decreases after the radiation and bin values increase
after radiation in the higher intensity region. The 5. Evaluation of the Diagnostic tool
difference between the bin reference value (before
radiation) and bin value after radiation is fuzzified to We have done experimental evaluation of our
derive a qualitative descriptor for this variation. diagnostic tool for patients who have undergone
In MR brain images, the intensity variation at radiation therapy after surgery in tumor cases. We have
boundaries in normal cases (between WM or GM or considered examination of 3 images or study per
CSF) is gradual and not sharp. Similarly in diseased patient. First study corresponds to pre-radiated
cases (white matter changes), the intensity variations in condition; other two studies depict post-radiated
the boundary region of white matter are gradual with conditions. These images were obtained from Rajiv
respect to diseased area, and healthy tissue of white Gandhi Cancer Institute and Research Center, New
matter, gray matter or CSF. The histogram is flat over Delhi. The report given by doctor on study2, and
the diseased area of white matter with no distinct peak. study3 of patient is compared with our results (white
So identification of different regions and distinct matter changes Y, N, or D) for assessing accuracy of
boundary within white matter region in MR images on our method. We first consider an example case study to
the basis of gray level distribution is difficult. A illustrate effectiveness of our scheme. This patient is a
connectionist approach has been used to generate case of left temporoparietal glioma whose three MR
probable pixel partitions from the 1-D histogram of studies were carried out, one before RT and two follow

128
up studies after RT. One such slice is shown in figure the accuracy of the tool. The tool can also be added
2. One side of the brain is affected by disease, and with more attributes in future to cover white matter
surgery is also performed. We manually cut a small changes due to other reasons
window of the image 25X75 size out of full for
changed white matter analysis. Same 25X75 window is 7. References
cut in the three slices (Figure 3). Features as discussed
above are extracted and fuzzified from the small 1. Moreno L, Pinerio JD, Sanchez JL, Manas S,
window MR image data (number of prototypes, bin Merino J, Acosta L and Hamilton A. “Brain
values, and number of pixels, etc.) Fuzzy rules are used maturation estimation using neural classifier”.
for interpretation of changes in the MR image. Features IEEE Transactions on Biomedical Engineering,
extracted are shown in table 3 along with conclusion Vol. 42, No. 4, April 1995.
drawn (last column). Image of slices and report of 2. Udupa JK, Wei L, Samarasekera S, Miki Y,
radiologist is indicated in figure 3. After analyzing M.A.Van Buchem and Grossman RI. “Multiple
study2, radiologist report a doubtful case of white Sclerosis-connectedness Principles”. IEEE
matter change. However, fuzzy rules infer a confirmed Transactionson Medical Imaging, Vol. 16, No. 5,
case of white matter change. After analyzing study 3 Oct 1997.
radiologists confirm changes in white matter and our
fuzzy rules corroborate this conclusion. This case study 3. Wong CC, Chun M and Lin NS. “Extraction of
clearly shows that our diagnostic tool can do correct Fuzzy control rules using genetic algorithm” Proc.
early prediction and can be of assistance to CFSA/IFIS/SOFT ‘95’ Fuzzy theory and
radiologists. Applications.

We have validated our results with the help of 4. Vinod VV, Chaudhury S, Mukherjee J and Ghose
leave-one out test on the collected examples. Prediction S. “A Connectionist approach for clustering with
results have been shown in table 2. It is found that the applications in image analysis”. IEEE Transactions
diagnostic tool successfully predicts the early white on Systems, Man, and Cybernetics. March 1994,
matter changes in most of the cases. 24 (3) pp.365-384.
5. Shi Y, Eberhart R and Chen Y. “Implementation
6. Conclusion of evolutionary Fuzzy Systems”. IEEE
transactions on Fuzzy Systems, Vol. 7, No. 2,
We proposed a demyelination diagnostic tool April 1999.
here to predict the early white matter changes due to
the radiation therapy given to the tumor cases. The tool 6. Buchi H, Nakashima and Murata T. “Three-
generates fuzzy rules based on the training pattern, objective genetics-based machine learning for
which is applied to new cases to predict white matter linguistic rule extraction”. Information Sciences,
changes. These training patterns are generated from the 136 (2001) 109-133.
labeled temporal study of patient before radiation and 7. S.B. Mehta, Soft-Computing for MR image
after radiation. The labeling is done with the help of Analysis, Ph.D Thesis, D.U, 2004
MRI expert. The training data set is generated from the
small number of tumor cases, but more data from the
tumor cases given radiation can be added to improve

0.5 0.5

0.4 0.4

0.3 0.3

0.2 0.2

0.1 0.1

0 0

Figure 1(a) Figure 1(b)

Figure 1.(a) Before Radiation (b) After Radiation
IF Bin3 IS medium AND Pix_WM IS high AND Pix_RM IS high THEN pos
IF Bin4 IS high THEN pos
IF Proto_No IS low AND Bin9 IS low THEN neg
IF Pix_WM IS low THEN neg
IF Pix_RM IS low THEN neg
IF Proto_No IS medium AND Bin9 IS medium AND Pix_RM IS high THEN doubt
IF Bin4 IS low THEN neg
IF Proto_No IS low AND Bin4 IS medium AND Pix_WM IS high AND Pix_RM IS high THEN doubt
IF Bin9 IS low AND Pix_RM IS high THEN doubt

Table 1: Some Fuzzy Rules generated by the GA

129
 Normal White
matter

White
matter
Edema

Surgery
Performed

Figure 2: A T2 weighted MR slice of tumor case before radiation therapy

MR report
No WM MR report
Change MR report WM
WM Change Changes
doubtful present

MR image before MR image after 3 MR image after 9

Radiation months of the
months of the
radiation
radiation

Figure 3: Portion (25X75) of T2 weighted MR image of tumor patient before radiation after radiation

Protto_Num Bin_Val48 Bin_Val64 Bin_Val112 Bin_Val128 WM_Num G_C_Num Changes

1 3 0.4 32 12 17 502 421
H L H M H H H Y
2 5 1.2 33 38 21 509 242
H L H H H H H Y

Table 2: shows output results of test case

No change in WM Change in WM WM change doubtful

Number of actual pattern 7 5 9

Number of correctly predicted 7 5 6

Percentage correctly predicted 100% 100% 66%

Table 3: Results of leave one out test

130
 DETECTION OF BLOOD VESSELS IN RETINAL IMAGES USING MODEL
BASED APPROACH

M. Ravi Kumar, S. Dandapat

Department of Electronics&communication,
Indian Institute of Technology Guwahati,
North Guwahati-781039, Assam, India.
e-mail: mrk@iitg.ernet.in, samaren@iitg.ernet.in

Abstract hypertension, and arteriosclerosis. Veins and arteries

have many observable features. In this paper, an
In this paper the problem of extracting the blood automated method to locate and outline blood vessels
vessels in retinal fundus images is investigated. It in images of the ocular fundus images is proposed.
involves the developing of feature/pattern extraction With this tool, eye care specialists can potentially
algorithm to detect blood vessels. These features will screen larger populations for vessel abnormalities.
help the physicians for detecting the disorders that Precise measurements may be more easily recorded,
appear in the retinal image data. An automated method for instance, for evaluation of treatment or for clinical
to locate and outline the blood vessels is described in study. Observations based upon such a tool would also
images of the ocular fundus. Detection method follows be more systematically reproducible.
model-based approach. The gray-level profile of the
cross section of a blood vessel can be approximated by The segmentation methods have three different
a Gaussian shaped curve. The concept of matched filter categories: Window-based method, classifier-based
detection of signals is used to detect piecewise linear method and tracking-based method. Window based
segments of retinal vessels in ocular fundus images. methods are the conventional edge detection methods
like sobel edge detector, prewitt operator and Laplacian
Key Words of Gaussian operator (LoG). The second category is
tracking based methods as proposed in [2], [3]. The
Retinal imaging, fundus image, retinal vessel, matched
drawback with this method is the seed points which
filter, blood vessel segmentation, feature extraction.
should be given by the user. This is a difficult task.
This method does not detect bifurcations, forks and
1. Introduction
joints precisely. The third method is the classifier based
Fundus retina image is the primary source to method. In the present work, we have used a classifier-
investigate the pathological changes in the retina. This based method. In this approach, blood vessels are
image gives the information about the localization of detected using a set of two-dimensional matched
features in it. Blood vessels, Optic disc, Fovea, filters.
Lesions, and Exudates etc. are the features in a retina
image. These features can provide information about 2. Methodology
pathological changes caused by some eye diseases such
as glaucoma, which may lead to loss of vision. The In the detection of blood vessels, Matched filter is
retinal vessels constitute the vascular network of a designed using model based approach. The filter is
retina. Blood vessels are the main objects in the retina designed according to the intensity profile model of the
considered by the medical professionals for diagnosis vessels. When the image is convolved with the filters
and monitoring the patient’s illness. Direct observation with appropriate direction, the response hits maximum
of retinal vessels in fundus image is difficult. It is value, whenever there is a match of signal passed
important to detect blood vessels in retinal images to through this filter. The concept of one-dimensional
detect the changes and abnormal structures during matched filter [4] can be extended to two-dimensional
clinical therapy. matched filters. The matched filter is modelled as
Gaussian filter. When the neighbourhood at each pixel
of retinal image is convolved with the matched filter
Edge detection has an important role in many the result is also Gaussian. It is observed that the blood
image-processing applications. But the conventional vessels’ cross-section profile follows Gaussian model
edge detection algorithms are not suitable for detection [5]. So the designed matched filter is a Gaussian filter.
of the features in the retinal images. Sobel and Prewitt In this work, twelve kernels are designed to detect the
operators are not able to detect the sharp edges lying in vessels that are lying in different directions. When the
different directions. The Laplacian operators are Gaussian filter is convolved with the vessels segment,
susceptible to direction rotation [1]. Blood vessel the convolution result hits maximum value when the
appearance is an important indicator for many vessel cross-section is in the appropriate direction of
pathological conditions, including diabetes, the matched filter. At each target pixel of the selected

131
segment position, the maximum response of the filtered
output is retained, while ignoring the remaining eleven
convolution results.
c 2w

3. Blood vessels segmentation

If we consider a cross section of a vessel L
segment, the vessel intensity profile will assume a
Gaussian shape as shown in fig. 1. The matched filter
is designed by exploiting the continuous properties of Figure 2: a segment cross-section
the blood vessels and their Gaussian intensity profiles.
The vessels may be considered as piecewise linear
segments, i.e. they have fixed width and fixed direction 4. Implementation of filter
for chosen length of the segment as shown in fig. 2.
Here, the direction of the vessel assumed to be aligned
along the vertical axis. Segmentation allows region The filter is implemented as follows. P=[x y] is a
based processing and noise can be suppressed. A vessel discrete grid in the two dimensional matched filter. θ
may be oriented at any angle between 0 to π radians. is the direction of the chosen vessel segment. The
For the detection, the filter is fitted to vessel segments matched filter will have its peak response only when it
instead of fitting a single profile. A typical segment is is aligned at an angle θ ± π / 2 . θ i is orientation of the
showed in fig.2, where C is the center of the vessel, W ith filter which is matched to a vessel at an angle θ i .
is width of the vessel from C, and L is length of the Angle resolution is 15 ° . Therefore it is required to
vessel segment in pixels. It is assumed that the filters design 12 filters. As the kernel is applied at the centre
are centered about the origin. Therefore the kernel will of the vessel segment, the filter needs to transform its
have maximum value at the origin, so that the mean co-ordinates to the center, and subsequently to rotate
value is set to be zero. about the origin. The two-dimensional rotation matrix
is given by [5]

 cos θ i − sin θ i 
R =  sin θ co s θ i  (2)
 i

The corresponding point in the rotated coordinated

P i = [s t] =[x y] × R . It is
T
system is given by
assumed that the kernel is centered at [0 0], as the filter
mask is applied at the centre of the segment. And the
designed filters are matched low pass filters of
Gaussian model. So, the kernel has its maximum value
at center. The mean value of the kernel is zero. A
Gaussian curve has infinite long double-sided trails in
the either side of its center. So we have to truncate the
Gaussian spread to a finite value s = ± 3σ . The
neighbour hood N can be defined as following.

{ (s , t )
Figure 1: vessel intensity profile
N= s ≤ 3σ , t ≤ L / 2 } (3)

A vessel segment is represented as the following

The corresponding weights in ith kernel is given by
mathematical equation [5]

K i ( x, y ) = − exp( − s 2 / 2σ 2 ) ∀ P i∈N (4)

K ( x, y ) = − exp(− x / 2σ ) For y ≤ L / 2 (1)
2 2

In the design of the filter it is assumed that the mean

where L is length of the segment. The maximum value value of the kernel is zero. But the filter is truncated to
of x is w from the center of the blood vessel, σ is the an optimal width. Therefore for every ith kernel, the
mean.
spread in the Gaussian intensity profile in a cross–
section of a vessel segment.

132
 Figure 4: Frequency response of 45° kernel

Figure 3: Frequency response of 0° kernel

5. Results and Discussions
value m i is subtracted from equation (3) in order to
keep the mean value of the kernel as zero. Thus the The method is tested with 50 images. Twelve
convolution mask used in the present algorithm is matched filter kernels are generated for every 15° angle
given as resolution. The chosen values of σ and L are 2 and 9
respectively. Frequency responses of two kernels are
ki' ( x, y ) = ki ( x, y ) − mi ∀ P i∈ N (4) showed in fig. 2, and fig. 3. It is observed that the
The mean value of the ith kernel is given by kernel rotated according to the direction, and the
resulting filters are low-pass filters. The size of the
mi = ∑ pi ∈N
K i ( x, y ) / A (5) kernel is 17 × 17. We can also select the higher size,
but it might lead to computational complexity.
A is total number of points in neighbourhood N.

A designed filter frequency response for 0° angle

rotation is shown in the fig. 3. It is observed that the
filter is resuming the low pass filter. The chosen values
of L and σ are 9 and 2 respectively by observing the
responses for different images. The filter has maximum
energy at low frequencies. The filter response is
showed for normalized frequency. Frequency response
for 45° is showed in fig. 4. The magnitude is decreasing
as the frequency in increased. As the Gaussian function
will have infinite spread, the filter is truncated to an
optimal width to keep the filter size. This truncation Figure 7: Graylevel mean Figure 8:Sobel edge
leads to conversion of an IIR filter to a FIR filter. Filter detec-
images for designed kernels for two angles are showed Filtered image ted image
in fig. 5 and fig. 6, the images are rotated according the
filter rotation.

Figure 9: Response using Figure 10:Response of

Figure 5&6: Structuring Images for 00 , 450 kernels
Prewitt Edge detector LoG operator

133
 6. Conclusions

The method is investigated and it is showing promising

results. It is a simple and reliable method. The
advantage of matched filter detection is the
improvement in results compared to the conventional
edge detection operators. This not only simplifies the
Figure 11: Response using Figure12:Response of diagnostic tool but also enable the physician to make a
des- confident judgement about the existence of eye
Canny edge detector cribed Matching filter diseases. The observation of fig. 12 and fig. 14 reveals
that we, indeed, get some response due to edges of
other bright objects such as optic disc, lesions, and
elliptical capture effect due to the sensing devise. The
described method is the part of work on detection of
blood vessels in retinal images using robust method. In
the future work we are working to minimise the above
said unwanted responses and to minimise the residual
errors due to model based approach. Applications of
above designed filters enhance individual vessel
segments in the image. Blood vessels can act as
Figure 13: Gray level fun- Figure 14: vessels landmarks for localizing the optic nerve, the fovea
detection (central vision area), and lesions.
dus image using matching filter
Acknowledgements

The blood vessels are detected by convolving the I would like to thank Clemson university
fundus image with each of the twelve filters at every authorities for providing database of images and
pixel, assuming it is the centre element. Before this, the clinical details.
colour fundus images are converted to grey level
images. This grey level image was mean filtered with References
3 × 3 mask in order to reduce the spurious noise
[1] Rafael C. Gonzalez, Richard E. Woods Digital
generated by the captured camera. Mean filtered image
Image Processing. (New Delhi: Pearson
is showed in fig. 7. At every target pixel the maximum
Education, 2002).
response of the twelve kernels is retained, while
ignoring the remaining responses. Response due to [2] Y. Sun, Automated identification of vessel
sobel edge detector and Prewitt detectors are not giving contours in coronary arteriograms by an adaptive
the best result, and not resuming blood vessels tracking algorithm. IEEE transactions on medical
imaging, vol. 8, march 1989, 78-88.
appearance as shown in fig. 8 and fig. 9 respectively.
If we observe the response of Laplacian of Gaussian [3] O. Chutatape, Liu Zheng, and S.M. Krishnan,
operator in fig. 10, it is not producing better Retinal blood vessel detection and tracking by
localization of the blood vessels. The best edge matched Gaussian and kalman filters. A tutorial
detector, Canny method as shown in fig.11, is also review, Proc. 20th IEEE Conf. on Engineering in
failing to produce the good results. Considerable Medicine and Biology Society, Piscataway, NJ,
1998, 3144-3149.
improvement can be achieved using the matching filter.
From the observation in fig. 12, it is apparent that the [4] S. Haykin, Communication Systems. (New Delhi:
matched filter is working well and giving the precise Wiley Eastern, 1979).
results. Another source image and its vessels detected [5] S. Chaudhuri, S. Chatterjee, N. Katz, M. Nelson,
image are showed in fig. 13, 14 respectively. and M. Goldbaum, Detection of blood vessels in
The method is implemented on PENTIUM-4 based retinal images using two-dimensional matched
filters. IEEE transactions on medical imaging,
computer system under Windows XP environment. The
vol.8, 1989, 263-269.
required algorithm was implemented using MATLAB
software tool. Fundus images are collected from [6] http://www.parl.clemson.edu/stare/nerve/stare-
Clemson University [6], which are made available to images.tar
the public of research interest. All reference images are
coloured indexed images of size 150 × 130 pixels.

134
 A UNIFIED IMAGE PROCESSING BASED MODEL FOR EARLY CANCER
DETECTION FROM MAMMOGRAM IMAGES
Rajib Mahapatra, P.K. Biswas
Dept of E&ECE, IIT Kharagpur
rajib_mahapatra@yahoo.com, pkb@ece.iitkgp.ernet.in

Abstract
There are different types of breast abnormalities
This paper presents a unified model for detecting which doctors observe as suspicious case in
different types of breast abnormalities from digital x- mammogram. Their edges, size, shape and texture help
ray mammogram. Different image processing the doctor to judge whether it is cancer or not. Masses
techniques are applied for pre-processing work. appear as dense regions of varying sizes and properties
Wavelet based enhancement technique has the ability and can be characterized as circumscribed (Fig. 1a &
to enhance the contrasts of mammographic features Fig. 1b), ill defined (Fig. 1c). Spiculated masses/lesions
while reducing the enhancement of noise also. Next, have a special feature. Its edges are distributed in all
multiscale segmentation is used, as histograms of 3600 direction from a central point (Fig. 1d).
mammogram images are not bimodal. At last textural Circumscribed masses are more uniform, circular,
features are used to characterize different types of although still irregular. On the other hand,
tissue in digital mammogram. Classification is microcalcifications (Fig. 1e) appear as small bright
performed by a top down approach using binary arbitrarily shaped regions/ spots on the large variety of
classifier tree. Experimental results using MIMS image breast texture background. These abnormalities are
database have shown that the algorithm is capable of difficult to detect. The great variability of the mass
detecting breast abnormalities and judgement from appearance along with the other abnormalities in digital
radiologists are very encouraging. mammograms is the main obstacle of building a
unified cancer detection method.
Key Word :
Digital mammogram, wavelets, multiresolution, feature
analysis, binary classifier tree.

1. Introduction
Breast cancer is the most common cancer among
women in the developed countries and the second most Fig. 1a Fig. 1b Fig. 1c
common in the developing countries. In India, the
incidence reported is 18 to 25 per 1,00,000 from
different urban centers and only 8.6 per 1,00,000 from
the rural cancer registry [1]. Indian Council of Medical
Research data shows that breast cancer is the most
common cancer to affect Indian females in
metropolitan cities [2]. Detection of the disease in early
stage can reduce rate of mortality. X-Ray mammogram
and ultrasound imaging are two main imaging test for Fig. 1d Fig. 1e
breast cancer diagnosis [3]. Among these two
techniques, mammogram is widely used technique for In this paper we propose a unified model, which
early detection. During several years a lot of researches will be effective to detect all types of breast
have been carried out on image processing based abnormalities from digital mammogram in an
computerized automated detection of breast cancer interactive way. The paper is organized as follows: In
[4,5]. The objective of this research is not to construct Section 2, a brief description of the overall diagnostic
a high-tech radiologist or cancer expert. The algorithm system is presented. In Section 3, wavelet based
is intended to serve as a second opinion for the contrast enhancement module is given. Section 4
radiologist. We focus our effort on the detection of presents multiscale segmentation technique. In Section
cancerous region from mammograms. Our goal is to 5, feature extraction method is described, a brief
locate suspicious regions in the mammogram for more description of classifier is given. Classification
detail examination by the attending physicians. Thus, algorithm is also shown in this section. Section 6 gives
development of tools for automatic classification of results and discussion. At last, section 7 draws
suspicious areas in mammograms is very important as conclusion.
it acts as an assisting tool to a radiologist to improve
the efficiency of screening process and avoid
unnecessary biopsies.

131
2. Proposed Computer Aided Diagnosis this context, considerable successes have already been
System achieved in the development of wavelet based
enhancement technique [9,10].
Basic block diagram of proposed system is given
in Fig. 2. The cancerous region has subtle contrast Here, nonlinear techniques for image
compared to normal breast structure making early enhancement are applied within the context of
malignancy detection a difficult task. So, image multiresolution representations. Enhancement
enhancement is an important step. Conventional techniques are applied only to the band pass sub-band
enhancement techniques such as contrast stretching, images of a multiresolution representation. 2-D dyadic
histogram equalization etc. lead to overall contrast wavelet decomposition at level L partitions images into
enhancement but generally fail to enhance local three sub-bands y0L, y1L, y2L. The high frequency sub-
feature. Here we use multiresolution enhancement bands are processed further with the objective of
method based on wavelet transform [6, 7]. The second enhancing the reconstructed image. Amplification of
step is image segmentation. Mammographic image is high frequency coefficients is a possible approach for
not bimodal. So single threshold generally gives poor enhancement. But this has the disadvantage that it
result. So we use multiscale segmentation method enhances the noise at the same time. To overcome this
based on histogram moment [8]. As a result, image is problem Laine et. al. [9] proposed a mapping function
partitioned in different clusters. Next step is image as shown in Fig. 3. Here, Gji and Tji are local gain and
analysis. In this part, different textural feature are threshold in the j-th sub-band at level i respectively.
extracted and here, our analysis is concentrated on each ŷji
partition resulted from multiscale segmentation. Last
part is classification that gives the ultimate decision. ŷji=yji + Tji(Gji-1) iff yij>T
-T ŷji= yji – Tji(Gji-1) iff yji<-T
3. Enhancement T yji ŷji=Gji *yji Otherwise

Multiscale/Multiresolution wavelet processing is

one of the most promising approaches for Fig. 3
mammographic feature enhancement. Generally, Microcalcification corresponds to high frequency
mammographic image is low contrast image. It is often component of image spectra. Suppressing low
needed to enhance weak signal features against a frequency subband gives microcalcification output in
background of relatively high noise. Our goal is to reconstructed image, built from only high frequency
achieve a balance between the enhancement of subtle subband (y1L & y2L).
image detail and avoidance of noise amplification. In

INPUT Different Sample Images

IMAGE

Image Enhancement
(Dyadic Wavelet)

Textural
Segmentation Feature Wavelet for
Extraction Microcalcification
Detection
False Color
Composit for
Classifier Design
Visualization

Enhanced Output Output Output Output

Output Image Image Image Image Image

Normal
Mass or Spiculated lession Detection Microcalcification
Mammogram
Detection
Detection

Fig. 2 Overall Block Diagram

136
4. Segmentation 1 4 *
p(i, j , d ) = ∑ p (i, j, d ,θ * 45)
4 θ =1
Segmentation of the possible cancerous regions is
really a challenging work. It obviously minimizes We extracted four features from GLCM with
errors in the further high level processing such as d=1.These features are F1.Energy, F2.Entropy,
classification. Mammograms have low signal to noise F3.Inertia, F4.Homogeneity.A fifth feature F5-
ratio (low contrast) and complicated structured Histogram Gradient Orientation is extracted which is
background. In a mammogram, the object of effective specially for spiculated lesion. It appears as
segmentation is the tumour (cancerous region), which an irregular tumor center surrounded by spicules that
generally has a higher gray scale value than the radiate in all direction where normal tissue radiate in
remainder of the image. Mammogram contains one direction i.e. nipple to chest wall. The differences
calcifications, dense tissue, and various kinds of noise between two is shown in Fig. 4. Let ∂Sij be some
etc., in addition to the normal tissues and possible neighborhood of pixel position (i, j). Also let histij be
tumors (targets). The normal tissues and possible the histogram of θ within ∂Sij, hence histij(n) = # of
tumours also contain some singular pixels. Moreover pixels in ∂Sij that have gradient orientations ∈ (-
the histogram is not usually bimodal. Hence, many π/2+nπ/256, -π/2+(n+1)π/256 ) where n = 0; 1; : : : ;
pixels may be misclassified using single threshold 255.
detection method on original mammograms. On the 1 255
other hand, we do not actually know the probability hist (i, j ) = ∑ histij (n) be the average of
256 n =0
distribution function (PDF) of any class as well as the a
priori probability of every class. Also, the probability histij.
distributions often vary greatly in different images so Now feature value standard deviation of gradient
that we cannot get a general distribution to describe all orientation histogram in ∂Sij :
images. Hence, it also creates problem in using the 1 255
Bayes classifier for mammogram image segmentation. σ hist (i , j ) = ∑
255 n = 0
( histij ( n ) − hist (i , j )) 2
One of the solutions is multiscale segmentation. Here
we use multisacle segmentation algorithm developed
by Anguh and Silva, 1997[11]. After extracting features classifier is designed. Here
our feature vector dimension is five and numbers of
5. Features Extraction and Classification classes are three [M-Class for mass, B-Class for
background, S-Class for spiculated lesion]. We use
Analysis of different partioned region after multiscale binary classifier tree (BCT) for classification which is
segmentation is very important step for this model. given in Fig. 5.
Here, our analysis concentrates on enhanced image
referencing each segmented partition individually, as Classification Algorithm Design:
suspicious cell may be present in each of the partitions. • First, we select window of size 30x30 pixels and
For this purpose, a set of features is extracted. This move the window in raster scan fashion in
approach addresses the difficulty to determine the overlapping way on enhanced image.
neighborhood size for feature extraction to characterize • In each window, we calculate the feature values
the object, which may appear in different sizes. Here (Energy, Entropy, Inertia, Homogeneity and
we take 30×30 pixels window which is large enough to Histogram gradient orientation).
fit entire mass or spiculated lesion. One of the most • Classifier is already trained by sample images.
well-known textural analysis method gray level co- Now, from this feature vector classifier express
occurrence matrix (GLCM) is used for textural feature whether the particular window belongs to cancer
extraction. The entry of the normalized GLCM matrix class or non-cancer (benign) class.
p(i, j) gives us the probability of two pixels having a
gray level i and a gray level j located at a distance d • For detail analysis and reducing time complexity,
we move the overlapping window only on
and a direction θ. Feature parameters are calculated
segmented region (where probability of masses be
from a symmetric normalized GLCM matrix.
present is high) instead of whole image.
C (i, j )
p (i, j ) = Normal Mammogram
Number of pixels

∑ C (i, j )
i, j Spiculated
In general, co-occurrence matrices are calculated for Lesion
four directions. A new matrix is formed as the average
of these matrices that is used for feature
extraction. So the extracted features will be rotation Gradient Direction
invariant at least for 45° steps of rotation. The final co-
occurrence matrix
Fig 4:Edge direction for Normal
formed will be
mammogram and Spiculated lesion

137
 F3=-6.00

F5=10. F2=40.00

F2=80.00 F4=0.30 B-class

F1=0.0020

S-class B-class M-class B-class

M-class M-class

Fig 10b, 11b & 12b give different classification output

Fig 5:Binary Classifier Tree
from original images (Fig 10a, 11a, & 12a).

6. Results and Discussion 7. Conclusion

Our algorithms are implemented in visual C++ based An interactive GUI based software has been developed
GUI environment. Experiment is conducted on one set for detecting different type of breast abnormities from
of data (mammogram images) and BCT is trained by digital mammogram. For this purpose, different
that set. After training another set is used for validation features are extracted and classification is done by
that classify mammograms. Fig 6 to Fig 7 represents simple binary classifier tree. Here, our features
different steps of image enhancement result. Fig. 8b dimension are five and class number are three. Further
and Fig. 9b express output of the multiscale stage, processing more features and neural network
segmentation result for original images Fig. 8a and Fig. based classifier may give better accuracy of results. As
9a respectively. the cancer cells are varying sizes in dimension, analysis
on different resolution (coarsest to finer) will be
applied on digital mammogram to improve the result.

References
[1] http://www.medivisionindia.com/cancers/
[2] http://www.tribuneindia.com/2004/
[3] www.cancer.org/docroot/CRI/content/
[4] K.Woods and K. Bowyer, “A general view of
Fig 6 and Fig 7:Enhancement Results detection algorithms,” Proceedings of the 3rd
International Workshop on Digital Mammography,
June 9-12 1996, Chicago, U.S.A., pp. 385-390.
[5] M. L. Giger, F. -F. Yin, K. Doi, C. E. Metz, R. A.
Schmidt, and C. J. Vyborny, “Investigation of methods
for the computerized detection and analysis of
mammographic masses,” Proceedings of SPIE, February
6 - 8 1990, Bellingham, Washington, pp. 183-184.
[6] Mallat S and Zong S,”Characterization of signal from
multiscale edges”IEEE trans. pattern anal. and mach.
intell. 14(7), July,1992, pp.710-732.
Fig 8a.Org. mammogram Fig 8b Segmented output [7] A F Laine, S Sculer, J Fan, and W Huda,
“Mammographic Feature Enhancement by Multiscale
Analysis”, IEEE transaction on medical imaging, 13(4),
December, 1994, pp.725-740.
[8] P Sallaropoulos, Phy in Medicine and biology, 2003
[9] A. Leine and S. Song, “Multiscale wavelet
representation for mammographic feature analysis” in
Proc. SPIE.conf.math. Method in Med. Imaging, San
Diego, CA. July,1992.
[10] A. Leine and S. Schuler, “Hexagonal wavelet
processing for digital mammography,” in Med. Img.,
Fig 9a.Org. mammogram Fig 9b Segmented output Newport Beach, CA, Feb. 1993.
[11] Anguh and Silva,” Multiscale segmentation and
enhancement in mammograms” Computer Graphics and
Image Processing, Proceedings., X Brazilian
Symposium on, IEEE, 14-17 Oct.1997, pp.136 – 139.

138
 LESION SEGMENTATION IN MAMMOGRAMS BY MEAN
SHIFT ALGORITHM
Nagaraju Odala_† and Ajoy Kumar Ray_†
_Department of Electronics & Electrical Communication Engg
Indian Institute of Technology Kharagpur, India.

†R&D Group,
Avisere Inc, Tucson,
Arizona, USA.
Email: odala.nagaraju@avisere.com

Abstract for lesion segmentation based on mean shift algorithm.

The proposed segmentation algorithm begins with a
The goal of seeded lesion segmentation is to separate seed point within the suspected lesion region. Many
suspected masses from surrounding tissue as current computerized mass detection schemes first
effectively as possible. Segmentation of lesion regions employ an initial detection algorithm, which returns
in a mammographic image is not easy due to the low locations that are used as seed points for the
contrast and the fuzzy nature of the transition of a segmentation algorithm. The mean-shift algorithm is a
malignant lesion from its core region to surrounding simple non-parametric procedure for estimating density
tissues. in this paper, we present a method for lesion gradients. Iteration of mean shift gives rise to natural
segmentation based on mean shift algorithm, a simple clustering algorithm. An Objective Measure was
nonparametric procedure for estimating density developed to differentiate a target T (lesion) from its
gradients. Iteration of mean shift gives rise to natural background B (Tissue). The Objective measure finds
clustering algorithm. An Objective Measure was the target to background contrast ratio using variance.
developed to differentiate a target T (lesion) from its The performance of proposed method is compared
background B (Tissue). The performance of proposed against RGI method.
method is compared against RGI method. Key Words :
cancer lesion, segmentation, mean-shift clustering. II. LESION SEGMENTATION

I. INTRODUCTION The lesion tissues are usually brighter than its

surrounding tissues and possess uniform density inside
Image segmentation plays an important role in the the area. The malignant lesion area possesses an
interpretation and understanding of Mammographic approximately uniform and more or less compact
images. Several studies has been made on the nature of circular shape of varying size [Lai et al [5] ]. Quite
segmentation of digital mammographic images, often they are observed to possess fuzzy edges. Given a
particularly on region of interest (ROI) images sub image or region-of-interest (ROI) of dimension n
consisting of malignant lesion. Numerous techniques by m containing the suspect lesion, the set of
have been developed to segment lesions from coordinates in the ROI is represented by
surrounding tissues in digital mammograms. Petrick et
al. [1] employed density-weighted contrast I = {(i, j) : i = 1, 2, . . . , n and j = 1, 2, . . . ,m} (1)
enhancement (DWCE) segmentation algorithm to
extract lesions from their surrounding tissues. Li et al. The function describing the pixel gray levels of this sub
[3] used Markov random fields to classify the different image is given by f(i, j) where i, j 2 I. The values of f(i,
regions in a mammogram based on texture. Despite the j), for this work, are bound between zero and one, with
difficulty and importance of this essential step of a zero representing black and a one representing white.
segmentation in a computerized mass-detection The pixel values for all images were normalized so as
schemes, there have been fewer attempts to analyze the to assume values within this range by dividing by the
performance of these segmentation algorithms alone. maximum pixel value. Lesions tend to be compact,
On the other hand, many investigations have been meaning that their shapes are typically convex. To
reported on the collective analysis of all the incorporate this knowledge into the creation of the
components in a mammogram image. Kupinski et al partitions, the original image is multiplied by a
[4] proposed two methods for segmenting lesions in function, called the constraint function that suppresses
digital mammograms: a radial gradient index (RGI) distant pixel values. For this study we chose to use an
based algorithm and a probabilistic algorithm. These isotropic Gaussian function centered on the seed point
techniques are seeded segmentation algorithms. In their location (µi, µj) with a fixed variance _2 c as the
work, RGI is a measure of the average proportion of constraint function. The function resulting from the
the gradients directed radially outward. Because of multiplication of the original ROI with the constraint
inherent nature of RGI, it gives more emphasis to function is given by
round shape lesions. In this paper, we present a method

139
h(i, j) = f(i, j)N(i, j; µi, µj , σ2c ) (2) mean shift is close to zero. The mean shift algorithm is
tool needed for feature space analysis. The uni-
where N(i, j; µi, µj , σ2c ) is circular normal modality condition, assumed during derivation of Eq 4,
distribution centered at (µi, µj) with a variance σ2c . can be relaxed and extended to multimodal conditions,
by randomly choosing the initial location of the search
A. Clustering based on non-parametric density- window. The algorithm then converges to the closest
estimation There are several nonparametric methods high-density region [7].
available for probability density estimation: histogram,
naive method, the nearest neighbor method, and kernel III. RESULTS & DISCUSSIONS
estimation. The kernel estimation method is one of the
most popular techniques used in estimating density. The width σ2c of the constraint function in (3) was
Given a set of n data points {xi}i=1,...,n in a d- determined based on knowledge of lesions and was not
dimensional Euclidian space Rd , the multivariate statistically determined. A value of 50 (pixel width)
kernel density estimator with kernel _ and window was empirically determined to work well for our
radius (bandwidth) hn is defined as follows [6] purposes. For all our experimentation r = 0.05 (kernel
width) is considered. The initial location of the search
1 n 1  x − xi  window in the feature space is randomly chosen. To
Pn(x) = ∑ Φ  ensure that the search starts close to a high-density
n i −1 Vn  h n  region several location candidates are examined. The
(3) random sampling is performed in the image domain
and a few, ( about M = 20 ), pixels are chosen.
After density estimation we identify candidate-clusters
by using gradient ascent (hill-climbing) to pinpoint A. Objective Measure for Segmentation
local maxima of the density pn(x) . Specifically, the k
nearest neighbors of every point is determined, The basic idea proposed in this paper involves
whereupon each point is linked to the point of highest developing an objective measure to differentiate a
density among these neighbors (possibly itself). Upon target T (lesion) from its backgroundB (Tissue) within
iteration, this procedure ends up assigning each point to the mammographic image. The target T has a greater
a nearby density-maximum, thus carving up the data density within the mammogram, thus having higher
set in compact and dense clumps [7]. mean gray level intensity component compared to the
surrounding background B. A good objective measure
B. Density Gradient Estimation and the Mean Shift should aim to yield high value at the point where the
Method Since its introduction by Fukunaga and contrast between target T and background B is high
Hostler, the mean shift method has been extensively [2]. The measure is initially computed by determining
exploited and applied in low-level computer vision the difference between mean gray values in the target
tasks [7][8][9] for its ease and efficiency. Applying the and background areas as
mean shift leads to the steepest ascent with a varying
step size that is the magnitude of the gradient. δµ = µT - µB (5)

Assuming that the probability density function p(x) of In addition, the objective measure should at the same
thepdimensional feature vectors x is multimodal and time yield low values with the spread of gray scales in
also assuming that a small sphere Sx of radius r, the target area compared with the background area.
centered on x contains the feature vectors y such that This reduction can be achieved by the ratio of the gray
||y - x|| _ r. The expected value of the vector z = y - x, level variances as
given x and Sx may be derived as
σT
δσ = (6)
r 2 ∆p( x ) σB
[ ]
E x x ∈ Sx − x =
p + 2 p( x )
(4)
The resulting target to background contrast ratio using
variance (Rv) can now be computed as
Thus, the mean shift vector, which is the vector of
difference between the local mean and the center of the
window, is proportional to the gradient of the  δµ 
Rv =   (7)
probability density at x. The proportionality factor is  δσ 
 
reciprocal to p(x). This is beneficial when the highest
density region of the probability density function is
This effective segmentation measure will lead to a
sought. Such region corresponds to large p(x), i.e.
large value of Rv where the contrast between target T
small mean shifts. On the other hand, low-density
and background B is high.
regions correspond to large mean shifts. The shifts are
always in the direction of probability density
maximum, which describes the mode. At the mode the

140
Since the values of f(i, j) are bound between zero and method compares favorably well with the one obtained
one, effectively the lesions will lie in the range 60 to using the RGI Index.
100 of maximum gray value. Fig (2) shows the
segmentation results; column (a) shows region of In order to quantify the performance differences
interest images, column (b) shows segmentation results between the two different segmentation methods, an
obtained by the proposed method, column(c) shows overlap measure has been used by us. An overlap
segmentation results by RGI index. From the results it measure O was calculated using the set returned from
may be observed that the performances of our proposed the proposed segmentation algorithm

141
 CO-OCCURRENCE BASED CLUSTERING FOR MIXED BIO-MEDICAL
DATASETS
Lipika Dey* Amir Ahmad^ Vipul Goel* Rajat Mangla*
*Department of Mathematics, Indian Institute of Technology, Delhi, New Delhi-110016.
^Solid State Physics Laboratory, Timarpur, Delhi.

of gene sequences from patients with malignant tumors

Abstract to sub-group the patients into three categories
depending on their chances of survival – high, medium
This paper presents an intentional clustering or low. Clustering techniques are also applied to group
technique based on K-means paradigm that is suitable genes according to their functional similarities.
for handling mixed data sets with both numeric and
categorical features. Co-occurrence based analysis is Traditionally, clustering algorithms used Euclidean
used for defining new distance measures between distance measures to judge the similarity between two
categorical values. We have shown how our proposed elements [3]. Though this measure is very appropriate
cost function and modified center concepts enable for clustering data sets in which all attributes are
intentional clustering of bio-medical data sets. We have numeric valued, the Euclidean distance measure fails to
presented results obtained with our method for a capture the similarity of data elements when attributes
number of –bio medical data sets and compared these are nominal valued or mixed. Most of the medical data
results with other well known methods. sets are mixed in nature. Some data sets like breast-
cancer data sets or gene sequence data sets are pure
Keywords categorical in nature. Thus the problem of clustering
has taken a new dimension with its enhanced
medical data analysis, mixed data set, Cost function, applicability in the domain of medical informatics.
Significance of attributes, Distance measure.
In this paper, we have proposed a distance
measure which can define similarity between elements
1. Introduction in a mixed data set very effectively. The measure also
works well for pure numeric or pure categorical data
With increasing automation in the health-care sector,
sets. We have established its effectiveness through
the bio-medical data repository has grown by leaps and
experimentations over different data sets. The results
bounds. This has facilitated the process of knowledge-
reported here have been derived through different
discovery – which aims at extracting hidden, but
implementations of the k-means clustering algorithm
potentially useful information embedded in the data
for intentional clustering. Another unique aspect of this
[1]. A large variety of machine-learning algorithms is
work is the re-definition of the cluster centers. These
employed to explore the relationships among the tuples
are defined in a way that helps in a complete
and characterize the nature of relationships that exist
visualization of the clusters and thereby study their
between them. Clustering refers to unsupervised
characteristics effectively.
learning algorithms, in which the aim is to discover
inherent partitions of a given set of data elements into
2. Clustering medical data – a review
homogeneous groups called clusters[2].
Lemke and Muller[1] has presented an
Along with the growth of the medical data analysis of the heart data set using Group Method of
repository, clustering techniques have gained Data Handling. Using all the 76 attributes of heart data
popularity because of their capability to create patients, they have reported about 80% accuracy in
homogeneous subgroups within a defined group. Let us identifying heart patients. Since medical data is usually
look at its use through some examples. Diagnosis of mixed in nature, hierarchical clustering techniques
diseases is an important and difficult task in medicine. have been very popular with bioinformatics research
Detecting a disease from several factors or symptoms groups and it has been applied for various applications.
is a many-layered problem that also may lead to false Li and Biswas [4] proposed the SBAC algorithm based
assumptions with often unpredictable effects. Again, on agglomerative clustering. Their method achieves
diagnosis of a disease does not imply that all patients around 75% accuracy with the heart data set. [4]
will have identical characteristics and therefore reports around 74% accuracy on the same data set
undergo the same treatment. Identification of sub- using ECOWEB, which is another hierarchical
groups with the help of intentional clustering is gaining clustering algorithm.
popularity among bioinformatics researchers to
characterize homogeneous sub-groups and determine The disadvantage with hierarchical clustering
effective medication for each sub-group. We have algorithms is that they are not scalable to very large
shown how intentional clustering can be applied to databases because of their high computational cost.
categorize heart patients and breast-cancer patients into Partitional algorithms like k-means on the other hand
sub-categories. Work has been reported on clustering scale well to large data sets. K-means and fuzzy k-

140
means algorithms have been applied to cluster yeast n mr
gene expression data which are pure numeric[5,6,7]. ζ= ∑ϑ(d ,C )
i=1
i j where ϑ(di,Cj)= ∑ (w (d
t=1
t it
r
− C jt
r
))
2

Traditionally k-means algorithm uses Euclidean

distance measure in its objective function, which can +
handle only numeric data sets. Thus this was not mc
suitable for studying patient data bases like heart or
breast_cancer data sets [8] or gene sequences. K-modes
∑ ( Ω(d itc , Ccjt ))
t=1
2
(3.1)

is a variation of k-means algorithm for clustering where wt defines the significance of the tth numeric
categorical data. But since it uses the mode only to attribute, which is computed from the data set.
represent cluster centers there is loss of information. In
[9] Huang proposed a cost function which can handle 4. Computing distance between two categorical
mixed data set and applied it for analyzing the heart values - δ(r,s)
data set. We will discuss this function here since our
function is motivated by this. The use of associations in deriving patterns from
grouped data is a well-established approach in data
Huang’s cost function for a mixed data set with n mining. Similarity between elements for a categorical
data objects and m attributes of which mr attributes are data set can also be based on frequently co-occurring
numeric, and mc are categorical, is defined as items [10,11,12,13,14].
n
ζ= ∑ϑ(d ,C )
i=1
i j (2.1) We denote the distance between two categorical
attribute values r and s of an attribute Ai as δ(r,s). Let
X be a data set containing n elements, each described
where Cj is the closest cluster center for data object di with k categorical attributes. Let ξj = {1,2,…….m} be
and ϑ(di ,Cj) is computed as the entire set of attribute values for another categorical
mr mc
attribute Aj. δ(r,s) is computed as follows:
ϑ(di ,Cj)= ∑
t=1
(ditr – Cjtr )2 + γj ∑ δ( d
t=1
c
it , Cjtc) (2.2)
m−1
ditr are values of numeric attributes of di and ditc are
values of categorical attributes for data object di. Here
δ(r,s)= ∑ dijrs / m-1 (4.1)
t =1
Cj=(Cj1, Cj2,…,Cjm) represents the central values of
attributes for cluster j. Cjtc represents the most common where dijrs is a function of two co-occurrence
value (mode) for categorical attributes t in cluster j. Cjtr measures defined as follows:
is given by the mean of numeric attribute values in
cluster j. To compute the distance between two objects, dijrs = Pir (w) + Pis (~w) (4.2)
the distance between two categorical values δ(p,q) has w denotes a set of values of attribute Aj. w is that subset
to be computed. Huang assumes δ(p,q)=0 for p=q and of attribute values of Aj, with which the value r co-
δ(p,q)=1 for p≠q for all categorical attributes. γj is a occurs maximally and the value s co-occurs maximally
user-defined weight for categorical attributes for with the complement of w. Equation 4.1 shows that
cluster j. computation of δ(r,s) takes into account co-occurrence
measures of r and s with all other attributes. Obviously
an attribute Ai plays a significant role in clustering
3. Shortcomings of Huang’s cost function objects into different groups, provided any pair of its
and our proposition attribute values are well separated i.e. have a high
value of δ. While computing, each value of Aj is
The above cost function has few shortcomings. checked for its co-occurrence with r and is inserted
1. Mode of a categorical attribute does not capture into w, if the a priori probability of occurrence of the
the information about the entire cluster value r of Ai is greater than or equal to the a priori
appropriately. probability of value s of Ai occurring with it. Quite
obviously a value cannot belong to both w and its
2. Binary distance value of 0 and 1 between any two complement also. Since each value is considered
categorical values is not correct. exactly once, there cannot be a set w that can yield a
3. Significance of all numeric attributes are taken as higher probability value for both the quantities. Thus
1, which may not be true. The significance of a the maximum value for δ(r,s) can be computed in
categorical attribute, γj, is a user defined parameter linear time.
and so can lead to inaccurate clustering results. For numeric attributes, we use the Euclidean
To overcome the shortcomings of Huang’s cost distance between two numeric values to compute
function, we present a new cost function as: object similarities and also distance between an object
and a cluster center. However, to find the significance
of an attribute, we first discretize the numeric attributes
and compute δ(r,s), where r and s are two intervals, as

143
earlier. If an attribute has n intervals, then δ(r,s) Fig. 6.1 presents a visualization for the dissimilarity
averaged over the nC2 pairs gives the significance of matrix for the clustered heart data. Obviously the
the attribute. clusters are well separated. Figure 6.1(b) and (c) show
the cluster centers obtained with further clustering of
5. Modified Cluster Center the heart patients – i.e. the bottom right quadrant of
6.1(a). All the attributes have higher values for cluster
The usual definition of a cluster center with mean 1, and is correct.
and mode does not characterize a cluster correctly.
Hence we define central value for ath categorical
attribute for a cluster c using the distribution of its
values within the cluster as

1 / N c {( N c a ,1 , ..., N c a , p , N c a ,h ),
( N c b ,1 , N c b , 2,.. , N c b ,m ),.., ( N c j ,1 , N c j , 2,.. , N c )} (a)
j .n
350 500
where Nc is the number of data objects in cluster c. Nca,p 300 400

is the number of data objects with ath atrribute value p

250
200
300

in cluster c, h is the number of distinct attribute values 150 200

100
for ath attribute, and there are j attributes. This 50
100

representation of a cluster represents the quality of a 0

1 3 5 7 9 11 13 15
0
1 15 29 43 57 71 85 99
cluster more effectively since it captures the overall
(b) (c)
distribution of all values in a cluster.
While clustering, the distance between any object Attributes Cluater1 Cluster2
and an existing cluster center is computed as Cholesterol- 250(avg) varying
Age – >50(avg) <=50
mr mc Resting blood 150-180 120
ϑ(di ,Cj)= ∑
t=1
(wt(ditr – Cjtr ))2 + ∑ (Ω( d
t=1
c
it , Cjtc))2 Pressure
Max heart_rate 100(avg) 170(avg)

where Ω( ditc, Cjtc) denotes the distance between

categorical attribute values in the center and one in the Fig 6.1 Sub clusters of heart patients – characteristics of
two types of patients are very different
object and is computed using equation 4.2.
Gene sequences - Gene sequences are purely
categorical. The set contains 3000 sequences belonging
6. Experiments to three categories – intron/exon, exon/intron and none.
In this section we present the results of clustering Figure 6.2 shows cluster centers for two clusters –
for three data sets all of which are obtained from cluster 1 contains intron/exon sequences while cluster 2
http://www.sgi.com/tech/mlc/db). The heart – disease contains intron/exon sequences. The two are distinct.
data set consists of 303 patients instances defined by 13 300

features. Table 6.1 shows that our method works with A 250

85% accuracy. Table 6.2 provides comparative analysis C

200

150

of our method with others. G 100

50
T 0

Table 6.1 Cluster recovery for Heart disease data -50

1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58

set with our proposed algorithm 50

45
40
35

Cluster No. Normal Heart Patient 30

25
20

1 126 19 15
10
5

2 24 101 0
-5 1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58

Table 6.2 Comparative study for Heart Data set Figure 6.2: (top) – exon/intron – showing predominance
Clustering results for different algorithms of sequence CAG. (bottom) – intron/exon –
predominance of sequence ATA or CTA (G missing).
Algorithm Number of data objects
in Desired Cluster Breast_cancer - It can be seen from Table 6.3 that
SBAC 228 cluster recovery by our method is very high for this
ECOWEB 224 data set. Figure 6.3 (top) shows the centers of the two
Huang’s Algorithm 200 clusters. Cluster 1 on the left has very low values of all
Proposed 257 attributes and represents elements without breast-
algorithm cancer possibility. Cluster 2 represents the malignant

144
class. Cluster 2 is further subdivided and the resulting
clusters are shown at the bottom line. As is expected,
many attributes have same values in both. But the
significant attribute that divides the data is denoted by
the pink line and represents “marginal adhesion” which
is low for cluster 1 and higher for cluster 2.
[3] Jain, A.K. and Dubes, R. C. (1988). Algorithms for
Clustering Data. Prentice Hall, Englewood Cliff, New
Table 6.3 Cluster recovery for Breast cancer data Jersey.
set with our proposed algorithm
[4] Li, C. and Biswas, G. (2002). Unsupervised Learning
Cluster No. Benign Malignant with Mixed Numeric and Nominal Data. IEEE
Transactions on Knowledge and Data Engineering,
1 459 8
vol. 14, no. 4, pp. 673-690.
2 7 240
[5] C. Arima and T. Hanai, (2003) “Gene Expression
500
Number of patients
150
Analysis Using Fuzzy K-Means Clustering”, Genome
Informatics 14: 334. {335 (2003)
Number of

400
100
patients

300 [6] Dembele, D. and Kastner, P., Fuzzy C-means method

200 50 for clustering microarray data, Bioinformatics, 19:973–
100
0
980, 2003.
0
1 3 5
Value 7 9
1 2 3 4 5 6 7 8 9 10 [7] S. Chu, J.Derisi, M.Eisen, J.Mulholland, D. Botstein,
P.O. Brown, and I. Herskowitz,(1998) “The
100 70 transcriptional program of sporulation in budding
80
60
50
yeast”, Science, 282:699–705, 1998.
60 40
40 30 [8] W.H.Wolberg and O.L.Mangasarian, (1990),
20
20
10
"Multisurface method of pattern separation for medical
0 0 diagnosis applied to breast cytology", Proceedings of
1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10
the National Academy of
Sciences, U.S.A., Volume 87, pp 9193-9196.
Figure 6.3 (top) – two clusters “benign” and “malignant”
[9] Huang, Z. (1997). Clustering Large Data sets with
(bottom) sub-clusters of malignant with low and high
Mixed Numeric and Categorical Values, In
“marginal adhesion”
Proceedings of The First Pacific-Asia Conference on
Knowledge Discovery and Data Mining. Singapore.
6. Conclusion and Future Work World Scientific.
In this paper, we have proposed a distance
measure which works for mixed data set, to cluster bio- [10] Ganti, V., Gekhre J. E. and Ramakrishnan R. (1999).
medical data. We have proposed a new representation CACTUS-Clustering categorical data using
summaries, In Proc. of 2nd Intl. Conf. on Knowledge
for cluster center which can capture cluster
Discovery and Data Mining (KDD), pp. 311-314.
characteristics effectively. We are currently working on
implementations of mixed algorithms with our [11] Das G. and Mannila H. (2000). Context-based
proposed measures. similarity measures for categorical databases,In
Proceedings of Principles of Data Mining and
Knowledge Discovery, 4th European Conference,
REFERENCES
PKDD, pp.201-210.
[1] F. Lemke and Müller, J.-A., Self-Organizing Data
[12] Stanfill, C. and Waltz, D. (1986). Toward memory
Mining. Libri, Hamburg 2000, ISBN 389811-861-4,
based reasoning. Communication of the ACM, vol.29,
http://www.knowledgeminer.net/pdf/medical.pdf.
no 12, pp. 1213-1228.
[2] Fayyad, U.M., Piatesky-Shapiro, G., Smyth, P. and
[13] Cost, S. and Salzberg, S. (1993). A weighted nearest
Uthurusamy, R. (1996). Advances in Knowledge
algorithm with symbolic features. Machine Learning,
Discovery and Data Mining. AAA1
Volume 10, pp. 57-78. Boston, MA: Kluwer
Publishers.
[14] E. Han, G. Karypis, V. Kumar, B. (1997) Mobasher,
Clustering based on association rule hypergraphs, In
Workshop on Research Issues on Data Mining and
Knowledge Discovery.

145
 IMPROVED DETECTION OF VENTRICULAR EJECTION TIME
FOR IMPEDANCE CARDIOGRAPHY
V. K. Pandey1 and P. C. Pandey2
1
BME Group, Bio School, IIT Bombay, Powai Mumbai 400 076, India
vinod@ee.iitb.ac.in
2
EE Dept., IIT Bombay, Powai Mumbai 400 076, India
pcpandey@ee.iitb.ac.in

Abstract DeMarzo et al [2] observed that B-point can occur at

any point on the ascending limb of the waveform in
Impedance cardiography is a noninvasive some patients. Further sensing of the variation in
technique for monitoring stroke volume (SV), based on thoracic impedance due to blood flow is influenced by
sensing variation in the thoracic impedance due to respiration and motion artifacts [4], [8], [11], [15],
blood flow. In this technique, first derivative of [16], [17]. These artifacts have large amplitude as
impedance (dz/dt) is used to calculate two parameters: compared to impedance variation due to cardiac cycle,
ventricular ejection time and (-dz/dt)max. Respiration and can cause a large base line drift. ICG signal
and motion artifacts cause base line drift in the sensed bandwidth extends over 0.5-20 Hz. Respiratory related
impedance waveform, and this results in errors in the artifacts extend over 0.03-10 Hz, while motion related
estimation of the two parameters. Ensemble averaging artifacts have band of 0.1-10 Hz [3], [11], [15], [17].
of ICG signal suppresses motion artifacts but it Therefore frequency band of the motion artifacts may
introduces distortion in the signal. In this study, partly overlap with that of the ICG. This causes errors
simultaneously acquired phonocardiogram (PCG) in detection of ventricular ejection time from ICG
signal is used to estimate the time difference between waveform.
the first and second heart sounds as a measure of
ventricular ejection time. Since PCG signal is less
influenced by respiratory and motion artifacts, it gives
a relatively error free measurement.

Key Words
Impedance cardiography; Stroke volume; Ventricular
ejection time; Phonocardiogram.

1. Introduction

Impedance cardiography is a simple, cost

effective and noninvasive technique for monitoring
stroke volume (SV), based on sensing the changes in
the electrical impedance of the thorax, caused by
variation in blood volume during the cardiac cycle [1]-
[13]. Time derivative of the thoracic impedance (-
dz/dt) is known as impedance cardiogram (ICG). The
parameter required for estimating the stroke volume,
using Kubicek, Bernstien, or Sramek formulas [5] are
left ventricular ejection time (Tlvet) and maximum
change in the impedance cardiogram, (-dz/dt)max.
As shown in Fig. 1, ventricular ejection time is
defined as the time difference between points B and X, Fig. 1 Typical impedance cardiogram (adapted from [9]). z(t)
in ICG waveform. The point B, often occurring at onset = Zo - Z(t) where Z(t) is thoracic impedance and Zo is the
of rapid upstroke of ICG, corresponds to opening of the basal impedance.
aortic valve. While the aortic valve closure is identified
as point X, defined as the minimum following the Ensemble averaging is most commonly used
maximum in ICG. Points B and X correspond to first technique for suppressing artifacts in the ICG signal
and second heart sounds respectively. There is no [4], [8], [11], [13], [16], [17]. From ensemble averaged
standard method for detecting point B of waveform. In ICG, ventricular ejection time and maximum change in
most of the analysis, B-point is identified as base line the waveform is measured. In ensemble averaging,
crossover point before maximum peak of ICG. time frames are decided with respect to some internal

147
reference point or with respect to a reference point in ICG waveform. The first and second heart sounds are
another waveform. Hence simultaneously acquired segmented and then cross-correlation is used to
electrocardiogram (ECG) R-points are used to decide measure time interval between them. For separating
time frames and ICG is ensemble averaged on beat-by- heart sounds, energy envelope of PCG is calculated and
beat basis, synchronized with the R-point of ECG. then by threshold detection, first and second heart
However, ensemble averaging suppresses beat-to-beat sounds are segmented. The location of peak in cross-
relation and transient changes in ICG signal. Because correlation gives the time difference between the two
of heart rate variability, ensemble averaging tends to heart sounds, and hence ventricular ejection time Tlvet.
blur or supress the less distinctive point B of the Recordings were done on five normal subjects
waveform and may result in error in its detection [4], with no known cardiovascular disease. Impedance
[11], [13], [14]. cardiograph instrument developed at IIT Bombay [12],
[21] was used for recording ECG and ICG waveforms.
Here we present phonocardiogram (PCG) as an
ICG was sensed by passing a high frequency (96 kHz),
alternative signal source to measure ventricular
low intensity (<5 mA) current into the thorax. Four-
ejection time for stroke volume calculation. The
electrode configuration, with spot electrodes, was used
relationship between heart sounds in the PCG with the
for reducing the effect of skin- electrode impedance. In
B and X point of ICG has been studied and reported
the physical arrangement of outer pair, one electrode
earlier [5], [7], [8], [9], [10], [11]. As shown in Fig. 1,
was placed around abdomen and the other around
points B and X of ICG waveform are synchronous to
upper part of the neck. For the inner electrode pair, one
first and second heart sounds respectively. Hence time
electrode is placed around the thorax at the level of
difference between first and second heart sounds can
joint between xiphoid and sternum and the other
be used as ventricular ejection time Tlvet. The PCG
around the lower part of the neck. The PCG was
signal is less affected by motion artifacts, hence it can
recorded to sense heart sounds by placing a phono-
give relatively error free estimation of Tlvet. We have
transducer (Pamtron, Mumbai, India) on intercostal
used cross-correlation technique to estimate the time
space just to the left of the sternum.
interval between the first and second heart sounds in
PCG signal, acquired simultaneously along with ICG.
The estimation of Tlvet from ensemble averaged ICG as
3. PCG Analysis
well as PCG are computed and compared.
Waveforms Zo, Z(t), ECG, and PCG are
2. Method simultaneously acquired at sampling rate of 1 k Sa/s
using a data acquisition unit (USBDAQ-9100-MS,
As mentioned above, the points B and X of the manufactured by Adlink Technology, Taiwan)
ICG waveform are used to calculate ventricular interfaced to PC through USB port. As shown in Fig. 2,
ejection time, and they are synchronous to first and first PCG signal is passed through a band pass filter
second heart sounds. Intra-subject variability results in (Butterworth low pass filter with fc1 = 7.5 Hz and
event latency which introduces distortion in ensemble Butterworth high pass filter with fc2 = 100 Hz, cascaded
averaged waveform. Visual inspection of ensemble together) to attenuate high and low frequency noise and
averaged waveform shows the ambiguity in detecting B physiological interferences. Further squaring is done
and X points. Therefore it is important to accurately and squared waveform is low pass filtered (fc = 12.5
detect B, C and X points, independent of base line Hz) to get energy envelope.
variation. PCG simultaneously acquired with ICG and Peaks in the energy envelope are located by
ECG can be used for estimating the interval between dynamic thresholding to give the position of first or
these points, because PCG is not much affected by second heart sound. The energy envelope is processed
motion artifacts. as windowed segments, with window length
The PCG signal received from human heart in corresponding to approximately one heart sound
one heart beat has four sounds [18], [19], [20]. duration (300 ms). If the peak within the window is
Generally only two heart sounds are perceived. Closure supra threshold, it is accepted as indicative of heart
of tricuspid and mitral valves generates the first heart sounds, and 60% of the peak is set as the threshold for
sound while closure of aortic and pulmonary valves the next segment. Next the decision process identifies
corresponds to the second heart sound. Third and first and second heart sounds. In the decision process,
fourth heart sounds have very low intensity, and hence the index of detected peaks are compared with previous
most of the time these are not audible. Opening of peak. If the time interval between peaks labeled i and i-
pulmonary and aortic valves coincides with the closure 1 is less than those labeled as i-1 and i-2, then the peak
of tricuspid and mitral valves. Hence the time interval i is marked as second heart sound peak (corresponding
between first and second heart sounds corresponds to to point X in ICG) and the peak i-1 is marked as first
ventricular ejection time. heart sound peak (corresponding to point B in ICG).
In this study, PCG signal is simultaneously After separating first and second heart sounds, cross-
acquired with varying impedance z(t), basal impedance correlation between the corresponding energy
Zo, and ECG for calculating ventricular ejection time envelopes is used to find the delay. The location of
Tlvet and comparing the same with that obtained from peak of cross-correlated waveform corresponds to the

147
delay between first and second heart sounds, and gives energy envelope based estimate was much more
the ventricular ejection time. Cross-correlation between consistent. Fig. 3 shows a typical PCG signal acquired
the segments of the signal waveform as well as those of from subject 'NSM' and output waveform after each
the energy envelope was used, and it was found that the processing step.

Band- pass Low- pass

PCG filter [ ]2 filter
fc1=7.5Hz , fc2=100Hz fc=12.5Hz

1st heart sound position Thresholding

&
2nd heart sound position decision

Fig. 2 Processing stages for separating first and second heart sound

Fig. 3 Output waveforms from processing of PCG, for recording from subject 'NSM' (x-axis: time in s, y-axis: arbitrary units)

4. Results
Processing as discussed in the previous section was MS 68 22.60 3.33 381 360
used to calculate ventricular ejection time from the (2.10) (0.47) (1.59) (13) (9)
ICG waveform as well as from PCG signal. The
recordings were done in normal conditions and post Table 1 gives the estimated value for the parameters
exercise relaxation condition at intervals of 5 min. The for stroke volume calculation: Zo, (-dz/dt)max and Tlvet.
number of ensemble averaging cycle for ICG The
waveform was taken as 8.

Table 1 Estimated values (mean, s.d.) of Zo, (-dz/dt)max,

and Tlvet from ICG and PCG, for 5 subjects under
resting condition. The number in parentheses indicates
the s.d.
Subject Beat Zo (-dz/dt)max Tlvet
/min (Ω) (Ω/s) (ms)
from from
ICG PCG
NSM 78 40.06 5.82 637 649
(1.18) (1.75) (0.22) (98) (79)
VKP 72 20.51 5.33 391 380
(3.12) (0.27) (0.87) (22) (23)
SP 70 22.66 4.49 340 359
(0.89) (0.89) (0.33) (3) (5)
PKL 83 24.17 12.08 305 304
(3.52) (0.42) (0.75) (18) (14)

148
 [5] J. Verdu, Electrical impedance method for the
measurement of stroke volume in man: state of
art, Acta et Comm.,Uni. Tartuensis, Jartu,
Estonia, 1994, 110-129.
[6] G. D. Jindal & J. P. Babu, Calibration of dZ/dt
in impedance plethysmography, Med. Biol.
Eng. Comput., 23(3), 1985, 279-280.
[7] R. P. Patterson, Fundamental of impedance
cardiography, IEEE Eng, Med. Biol. Mag., Mar.
1989, 35-38.
[8] W. G. Kubicek, F. J. Kottke, M. U. Ramos, R.
P. Patterson, D. A. Witsor, J. W. Labree, W.
Remole, T. E. Layman, H. Schoening, & J. T.
Fig. 4 R-R interval and Tlvet vs time plot in resting and post- Garamela, The Minnesota impedance
exercise rest condition for subject ‘VKP’.
cardiograph- theory and application, Biomed.
values of Tlvet, obtained by the two methods, closely Eng., 9, 1974, 410-417.
match, and standard deviations for the two also are
[9] J. Malmivuo & R. Plonsey,
similar
Bioelectromagnetism (2nd ed., New York:
However visual inspection of plots of these values as a Oxford Univ. Press, 1995).
function of time indicates fluctuation in Tlvet estimated
[10] J. Nyboer, Electrical Impedance
from ICG. This was much more noticeable during the
Plethysmography
plots for post-exercise measurements. Fig. 4 shows a
(2nd ed., Springfield, Massachuates: Charles C.
plot of R-R interval and Tlvet for a typical subject.
Thomas, 1970).
During resting condition, we see a few deviation in Tlvet
values and these could be related to error in locating B [11] M. Qu., Y. Zang, J. G. Webster, & W. J.
and X points in ensemble averaged ICG. These Tompkins, Motion artifacts from spot and band
deviations are much more visible in the post-exercise electrodes during impedance cardiography,
plot. This may be because of shift in the B and X points IEEE Trans. Biomed. Eng., 33, Nov. 1986,
due to ensemble averaging when stroke volume and 1029-1036.
cardiac activities are changing.
[12] S. M. Joshi & P. C. Pandey, A cardiac output
monitor using impedance phethysmography,
5. Conclusion Proc. Int. Conf. on Recent Advances in
Biomedical Engg., Hyderabad, Jan 6-8, 1994,
A method for estimation of ventricular ejection time 157-160.
from PCG has been investigated. It may be more
reliable than the estimation from the ensemble [13] Y. Zhang, M. Qu, J. G. Webster, W. J.
averaged ICG waveform, particularly during exercise, Tompkins, B. A. Ward, & D. R. Bassett,
or post-exercise duration, when cardiac activity is Cardiac output monitoring by impedance
rapidly changing. cardiography during treadmill exercise, IEEE
Trans. Biomed. Eng., 33, Nov. 1986, 1029-
References 1041.
[14] G. D. Jindal, S. N. Nerurkar, S. A. Pednekar, J.
[1] L. E. Baker, Principle of the impedance
P. Babu, M. D. Kelkar, & A. K. Deshpande,
cardiography technique, IEEE Eng. Med. Biol.
Corrected formula for estimating peripheral
Mag., Mar.1989, 11-15.
blood flow by impedance plethysmography,
[2] A. DeMarzo & R. M. lang, A new algorithm for Med. Biol. Eng. Comput., 32, 1994, 625-628.
improved detection of aortic valve opening by
[15] B. H. Brown, D. C. Barber, A. H. Morica, & A.
impedance cardiography, Comp. in Cardiol,
D. Leathard, Cardiac and respiratory related
1996, 373-376.
electric impedance changes in the human
[3] J. G. Webster, Medical Instrumentation- thorax, IEEE Trans. Biomed. Eng., 41(8), Aug.
Application and Design (3rd ed., New Delhi: 1994, 729-723.
John Wiley, 1998). [16] H. Riese, P. F. C. Groot, M. V. D. Berg, N. H.
[4] B. E. Hurwitz, L. Y. Shyu, S. P. Reddy, N. M. Kupper, E. H. B. Magnee, & E. J. Rohaan,
Schneiderman, & J. H. Nagel, Coherent Large-scale ensamble averaging of ambulatory
ensemble averaging techniques for impedance impedance cardiograms, Behavior Research
cardiography, Proc. 3rd Annu.. IEEE Symp. Methods, Instruments, & Computers, 35(3),
Comp. Based Med. Syst., June 1988. 2003, 467-477.

149
[17] J. Rosell & J. G. Webster, Signal-to-motion Conf. IEEE Eng. Med. Biol. Soc., 20, 1998,
artifacts ratio versus frequency for impedance 220-222.
pneunmography, IEEE Trans. Biomed. Eng.,
[20] A. H. Mood & J. N. Torry, Time-varying
42, 1995, 321-323.
filtering of first and second heart sounds, Proc.
[18] A. Djebbari & B. Reguig, Short-time Fourier 18th Annu. Int. Conf. IEEE Eng. Med. Biol.
transform analysis of the phonocardiogram Soc., Amsterdam 1996, 950-951.
signal, Proc. 7th Annu. Int. Conf. IEEE
[21] N. S. Manigandan, Development of hardware
Electronics, Circuits & Systems, Dec 2000,
for impedance cardiography, M.Tech. Dissertation,
844-847.
Supervisor: P. C. Pandey, Biomedical Engg. Group,
[19] P. Zhou & Z. Wang, A computer algorithm for IIT Bombay, 2004.
ECG, PCG and CAP, Proc. 20th Annu. Int.

150
 A WAVELET BASED METHOD FOR IDENTIFICATION OF TAGS FROM
SPAMM CARDIAC MR IMAGES FOR HARP MRI
Ajay V. Deshmukh1, Tanish Zaveri2, Vivek Shivhare3, Alok O. Modak4, Vikram M. Gadre5,
Deepak P. Patkar6 and Sona Pungavkar7
1, 2, 3, 4, 5
Indian Institute of Technology, Bombay, Powai, Mumbai 400 076, India
6, 7
New MRI Centre, Nanavati Hospital, Vile-Parle (W), Mumbai, 400 056, India
deshmukh@ee.iitb.ac.in1, ztanish@yahoo.com2, shivhare@ee.iitb.ac.in 3, alokmodak@yahoo.com 4,
vmgadre@ee.iitb.ac.in 5, drdppatkar@yahoo.com 6, drsonap@yahoo.co.in7

Abstract information about heart motion needs that the spectral

peak be extracted appropriately.
Tagged cardiac magnetic resonance images are useful
for making the diagnosis of heart disease. Analysis of
tagged cardiac magnetic resonance images involves
separation of tags, since the tag pattern is deformed by
the heart motion. For this purpose a spectral peak is
extracted from the spatially modulated magnetization
(SPAMM) tagged cardiac MR images. This facilitates
estimation of non-rigid motion and deformation of the
heart. We present, in this work, a novel application of
wavelets for identification of tags from the SPAMM-
tagged MR images of the heart. The method involves
feature extraction in the direction of spatial frequencies
and then reconstruction using selective data from the
wavelet domain. The reconstructed data consists of Figure 2 Tagged Images (1- D. SPAMM) and
features in only one direction. Thus, it has been Magnitude of its Fourier Transform
possible to separate tags in a given direction.
Extraction of more than one spectral peak has been
demonstrated. The method can work with k-space data
which is available directly from the existing MR
scanners.

Key Words
SPAMM MR Tagging, Wavelets, HARP.

1. Introduction
Figure 3 Tagged 2-D SPAMM Image and
Motion and Strain estimation from the cardiac Magnitude of Its Fourier Transform
tagged magnetic resonance (MR) images is a basic
need for the diagnosis of heart diseases. Tagged MR
images are produced by a technique called spatial Ultimately, it becomes a problem to design a two
modulation of magnetization (SPAMM) [1][2]. The dimensional band pass filter to get the spectral
SPAMM technique modulates the longitudinal information. In practice the spectral peaks do not
magnetization of the subject before acquiring image have finite support. Hence efforts are made to
data. Tags are also deformed as per the heart motion. extract the spectral information in a best possible
Therefore, detection of tags from SPAMM images is manner. In this paper, we describe a method for
required for calculating the deformation of the heart. extracting the spectral information from SPAMM
Figure 1 shows a 1-D SPAMM cardiac image of left cardiac MR image, using the DWT and Hilbert
ventricle and its Fourier transform magnitude. The Transform approach.
spectral peaks are oriented in the direction
perpendicular to the tagging lines. Figure 2 shows the 2. The Discrete Wavelet Transform
tagged MR image of the heart with a 2-D SPAMM
pattern. One can see nine spectral peaks. Eight spectral The 1-D discrete Wavelet transform (DWT)
peaks are due to four sinusoidal patterns and a peak decomposes a real valued continuous signal s( x ) in
corresponding to the DC component. Extraction of terms of translations and dilations of a lowpass scaling

151
function þ(x) and bandpass wavelet (x) . Thus, the
dilations and translations of these functions form an
orthonormal basis for the space of finite energy
functions, L2(IR) [5]. The signal s(x ) can be
expressed as a sum of coarse and detail coefficients.
The coarse information is given by

P J
àJ (1)
cJk 2à 2þ (2 x à k)
k
Figure 4 Eight frames of Cardiac (Left Ventricle)
The detail information is Tagged Images (2-D SPAMM Pattern)

PJ P j à j
à j (2) 5. Signal Processing for Tag Identification
dk2 2 (2 x à k)
j= à 1 k
The SPAMM cardiac MR image can be considered to
Where the scaling coefficients are be a magnetization sinusoidal pattern superimposed on
J
8
à the anatomic image. The magnetization pattern can be
c Jk = 2 2 ; s( x ) þ ( 2 àJ
x à k ) dx (3) seen as a number of spectral peaks in the Fourier
and the wavelet domain. Each of these spectral peaks contains the
J
8 coefficients are
à information about the cardiac motion in a certain
d jk = 2 2 ; s( x ) ( 2 à j x à k ) dx (4)
direction. If a spectral peak is extracted then after
reconstruction, its phase directly gives the motion. The
algorithm consists of the following steps.
3. The Hilbert Transform • Decompose the SPAMM MR image at scale
‘j’ using the DWT.
For a real signal x(t) the corresponding analytic
function [5] is given by • The DWT expansion consists of vertical,
horizontal and diagonal coefficients along
z (t ) = x(t ) + i * y (t ) (5) with the coarse information.
• To extract the vertical features only the
Where y(t) is the Hilbert Transform of x(t).
vertical coefficients are considered, for image
reconstruction using the inverse DWT.
The phase of this analytic signal z(t) is given by
• The reconstructed image thus contains only
 y (t )  the vertical tagging lines.
ϕ (t ) = tan −1   (6)
 x(t )  • For horizontal features similar reconstruction
procedures can be followed with only
In case of Cardiac Taggd MR images the phase horizontal coefficients at all the scales.
contains the information about the motion of the heart.
• The reconstructed image is complexified
using the Hilbert Transform and its phase is
4. SPAMM Cardiac MR Images nothing but the angle image.

The data set was obtained from the Nanavati Hospital • A magnitude mask is obtained from the
Vile Parle Mumbai. The first data set consists of a magnitude of the complexified SPAMM
sequence of 10, 1-D SPAMM images of the same slice. image and applied to the angle image obtained
The second data set was recorded for 40 frames with 2- in the last step.
D SPAMM pattern. Figure 3 shows the first 8 images
from the second data set. 6. Results
The algorithm was executed first on the 1-D SPAMM
image from the first data set. Figure 4 shows the DWT
expansion of the first image from the first data set,
with the HAAR wavelets at scale 1. The vertical and
horizontal features can be seen very clearly. Further the
data set was decomposed at higher scales for the
purpose of analysis.

152
Figure 5 DWT Expansion of CardiacTagged image
at scale 1 Figure 8 HARP image obtained after the
application of Magnitude mask
The reconstruction using the inverse DWT was carried The second data set was processed in the similar
out from scale 3, with only vertical coefficients. The manner. Figure 8 shows the DWT expansion of the 2-D
resulting image is as shown in Figure 5 (left). Figure 5 SPAMM image of Figure 2, with Haar Wavelets at
(right) also shows the Fourier transform magnitude of scale 1.
the reconstructed image. If it is compared with the
Fourier transform magnitude of Figure 1, one
difference could be observed that the DC component
has been completely removed by the algorithm and not
seen in the Fourier Transform magnitude of Figure 5.
This indicates the clear extraction of vertical features
from the SPAMM image data. After extracting the
features, the Hilbert transform of the reconstructed
image was obtained. The phase of this complexified
image is shown in Figure 6. Next the magnitude mask
was obtained by using the Hilbert transform of the
original SPAMM image of Figure 1. The mask was
then applied to the angle image to get the HARP image
of Figure 7. Figure 9 DWT expansion of 2-D SPAMM image at
scale 1, with the Haar Wavelet

Figure 10 Vertical features extracted from the DWT

Figure 6 Vertical features extracted using the DWT reconstruction at scale 2 (left), and its Fourier
at scale 2, with the HAAR Wavelets (left) and its transform magnitude 9right)
Fourier Transform Magnitude

Figure 11 Angle image for image in Figure 2, using

the Hilbert Transform (left), Magnitude image
Figure 7 Angle Image Obtained using the Hilbert
obtained from the 2-D SPAMM (right)
Transform (left), Magnitude mask (right)

153
The vertical features were extracted using the 8. Acknowledgements
reconstruction with the inverse DWT. These are shown
in Figure 9. Angle image was obtained using the
The authors thank Dr. Nagraj G. Huilgol for
Hilbert Transform of the reconstructed image as shown
in Figure 10. Figure 10 also shows the magnitude encouraging this research and making the platform of
reconstruction from the original 2-D SPAMM image Society for Cancer Research and Communications,
for obtaining the mask. The mask was obtained by India, available for the research work.
thresholding this magnitude image. The final HARP
image with only vertical tags is shown in Figure 11. The authors acknowledge with thanks the support
provided by the New MRI Centre Nanavati Hospital,
Mumbai, India for carrying out the fMRI experiment,
and verifying the results.

Ajay Deshmukh thanks the Principal, Prof. H. K.

Abhyankar, and the management of his parent
institution, Vishwakarma Institute of Technology, Pune
for sponsoring him to pursue the Ph.D. programme at
IIT Bombay, Powai, Mumbai, India, on full time basis
and with financial support.

References:

Figure 12 HARP image obtained by applying the [1] E.A. Zerhouni, D. M. Parish, W. J. Rogers, A.
magnitude data mask Yang, and E. P. Shapiro, Human Heart: Tagging
with MR imaging-A Method for Noninvasive
Assessment of Myocardial Motion'', Radiology,
7. Conclusion
169, 1988, 59-63.
A Wavelet based method to identify tagging lines [2] L. Axel and L. Dougherty, MR imaging of motion
from SPAMM cardiac MR images has been
with spatial modulation of magnetization',
demonstrated in this paper. The method identifies tags
in one of the two orthogonal directions. The resultant Radiology, 171, 1989, 841-845.
harmonic phase image is useful for analysing the
[3] J. L. Prince and E.R. McVeigh, Motion estimation
Cardiac motion. One distinct advantage of this method
is that it extracts all the spectral peaks in a certain from MR tagged image sequences, IEEE Trans.
direction. Ultimately it serves as a two dimensional Med. Imag., 11, 1992, 238-249.
filter in the Wavelet domain. This kind of two
[4] Nael F. Osman, Elliot R. McVeigh, and Jerry L.
dimensional filtering is one important step in spectral
Prince, Imaging Heart Motion Using Harmonic
peak extraction for HARP MRI. The algorithm needs
Phase MRI, IEEE Trans. Med. Imag., 19(3), 2000,
spectral peak locations as the prior information, that is
186-202.
known in any tagged MRI experiment. The advantage
of this method is that it extracts more than one spectral [5] Patric Flandrin, Time Frequency/ Time-Scale
peaks at a time, in a given direction. Analysis, Academic Press, 1999

154
 STANDARDISATION OF TRADITIONAL TONGUE DIAGNOSIS AIDED
BY COMPUTERIZED DIGITAL PHOTOGRAPHY- A STUDY
*
Dr. Debasis Bakshi, Dr.Sujata Pal
Indian Research Institute for Integrated Medicine (IRIIM)
Mourigram Stn.para, P.O.Unsani, Howrah 711 302, W.B. India
Ph.: 033-2669-6671 / 0652,Mob.098311-11317, E-Mail: dbakshi123 @ rediffmail.com, acuyoga@vsnl.net

Abstract a. The Human Body as a Whole

Introduction: Tongue Diagnosis is a non-invasive The different body- parts, zang-viscera and fu-
diagnostic procedure of different traditional medical viscera, the meridians of the body are linked as an
systems like Acupuncture, Ayurveda etc. In modern inseparable whole. The inner abnormality can present
medicine tongue gets importance but no qualitative on outer parts, while the outer disease can turn into the
evaluation has been performed. In Traditional inner parts. For example, the sore on body surface can
Medicine, along with other traditional diagnostic result from spleen dampness or heart fire; the liver fire
criteria, TONGUE DIAGNOSIS contributes enormous can manifest as the red eyes and hypochondriac pain,
importance in identification & evaluation of and so on.
pathological conditions as a whole, also very useful in
assessing the structural (Yin) and functional (Yang) b. The Unification of Human Body and the
status of organ systems. Surrounding

Methodology: Present study aims at to grab During the long time evaluation, the
tongue images of Normal and already diagnosed physiological function of human body has adapted
Diseases i.e. High Blood Pressure, Anemia, Upper itself to the general changes of surrounding and nature.
Abdominal pain with Digital Camera and On diagnosis, we must consider disease with the
subsequently transferred to Computer database surrounding.
along with clinical findings.
Besides, the tiredness, mental stress and improper
Result: Normal tongue images (including variations diet can all affect the visceral functions and cause
within normal range) show normal clinical findings. In diseases. And the diseases of viscera will result in a
Diseased group, tongue features (color, coating & lower endurance to tiredness, mental changes and poor
characteristics) along with other clinical findings vary appetite, etc.
from normal group. Abnormal tongue features also
vary amongst different diseased groups.
Geographical Seasons & Living Mental
Conclusion: Primary findings of Tongue-data obtained conditions climates conditions surroundings
show Tongue features vary with clinico-pathological
conditions, which can well be utilized as an important
non-invasive diagnostic tool through developing a
standard database. Further study will enrich the science Outer
of non-invasive, low-cost, easily accessible diagnostic
potential of both traditional & mainstream medicine.
Body
Key Words
Tongue diagnosis, Traditional diagnosis, Acupuncture,
Inner
Ayurveda, eHealth, computer aided diagnosis

Viscera Meridians Skin, Muscles, Vessels,

1. Introduction
Tendons and bones
Tongue-Background
The Holism
The holism is a basic concept of Traditional To sum up, the importance of holism (inspecting
Medicine (TM); the holism in diagnosing refers to the both the inner and outer) lies not only on the relation
comprehensive consideration of human body (the between local disorders and whole body, but also on
inner) and the natural surrounding (the outer). It is also the effect on human body from the surrounding. [1]. [2]
called “inspecting of both the inner and outer”. It
includes two aspects as follows:

155
Diagnosis of Traditional Medicine (TM) is a study
concerning the examination of body & mind including
tongue & pulse, differentiation of syndromes and
diagnosis of diseases under the guidance of the basic
theories of TM.
During the long period of medical practice,
doctors of past ages accumulated rich experience of
diagnosis, which formed the comprehensive diagnostic
system specially for TM, i.e., four diagnostic methods
(observation, auscultation & olfaction, interrogation
and palpation) and syndrome differentiation. And
based on those, they established diagnostics of TM. Representation of Internal Organs at Different Parts of
From ancient time to now, the diagnostics of TM has Tongue
been playing its constant role in clinical practice.
Recently, modern scientific methods have been applied The significances of Tongue Diagnosis are as follows:
to conducting the four diagnostic methods and
syndrome differentiation in combination with 1. To judge the exuberance or decline of the genuine
qi: The exuberance or decline of visceral qi and
traditional methods. It is developing in the way of
blood can be shown in tongue. For example, the red
keeping the distinctive features and combining with
and moist tongue means the exuberance of qi and
modern sciences.
blood; while the pale tongue is a sign of deficiency
of both qi and blood. The white, thin and moist
Introduction of Tongue coating indicates an exuberant stomach qi, while no
It is a voluntary muscular organ of the oral cavity coating is due to the decline of stomach qi or
supplied with specific vascular, lymphatic, impairment of stomach yin.
neurological connections. It is involved in varied
physiological functions like deglutition, mastication, 2. To distinguish the nature of disease: Evils of
taste sensation & speech. different natures will make different changes in
tongue. For example, if there is no prickle on
The Relation between Tongue and Viscera (as per TM) tongue surface and the coating is white and moist,
or the tongue is bluish black without prickle, it is
Tongue is the sprout of the heart and the heart is due to cold evil. If there is a red and dry tongue
the supreme monarch of all organs. So disease of with yellow coating, or red prickle tongue with
viscera can influence not only the heart, but also the yellow, thick and greasy coating, it is due to warm
tongue. Tongue is called the out-show of the spleen. or heat evil. The greasy or putrid coating indicates
The spleen dominates transportation and the food-retention. The blue macule or spot on
transformation. So, the tongue is closely related to tongue suggests blood stasis.
splenic function. Tongue coating has special relation 3. To detect the location of disease: In exogenous
with stomach qi. It is said: “In healthy body, there is a diseases, the thick or thin of the coating can reflect
little thin coating like grass roots. It is the embodiment the deep or shallow of disease location. For
of stomach qi activity.” For example, in Acupuncture example, the thin coating suggests that the disease
Meridians, three yang meridians & three yin meridians is in its initial stage, the disease is located in
of foot have connection to the tongue. Traditionally it shallow part and it is an exterior syndrome; while
is said: “ All qi of meridians flow up to tongue. So we the thick coating suggests that the evil enter into the
can know the deficiency or excess, cold or heat of inner part of the body, the disease is located in deep
viscera and meridians by observing tongue.” part and it is an interior syndrome. The crimson
tongue means heat evil is in nutrient or blood-
As because tongue is closely related to viscera phase, disease is very deep, it is a critical condition.
and meridians, so viscera have their representative
areas on tongue surface: 4. To infer the tendency of disease: The changes of
tongue usually follow the changes of genuine qi and
i) Tip of Tongue Represents -- Heart (HE) & Lung evils, and disease location. We can infer the
(LU) tendency of disease by observing tongue, especially
ii) Central Part Represents-- Stomach (ST) & Spleen in exogenous febrile diseases. For example, the
(SP) turning of coating from white to yellow, and from
yellow to black is usually due to the transferring of
iii) Sides Represent -- Liver (LIV) & Gall Bladder evils from exterior to interior, or from cold to heat.
(GB) It shows the deterioration of disease. The change of
coating from thick to thin is a sign of improving or
iv) Root Represents -- Kidney (KI) & Bladder (BL) curing. [3]

156
 In modern western medicine, tongue plays very High Blood Pressure - 120
limited role in disease diagnosis. But in traditional Anemia – 101
medicines as in acupuncture and ayurveda, tongue is Upper Abdominal Pain - 81
considered as a unique diagnostic system for the Images are captured with high-resolution
identification of various pathological conditions. A digital camera under fixed setting and subsequently
great deal of similarity exists in the methodology of transferring the images (along with other clinical
ayurveda & acupuncture in respect to tongue diagnosis. findings) to the computer database.
Results:
Highest % of ‘Blackish patch / spot on sides
Aspects of Tongue Observation of tongue’ mainly at ‘Liver – Gall Bladder area’ is
Vitality of Tongue Colour significant for High Blood Pressure disease group as
Tongue Body Colour shown in fig.2. Presence of blackish/purplish spot or
Tongue Body Shape patch indicates stagnation of blood circulation. As
Tongue Coating these pigmentations lie on the both sides of the tongue
Tongue Moisture i.e. in Liver- Gallbladder area, it means stagnation of
flow in Qi and blood in these particular organs.
Normal Tongue Features
Proper size
Soft in quality
Free in motion
Slightly red in Colour
Thin white coat
Neither dry nor over moist [1], [2]

Role of Information Technology(IT) for

Standardization of TONGUE Diagnosis
In this study, images of Tongue are taken by high 1. Normal Tongue
resolution Digital Camera and transferred to Computer
Database along with detailed case history in Normal &
different Diseased conditions. This standard
technology is applied to analyse the specific detail of
Tongue characteristics to identify different diagnostic
patterns correlated with varied clinical disorders and
healthy conditions.

2. Aim:
This study aimed to standardize & develop 2. Blackish patch both sides in High Blood Pressure
traditional medical wisdom by preparing a
computerized database along with tongue images of
Normal and different Diseased conditions which can be
applied as one of the non-invasive diagnostic
procedures.
By integrating IT with TM, this simple
technology will provide a low cost, easily available &
accessible, non-hazardous diagnostic procedure for the
service of the broad masses of people to develop and
improve health care service. This study might help in
the following: 3. Pale, flabby, swollen with imprint of teeth
in Anemia group
• To enable Traditional Health Practitioner to diagnose
diseases properly in a holistic manner
• To foster the development of computer based health
technology in India
• To develop vital interaction between Computers,
Traditional Healers & Medical Scientists.

Methods:
In this present study, out of 2030 images, 404
tongue images are selected:
Normal - 102

157
 4. Peeled off coating in Upper Abdominal Pain Primary findings of data obtained from the study
Group show Tongue features changes with clinico-
Highest % of ‘Flabby Tongue’ and ‘Pale color of pathological changes, which can well be utilized as an
whole tongue’ correlates mostly with Anemia disease important non-invasive diagnostic tool.
group. As shown in the figure 3, these conditions,
Tongue diagnosis has several distinctions, which
reflects the problems of functional deficiency (Yang) in
make it of critical importance in Traditional
the organs like, Spleen and Kidney, which leads to
diagnostics. In some respects it is more reliable than
increased size of the body and Pale colour of Tongue.
pulse diagnosis.
This deficiency of the two organs, on the other hand,
leads to general weakness, less appetite, dizziness etc. However, it should be pointed out that
and ultimately Anemia. sometimes the tongue is only slightly changed in
some severe cases, and the abnormal changes of
As shown in Fig.4, in Upper Abdominal Pain
tongue are seen in normal people. So, the tongue
Group, highest % of either partial or entire peeled off
observation should be used in combination with
coating (glossy tongue) indicates exhaustion of
other examinations. Only by comprehensive
Stomach (Yin) and severe damage of Stomach (Qi).
analysis on whole clinical data, can we make a
Patients make complaints of acidity, indigestion,
correct diagnosis.
feeling of heaviness and burning sensation in upper
abdominal area specially epigastric region. In the medical world, concomitant development
of non-invasive diagnostic technology has been heavily
marked with the advancement of knowledge regarding
Tongue Study group the limitation of interventional pathological techniques
Features Normal High Anemia Upper Abdo- in regard to transmission of disease through mechanical
Blood minal Pain intermission.
Pressure
Body - 68/30 72/29 48/2 More in depth study is required in each study
Pale + / ++ group to differentiate the varied Tongue features due to
Body Pink 102 13 - 29 the variation of origin of disease and pathogenesis.
Body Deep - 9 - 2
Pink If one wants to establish the science behind the
Flabby 29 48 51 33 tongue diagnosis exhaustive study needs to be
Glossy 2 9 10 10 undertaken.
Imprint 24 78 47 42
Crack 7 39 37 28
Black patch 1 24 7 5
White coat 7 104 69 63 4. Acknowledgements
Yellow - 10 10 10
coat
Sticky coat 16 20 32 24 References:
Peeled-off 6 3 5
coat [1] Giovaani Maciocia, Tongue diagnosis in chinese
medicine ( Seattle, Eastland Press, 2000)
3. Conclusion
[2] Cheng Xinnong, Chinese acupuncture and
From above observations, we can come to a moxibustion 1st ed. ( Beijing, China: Foreign
conclusion that Normal tongue images (including Language Press, 1993)
variations within normal range) show normal clinical
findings. [3] Beijing University of Traditional Chinese
Medicine (Editor), Diagnostics of traditional
In Diseased group, tongue features (color, coating chinese medicine (Beijing China: Academy Press,
& characteristics) along with other clinical findings 1995)
vary from normal group. Abnormal tongue features
also vary amongst different diseased groups.

158
MEDICAL INSTRUMENTATION
 ECG DATA ACQUISITION AND MONITORING SYSTEM FOR
TELEMEDICINE APPLICATION

Prashant Agrawal, Arundhati Jana and Ajit Pal

Indian Institute of Technology Kharagpur
Department of Computer Science & Engineering, IIT Kharagpur
Kharagpur 721302 (West Bengal), India
pagrawal@cse.iitkgp.ernet.in, ajana@cse.iitkgp.ernet.in, apal@cse.iitkgp.ernet.in

Abstract The importance of Telemedicine can be ascertained

from initiatives taken by the Department of Space,
This paper presents the development of a system that Department of Information Technology and other
will enable integration of analog ECG machines with public and private organizations [1]. Our project is also
the Telemedicine technology. The system is being being supported and sponsored by the Ministry of
developed in the context of the medical standards and Human Resource Development, New Delhi.
practices in India, addressing problems specific to
Indian health care scenario. Basic operations along Specific clinical systems are developed and put in
with hardware and software architecture implementing operation for use by physicians and other healthcare
the system are described. personnel. One such important system is the ECG
monitoring system. This system continuously records
the ECG signal of the person under observation and
Key Words stores it. In the Telemedicine context, an ECG
Telemedicine, ECG signal, Patient monitoring,
monitoring system can be used for Tele-diagnostics but
Embedded System
it requires a digital ECG machine. However, the
majority of the ECG devices used in India are analog in
1. Introduction nature, which cannot be directly interfaced with the
Telemedicine technology.
For more than 30 years, clinicians, health services
researchers, and others have been investigating the use This paper presents development of an interface
of advanced telecommunications and information system for the analog ECG machines for their
technologies to improve health care. At the intersection integration with the Telemedicine technology. The
of many of these efforts lies telemedicine—a system presented in this paper is being developed as a
combination of innovative and mainstream prototype to understand the complete functionality
technologies. Telemedicine is the use of electronic which will help us in the development of a System-On-
information and communication technologies to Chip for ECG data acquisition and monitoring system.
provide and support health care when distance Our system is a microcontroller-based system,
separates the participants [5]. which acquires signals from an analog ECG machine at
In Indian context, its 1 billion populations are a specified sample rate. The acquired signals are stored
predominantly rural and distributed in distant in memory and simultaneously displayed on a standard
geographical locations apart from the high-density oscilloscope for monitoring purposes. For closer
urban areas. The health and medical facilities presently observation, the display can be made stationary. The
available in the non-urban parts of the country is acquired data can be transmitted to a local computer for
inadequate and there are wide disparities in terms of further analysis and diagnosis. On the PC side, the
health care infrastructure, facility, manpower and funds transmitted data is stored in a database, which can be
between rural and urban communities, and between retrieved later and displayed graphically. The system
different states and even districts within states. This has been developed using the cheaper components,
calls for innovative methods of utilization of science which are also easily available in India.
and technology for the benefit of our society and
The paper is organized as follows. Section 2 gives
telemedicine assumes a great significance to
a description of the system. There are two subsections
revolutionize the health care system in India.
in it. The subsection 2.1 describes the hardware sub-
Adaptation of Telemedicine technology offers one of
system and the subsection 2.2 describes the software
the best options for delivering healthcare for rural &
sub-system. Section 3 provides the concluding remarks
geographically distant population spread across
besides outlining the future enhancements and
India [1].
modifications planned for this system.

159
 We have used a multi-channel ADC because in the
EC G Input AD C subsequent versions of the system we shall be
acquiring signals other than ECG. The conversion time
for the ADC is about 100µs.
LC D
Key board Microcontroller Counter
Display Data Storage and Display Unit: The system acquires
4 cycles of the ECG signal at a sampling rate of 300
Line Driv er/ Bidirectional
Receiv er
RAM
Buf f er samples/ sec, i.e. approximately a total of 1K sample
values are acquired. The acquired samples are stored in
To PC/ Modem
DAC To CRO the external RAM. For the purpose of local monitoring,
the acquired sample values have to be displayed on an
Figure 1: Block Diagram of the Hardware Sub-System oscilloscope. The time needed for the acquisition of the
frame, consisting of four cycles of ECG signals is
2. System description about 60 x 4/72 = 3.3s (normal heart rate is 72 cycles/
2.1 Hardware Sub-system min). A standard oscilloscope does not produce a
flicker-free display of such a low-frequency signal.
The hardware sub-system is the mobile, Since the captured data are already available in the
battery operated field device. This sub-system is RAM, these data can be accessed at a much faster rate
interfaced to the ECG machine and it acquires and the data can be applied to a DAC to generate the
analog signals from it, digitizes them and displays same ECG signal with frequency translation. This
them for local monitoring. The acquired signals frequency translated signal will produce flicker-free
can also be transmitted to a local computer for display on a standard oscilloscope. A display rate
further analysis and diagnosis. which is 1000 times faster than the data acquisition rate
will be sufficient for this purpose. Therefore, data are
The hardware sub-system is a microcontroller stored in the RAM (while acquiring) at the rate of
based system. Its main functionalities are: 3.3ms and data are read from the RAM at the rate of
• Data Acquisition 3.3µs for flicker free display. To perform both
• Data Storage and Display acquisition and display concurrently, for each data
• Data Transmission written into the RAM, the entire 1K byte RAM is
scanned and displayed. The data reading operation is
Fig.1 shows the main components of the too fast to be achieved by any programmed data
hardware sub-system which can be grouped into the transfer technique. Hence, we have used a DMA-like
following units: CPU, user interface unit, data data transfer technique in this system.
acquisition unit, data storage and display unit, and
communication unit. We have used preset counters to implement
the DMA-like technique. In the COUNT mode, the
CPU: Atmel AT89C52 microcontroller has been used preset counter acts as a 10-bit ripple counter generating
as the CPU for the system. Atmel AT89C52 is a low- consecutive addresses for the display RAM. The clock
power, high-performance CMOS 8-bit microcomputer has a time period of 3.3µs. Display RAM data are read
with 8K bytes of Flash programmable and erasable at this rate and fed to a Digital-to-Analog Converter to
read only memory. Besides, it is also ideal for the generate the frequency translated ECG signal to be
development of flexible and cost-effective embedded applied to an oscilloscope.
systems.
For each scan through the entire 1K byte of the
User Interface Unit: The user interface consists of a data, two processes take place:
4x4 matrix keyboard and an alphanumeric LCD.
a. The counter is forced to the LOAD mode. In this
Keyboard is used to take inputs like Patient’s Id,
mode the counter is presetted with RAM location
Channel No., and ECG lead No., Start/stop monitoring
value from where the scanning starts in the
and Scroll/ Freeze display. The status of the system and
the input by the user is displayed on the LCD. following scan cycle. Each time the scanning of
the display RAM data starts from one location
Data Acquisition Unit: Data acquisition involves higher than the previous scan with a fold back to
capturing of signals from a standard single-lead ECG 000h after 3FFh.
machine using 12-lead system. The range of the b. The register, which keeps track of the address at
frequency of the ECG voltage signal is 0.5 to 100 Hz. which writing into the display RAM takes place, is
The signal is sampled at a rate of 300 samples/ sec [1]. incremented by one to point to the next RAM
The ECG signal is acquired by an 8-bit, 8- location and AD conversion for new data is
channel multiplexer analog-to-digital converter. The started.
microcontroller compatible ADC0808 has been used.

160
 Therefore, as acquisition of one ECG data is
done, previously acquired data are scanned for display User
purpose.
Login
There are 2 modes of display – Scroll mode Information
and Freeze mode. The scroll mode in which the display
on the screen slowly moves towards the left giving a Software
Hardware Medical Graphical
scrolling effect as discussed above. Subsystem Data1
Subsystem
Display
User
1.0
Freeze mode is used for a closer observation of the
ECG signal. In this mode, the display on the screen is
made stationary by bypassing the write operation and
scanning the display RAM from the same location Figure 2: Level 0 Data Flow Diagram
repeatedly [2].
Hardware
Communication Unit: The data acquired for local Subsystem
monitoring and details of patient can be transmitted to Medical Data1
a local PC. Serial port communication conforming to User

RS232C standards and at a baud rate of 9600 is used Login

Receive
Data
Medical Data2
Receive Data
for the transmission. 9 pin D-type connector in Null- Information Request 1.3

Modem configuration is used [4]. The format of the MedicDB

transmission packet is shown in Table 1 [1]. Verify
Login Info
Verification
Request
Handler
OK
1.1 1.2 Graphical
The transmission is initiated by the user by hitting Display User

the TxD (Start Transmission) key on the keyboard. The Login

Display Data
Request Display
system transmits using packets of size 1038 bytes. Information Data
Login Id 1.4
While transmitting data, acquisition of ECG signals is Patient Medical
Details Data2
halted and it is resumed once the transmission is over.
MedicDB MedicDB

Byte No Description
1 Start Byte (96h)
2 – 11 10 Digit Patient Id Figure 3: Level 1 Data Flow Diagram
12 Channel No
3. Conclusion
13 ECG Lead No
14 – 1037 1024 Bytes of sampled values This system is being developed using the
1038 Stop Byte (97h) prototyping approach of system development. At the
Table 1: Format of Data Packet time of writing, we have been able to implement the
basic system. Currently, the system is able to acquire,
2.2 Software Sub-system display and transmit ECG signals. Fig. 4 shows the plot
of ECG Data acquired using the present system.
The software sub-system is a Microsoft Windows We are working towards improving the system by
based PC application performing the following adding the following features and by implementing the
functions: following modifications:
a. Provides access of the software to the authorized a. Acquisitions of signals other than ECG
users only. The method of password protected
access should be used. b. Transmission of data to a remote PC using a dial-
up modem
b. Receives the packets transmitted by the hardware
sub-system. The software attaches date and time c. Data communication with PC using IrDA interface
stamp to the packets received and store them in a d. Optical isolation of power supply and data
database. communication interface
c. Enables user to retrieve data from the database and e. Use of a graphical LCD instead of oscilloscope for
display it graphically. displaying the acquired signals
The Graphical User Interface of the software has f. Implement the system using FPGA based
been developed using Microsoft Visual Basic 6.0. Integrated Circuit followed by an ASIC.
Microsoft Access 2002 has been used as the back end
database [1]. The top level description of the software g. Development of analog subsystem to replace the
is illustrated using the data flow diagrams in Fig. 2 analog ECG machine.
and 3.

161
4. Acknowledgements Appendix A - Data Dictionary

We extend our sincere gratitude to the a. Login Information

Ministry of Human Resource Development, New Delhi Login Information = Username + Password
who are supporting this work by a grant. Username = ((Char)|(Digit))((Char)*|(Digit)*)
Password = ((Char)|(Digit))((Char)*|(Digit)*)
We also take this opportunity to thank Mr. P. K. Char = [a-z A-Z]
Purkait and other members of staff in the Hardware Digit = [0-9]
Lab, CSE Department (IIT Kharagpur) for their
b. Verification OK
continuous help and ever willingness to help us during
Verification OK = [0|1]
the development of the project.
c. Receive Data Request
References Receive Data Request = 1
[1] Technical Working Group for Telemedicine d. Display Data Request
Standardization; Recommended Guidelines and Receive Data Request = 2
Standards for Practice of Telemedicine in India e. Medical Data1
(New Delhi: Department of Information Medical Data1 = Patient Id + Channel No + Lead
Technology, Ministry of Communications and No + Data
Information Technology, 2003) Patient Id = (Digit)10
Channel No = (Digit)1
[2] Ajit Pal; Microprocessors Principles and Lead No = (Digit)2
Applications (New Delhi: Tata McGraw-Hill Data = [0-255]1024
Publishing Company Limited, 1990)
f. Medical Data2
[3] M A Mazidi, J G Mazidi; The 8051 Medical Data2 = Patient Id + Channel No + Lead
Microcontroller and Embedded System (Delhi: No + Data + Date + Time
Pearson Education Asia, 2000) Date = Day + Month + Year
Time = Hour + Min + Sec + Millisecond
[4] Jan Axelson; Serial Port Complete (Mumbai:
Day = [1-31]
Penram International Publishing (I) Pvt. Ltd.,
Month = [1-12]
1998).
Year = (Digit)4
[5] Marilyn J Field; Telemedicine – A Guide to Hour = [0-23]
Assessing Telecommunications in Health Care Min = [0-59]
(Washington DC: National Academy Press, Sec = [0-59]
1996) Millisec = [0-9]
[6] Intel; MCS 51 Microcontroller Family User g. Patient Details
Manual (February 1994) Patient Details = Patient Id + Channel No + Lead
No + Data + Date + Time
[7] Atmel; Atmel 8051 Microcontrollers Hardware
h. Graphical Display
Manual (2003)
Graphical Display = A graph plot of data

Figure 4: Plot of ECG Data Acquired on a PC using the Developed System

162
CHARACTERIZATION OF MOTION ARTIFACTS USING WAVELET TRANSFORM AND
NEURAL NETWORKS

Vijay S Nimbargi, Vikram M Gadre and Soumyo Mukherji

IIT-Bombay, Powai, Mumbai 400076, India
vinim_s@rediffmail.com, vmgadre@ee.iitb.ac.in, mukherji@iitb.ac.in

Abstract 2. Wavelet Transform

Motion artifacts, in ECG-s, have been traditionally An ECG signal corrupted with motion artifacts is a
considered to be a problem in ambulatory recordings. non-stationary signal which cannot be analyzed
However, they may contain information regarding the effectively with Fourier transform (FT). FT gives very
movements of the patient, which may help diagnosis, good frequency localization, but no time localization.
particularly in telecardiology applications. In this paper Short term Fourier transform (STFT) may be used for
ambulatory ECG-s were analyzed to extract the analysis of such signals. In this technique the signal is
signatures of three typical movement patterns by divided into small parts by windowing and the Fourier
characterizing the low frequency artifacts. Daubechies- transform of those signal segments are taken to obtain
3 wavelet has been used for analysis and denoising. STFT. The STFT provides time localization as well as
Backpropagation based neural network is used for frequency localization but due to the fixed window
characterization. The effects of both amplitude and length performance is limited and there is a lower limit
phase of ECGs on the performance of the neural on time and frequency resolutions. Taking these
network has been observed. shortcomings into consideration as well as the nature
and possible characterization to be performed on the
signal, Wavelet Transforms have been used in this
Key Words study.
Telecardiology, motion artifacts, wavelets and neural
networks The discrete wavelet transform (DWT) can be
calculated using filter banks, which are a series of,
filters having a relationship between them. The figure
1. Introduction below shows the steps involved in computation of
DWT.
Telemedicine relies heavily on data collected from
remote sources for analysis by medical practitioners. The filters h1 ( n) and h0 ( n) are related through the
To arrive at an accurate diagnosis the physical state and following equation,
activity level of the patient is required. One typical
example of such a situation is ambulatory ECG-s,
which is prescribed and obtained from patients in
whom abnormalities appear only during physical
activity. Recordings of ECG-s during such activity
contain motion artifacts (low frequency noise in
surface electrograms occurring due to the disturbance
of the electrode-skin interface [1]). These artifacts
make diagnosis a difficult job and, traditionally, the
emphasis has always been on elimination of these
artifacts. However, these artifacts may also contain
information regarding the movements of the patient, Figure 13 : Decomposition steps of DWT [2][3].
which, if extracted, might aid diagnosis.

In this study ECG signals containing motion artifacts

h1 ( n) = (− 1) h0 (1 − n ) ….…(2.1)
were decomposed using wavelet transform into pre- n
decided wavelet levels. Coefficients, characterizing the
The filters above have the properties of lowpass and
motion artifacts at these levels, were chosen. A neural
network model using the back-propagation algorithm highpass filters. The filter h0 ( n) has the structure of a
was used to classify the artifacts using the wavelet- lowpass filter, performing the action of a scaling
derived coefficients. Effect of phase of the input function and h1 ( n) has that of high pass filter
signals on network performance was analyzed.
performing the action of a wavelet.

163
 The input signal to be decomposed is passed Basic model of neuron is as shown in the figure 2
through a series of lowpass and highpass filters to get consisting of input nodes, output node, synaptic
approximation and detailed coefficients. The outputs of weights and a bias term. The input signals are given to
lowpass filters are downsampled by 2 to get the input nodes and the network learns by changing the
approximation coefficients and outputs of highpass synaptic weights in accordance with a learning rule.
filters are downsampled by 2 to get wavelet m
coefficients. u k = ∑ wkj x j (3.1)
Wavelet transform has the property of j =1
concentrating the information content in the signal in v k = bk + u k (3.2)
fewer number of wavelet coefficients. This property
has been often exploited in signal compression and y k = ϕ (v k ) (3.3)
denoising methods. Wavelet based denoising methods
are used for elimination of motion artifacts. Where, y k is the actual output of the neuron,
There are two types of thresholding methods [4], v k is called local induced field, ϕ is activation
which are widely used for denoising,
function, bk is the bias which has he effect of
• Hard thresholding
changing the net input to the activation function
• Soft thresholding depending on whether it is positive or negative and
In these thresholding methods, the wavelet wkj is the synaptic weight connecting from neuron j to
coefficients are passed through a pre-calculated k.
threshold, and a specific action is performed depending
upon the type of thresholding. Soft thresholding is Multilayer perceptron with one hidden layer is
shown to produce better denoising and visual quality used; the activation function used is a hyperbolic
than hard thresholding and hence has been used. tangent function in all the layers.
Soft thresholding method is applied using the Backpropagation algorithm is used for learning
following formulae, process, where we try to minimize the cost function
x = ( x − γ ) if x > γ defined as average error function or global error ε av ,
0 if x ≤ γ . (2.2)
1 N
Where, γ is the threshold, calculated using the ε av = ∑ ε ( n)
N n =1
.(3.4)
formula,
γ = K 2 log(n ) (2.3) m
1 0 2
Median absolute value
ε ( n) = ∑ e ( n)
2 j =1 k
(3.5)
K= (2.4)
0.6745
m0 is the number of output nodes where
3. Neural Networks instantaneous error is calculated by taking the
difference between the desired response and actual
Neural networks are designed to mimic the response. ε ( n) is the summed instantaneous error at
functions of human brain and are widely used for the output layer. In the above equations N represents
pattern recognition and classifications. Neural the number of training samples used for training the
networks are made up of basic computational units network
known as neurons interconnected through synaptic
weights, which form the core part of learning process. The weights were updated after every epoch of
The advantage of using neural network is that it learns data is presented to the network and global error is
the complex problems with ease and also generalizes updated by calculating the error at output nodes, the
from it. weights are updated by back propagating the error till a
stopping criteria used is met. Three different stopping
criteria have been used such as; rate of change of
global error is less than 0.01, 0.0001 and 0.000001
percent. For more detailed explanation on
backpropagation algorithm can be obtained from texts
[5][6].

4. Data Recording
A single lead ambulatory ECG recorder (under
development at IIT-Bombay) was used to record Lead
Figure 14: Basic model of neuron [5].

164
 8
II ECG from 5 subjects while they performed pre-
determined motions. Subjects performed the following 7 7
manoeuvres: Walking on a flat surface (W), climbing 6
stairs (C), moving the arm up and down (H). The 5
6

sampling rate was 363Hz.

NPC
4
A total of 30 data files of each pattern were 3
recorded, each file consisting of 1024 sample points. A 2
2
total of 25 files of each pattern were used for training
the network and 5 were used for testing the network. 1
Some data files with motion artifacts such as finger 0
movements, scratching around the electrode- skin 5 10 15 20
interface were also done at a sampling rate of 250Hz Hidden nodes
using a desktop ECG machine.
CW CH WH

Figure 5. No. of proper classifications(NPC) versus hidden

nodes, with a learning rate of 0.005 and momentum of 0.3,
stopping criteria 0.0001%.

W6, W7 and W8 levels that corresponded to a frequency

band of 0.4 to 3.5 Hz. From this it may be concluded
that the motion artifacts will primarily reside in this
frequency band. The coefficients at these levels for
each signal were used as inputs to the neural network
for motion artifact classification.
Figure 3 shows an example of elimination of the
finger movement artifact, from where the wavelet
levels were determined. The wavelet coefficients were
Figure 3. Elimination of finger movement artifact. normalized to keep the values between -1 and 1;
synaptic weights were initialized from uniform random
distribution of 0 and 1.
Figure 4 shows a representative data of the
recordings made on a subject. In order to choose the
optimum number of hidden nodes, learning rate and
momentum of the network, data sets were used
(besides the training data set) for which the movements
were known. Each of the parameters was chosen based
on the ability of the network to properly classify (NPC)
the movements viz. CW (Climbing versus Walking),
CH (Climbing versus Hand movement) and WH
(Walking versus Hand movement). Figures 5 to 7 show
the effect of varying each of the three parameters
whilst keeping the other two constant. From figure 5 it
can be seen that the number of proper classifications is
best for 6 hidden nodes (7,6 and 2 proper
Figure 4. ECG recordings of subjects, resting, classifications for classes respectively). Hence, 6
walking, climbing stairs, and up down hand hidden nodes were used for refining the learning rate
movements of ECGs in the order from top to bottom. and momentum, which were chosen, based on data
represented in figures 6 and 7.
5. Results
The final optimized network consisted of,
The objective of this study was to classify • 6 hidden nodes
motion artifacts. Since the exact characteristics of these • 0.005 learning rate
artifacts were not known, wavelet based thresholding • 0.3 momentum
methods were used to determine the wavelet levels • 0.0001% stopping criteria
corresponding to these artifacts. The input signals was
decomposed into 8 levels using wavelet transform. The Besides evaluating the performance of the network
motion artifacts could be effectively eliminated from for ambulatory ECG-s the performance of the network
the ECG-s by thresholding the wavelet coefficients at was also tested with the resting ECG as a verification.

165
5 data files were used as testing data. A threshold of Table 1
0.15 was chosen i.e. tolerance of ±0.15 was used from Number of proper classifications
the desired responses. The number of wrong Movements
Non-referenced Referenced
classifications was 1 in 5, for each of the types of CW 7 10
movement versus a resting ECG. CH 6 9
WH 2 5
8
7 7
6. Conclusion
6 6

5 The wavelet based neural network algorithm

discussed in this paper was able to classify and
NPC

4
3
segregate some common human movements recorded
as motion artifacts in an ECG signal. The network
2
2 showed excellent performance in discerning two types
1 1
of movements (using a learning rate of 0.005,
0 0 0 momentum of 0.3 and weight initialization between 0
0 0.01 0.02 0.03 0.04 0.05 0.06 and 1 chosen from uniform random distribution). This
Learning rate could even be achieved with recordings from a single
lead where some of the artifacts from different classes
CW CH WH
of movement might present themselves as similar
Figure 6. Number. of proper classifications (NPC) versus signals. The comparatively poorer results derived for
learning rate, hidden nodes=6, momentum=0.3, stopping walking versus hand movement may be attributed to
criteria 0.0001%. this effect.
It may be surmised that such analysis done with a 3-
lead ECG recording will yield much better
8
7
classification of physical movements of the subject. An
7 automated classification of patient movement may be
6
6 an important tool in the hands of a cardiologist, who
5 might be able to correlate the occurrence of certain
types of arrhythmias with certain physical activities.
NPC

3
2
7. Acknowledgements
2

1 The authors wish to acknowledge the support of

0 Prof. R. Lal and the Silicon Locket group in Electrical
0 0.2 0.4 0.6 0.8 1 Engineering Department of IIT-Bombay and TCS in
Momentum
performing the above study.
CW CH WH

References
Figure 7. Number of proper classifications(NPC) versus
momentum, hidden nodes=6, learning rate=0.005, [1] J. G. Webster, Medical Instrumentation (John
stopping criteria 0.0001%. Wiley & Sons, New York, 1998).
[2] M.Vetterli & C.Herley, “Wavelets and filter
banks: theory and design”, IEEE Trans. Signal
Effect of Phase processing, 40(9), 1992, 2207–2231,.
[3] C.S. Burrus, R.A.Gopinath, & H.Guo,
In the formulation of aforementioned network the Introduction to Wavelets and Wavelet Transform,
effect of location of the sample window in the time A primer (Prentice-Hall International, New Jersey,
domain was not considered. However, the predominant USA, 1998).
features in the input signal (viz. the characteristic ECG [4] D.L. Donoho, “Denoising by soft thresholding”,
waves, P,Q, R, S and T) might play a significant role in IEEE Trans. Inform. Theory, 41(5), 1995, 613–
the classification process. They may contribute 627, .
negatively if the window was not initialized at a fixed [5] S. Haykin, Neural Networks: A Comprehensive
reference point. To consider this effect the data files Foundation (Pearson Education Asia, Delhi, India,
were further -analyzed using a window that started at 2001).
the peak of the R-wave The results show a marked [6] M .J. Zurada, Introduction to artificial neural
improvement in classification as shown in Table 1. systems, (PWS publication, 1992).

166
EEG SIGNAL PROCESSING FOR MONITORING DEPTH OF ANESTHESIA
Anil Srivastav1, Amod Kumar2, Y.L. Narayanan3 and Sneh Anand1
1Center for Biomedical Engineering, Indian Institute of Technology & All India Institute of Medical Sciences,
New Delhi
and 2CSIO, Chandigarh, 3 PGI, Chandigarh

167
168
169
170
 A NEW LOW-COST MUSCLE STRENGTH TESTING SYSTEM FOR
NEUROLEPTIC PATIENTS
P.Thirusakthimurugan and P.Dananjayan
Dept. of Electronics and Instrumentation Engineering, Pondicherry Engineering College,
Pondicherry-605014, India.e-mail: thirusakthimurugan@rediffmail.com

Abstract Pinch Force DISPLAY

Patient Sensor
The design and implementation of a client based
testing system to rehabilitate and strengthen muscles
for a patient with Neuroleptic diseases is described. Instrumentation
The patients with neuroleptic diseases like multiple Amplifier
User Check

Parallel Port
sclerosis are used to go to the hospital at least once in Personal
Through
every three months in order to test their strength and Key Board Computer
endurance using a variety of traditional Instrument. ADC
Also, it is very difficult and expensive for patients to Interface
travel to the hospital for regular checkup. Hence this Data Storage
paper is focused on the design and development of a
client based muscle strength testing package, which Figure.1. Block diagram of pinch strength testing
includes both low cost hardware and user-friendly device
software for their use in the home. In-home testing is
advantageous to increase the availability and frequency Testing at the hospital is very expensive and also
of testing if they have a personal computer (PC). In this difficult for the patient to reach in multistoried
paper a new transducer is designed and integrated with building, because they cannot go to the hospital by
a PC through a low cost Analog to Digital interface. themselves and they need another person to escort
Visual Basic (VB) software is used for the system them to go to the hospital. As a result, an interactive
integration to record, display and store information that PC based muscle strength testing system is proposed,
was initiated from the transducer to document and which allow the patients to do the muscle testing
analyzes patient progress. The experimental results privately in their home itself. In this project, it is
show the effectiveness and feasibility of the proposed assumed that patients have home computers. The client
system. specifies that the tests to be run by applying pinch
force on a transducer that, in turn, generates electrical
Key Words signals through an amplifier into an A/D converter. The
converter translates the electrical signal into a format
Sensor fabrication, hardware design, software design the computer can read, store, and send to the
Neurology specialist. The main concentration in the
project is that each patient would need to be evaluated
1. Introduction in their ability to complete this task on their own or
with help of their caretaker. The block diagram shown
One of the neuroleptic diseases is a due to loss of in Figure.1 describes the overall setup of the interactive
fatty tissue called myelin in multiple areas, leaving scar muscle strength program.
tissue known as sclerosis. In some cases, the nerve
fiber itself is damaged or destroyed and creates
neurolysis [1]. When myelin or the nerve fiber is 2. Hardware Design
damaged or destroyed, the ability of the nerves to
conduct electrical impulses to and from the brain is 2.1. Fabrication of Pinch transducer
disrupted and this produces the various symptoms like
blurring, graying, and loss of vision, usually in only
one eye [2]. A patient has to go to hospital in a regular The pinch force transducer is a device that
interval of time. In each visit, the physician tests the converts the mechanical energy generated by the
patient’s muscle strength and coordination. The result patient’s finger muscle into electrical signals [4]. It is
is then compared to the patient’s previous results in designed with two strain gauges bonded on a ‘S’
order to determine whether the muscles’ strength and shaped aluminum member, two commercially available
coordination are improving or worsening [3]. The strain gauges have been glued to the upper and lower
physician may detect a relapse and perform necessary rectangular section of the central limb where maximum
actions, such as making changes in medication by strain is developed as shown in Figure.2.
examining the patients regularly.

171
 low power consumption, high common mode rejection
ratio and high slew rate are desirable for superior
performance. So in this paper, a high precision low
noise instrumentation amplifier AD 624 is used [6].
The complete hardware circuit used in this work is
shown in Figure.3. Here voltage source given to bridge
is maintained constant by the op-amp and transistor
circuit. For balancing there is a provision for coarse
and fine adjustment. Bridge circuit consists of two
strain gages of 240 ohms in push pull configuration and
other arms contain 240 ohms resistor. Output of the
Figure.2. Pinch force sensor bridge circuit is given to the instrumentation amplifier
(AD624).Then the resultant output is given to the A0
Pinch force sensor converts mechanical stress terminal of the ADC interface. The designed pinch
exerted by the patient into electrical signal. An force sensor is calibrated for zero, span and linearity
Instrumentation amplifier (IA) is used to measure the adjustments. The output response of the muscle
very small output signal produced [5] by the force transducer is almost linear for the entire range of
transducer. The IA designed in this work is intending measurement. Thus it can be used for practical
for precise, low-level signal amplification where low applications, especially for muscle force measurement
noise, low thermal and time drifts, high input resistance purpose.
and accurate closed loop gain are required. Besides,
10 Ω
+12V
+5V
OP-07 SL100

DISPLAY
Monitor

CPU
240 Ω 240 ±∆Rg Ω
1MΩ 100 KΩ
10KΩ IA ADC
Zero adjust Interface
(Fine) 240 Ω 240 ±∆Rg Ω

10 KΩ KEY BOARD
Zero adjust
(Coarse)
Pinch
Sensor

Figure.3.Diagram of the complete hardware setup

2.2. ADC Interface design

In the paper, an extremely simple Analog to It is connected to the parallel printer port, and runs
Digital Converter IC ADC0833 is chosen for the from a 5V supply. Analog input voltage range is 0 to 5
Interfacing. The ADC0833 is an 8-bit successive V. Driver software is written in Turbo C. The
approximation A/D converter with a serial I/O and measured speed is roughly 1200 samples/sec on an 8
configurable input multiplexer with 4 channels [7] is MHz XT, and 5700/sec on the same PC with a 10 MHz
shown in Figure.4. 80286/cache accelerator card turned on. For signals,
1KΩ which are noisy beyond 1 LSB (roughly 20 mV),
+5V
taking several readings and averaging them will
1 improve the effective resolution by the square root of
12 D0 (pin 2 of DB25 male plug)
the number of readings. The 4-channel multiplexer is
A0 3 software configured for single-ended or differential
2 D1 (pin 3 of DB25 male plug)
A1 4 D2 (pin 4 of DB25 male plug) inputs when channel assigned by a 4- bit serial word.
ADC 13
A2 5 0833 Acknowledge (pin 10 of DB25
10 male plug) 2.3 Design of Digital Display
A3 6
7 Ground (pin 18 of DB25 male plug)
Ground 8 The signal coming from the AD624 is also converted
7 9
into digital signal by another A-D converter, then fed
to display circuit, which has BCD to seven segment
0.1µf LM336Z-2.5V decoders, display drivers etc., instead of going for
separate A/D converters and display circuit, a single
Figure.4.ADC interface to LPT1 CMOS IC, which include A/D converter, seven

172
segment decoders, display drivers, a reference and a The software is developed to process the patient
clock, i.e. ICL 7107 is used [8]. The A to D conversion data and give their present condition. The photograph
is done utilizing dual slope conversion technique. The of the working model of the system is shown in
power supply frequency noise is suppressed. The figure.6. There are several ethical concerns
output value is displayed on a 3½-digit common anode surrounding the implementation of an in-home
seven-segment LED display [8] strengthening/testing program for MS patients. One
issue is that unauthorized people will use the program,
3. Software Design to avoid this problem, warning labels could be attached
to the hardware and the software indicating that the
Visual Basic is chosen as a programming program is for the use of the patient only. Another
language [9] for user-friendly operation. Security concern is that the patient may try to squeeze the
issues are also implemented in this platform. The transducer more softly in order to continue taking
created program allows the patient to perform all medication or receiving disability insurance. These
exercise and testing tasks at home using this transducer problems could be avoided by continuing to check the
connected to a personal computer via parallel port. The patient in the lab setting. It is important that a patient’s
patient would be able to interact with the program data must be encrypted so that it remains confidential.
enabling them to perform their tests at home. The In addition, the data collected and sent to the hospital is
implemented system is very user-friendly for sustained to be used under the doctor’s discretion only.
use by patients. Due to the many different types and
stages of nerve diseases, each system must be tailored 4. Experimental Results
for the individual patient and their condition at a
certain time. The physician based on patient progress The effectiveness of the device is tested with ten
could easily change the system settings and exercise numbers of physically wealth people in each category
routine. The flowchart shown in Figure.5 encompasses based on age and mean and standard deviation are
the general idea of what a neurolysis patient would computed and presented in the table.1.The normalized
experience while running the Visual Basic program. results are stored in the look up table. The result
First, the muscle strength of the patient is determined obtained during execution of the muscle strength-
using a strain gauge based muscle force transducer. testing system for the patient with neurolysis compares
Then the output from the strain gauge transducer is the look up table and displayed their present condition.
intensified using an instrumentation amplifier. This The performance of the setup has been evaluated with a
improves the processing capabilities of the signal for 51 years old male sclerosis patient and the interactive
further stages. The ADC interface used here is screens in similar to the screen shown in Figure.7 were
transmitting the patient data to the Personal Computer. obtained.
START

Enter your login, password and age

Is correct?
No Sorry! Enter correct
login & Password
Yes

Screen greeting the patient

Operational Instructions to the patient

Are you ready?

Squeeze the ‘S’ beam as per the instructions

Figure.6 Photograph of the working model of the

Stop squeezing! system

Patient condition is displayed with advice The result obtained from the system is compared
[10], which shows the viability and consistency of the
scheme.
Do you
Want to try
again?
The executable screens encompass the general
idea of what the patient would experience while
running the program. The prompt asks for a login and
Thank you for taking the test password if it is not entered properly and the system
never allow any body to use the set up. This is to
STOP
protect a patient’s information, along with making the
interface tailored in name and test applications. The
Figure.5.Flow chart for the system execution data is also encrypted so others could not access a

173
patient’s testing information. Problems could arise in a results show that the proposed scheme could provide
patient forgetting the password or entering the good guidance to strengthen their muscles in spite of
password incorrectly. But some of their relatives or variations in the age and disease level. This testing
caretaker who took the patient to the hospital is system could be customized for any other specific
assumed to help them with the home test, along with therapeutic needs rather than this application. The
the assumption of a computer, which otherwise would author planned to design web based muscle strength
void the warrant of the former assumption. The screens testing system for bio-telemetering purposes along with
also tell the patient to keep squeezing until the system some additional features, will be presented in future.
stabilizes, but without such precise instructions, the
user may think the test is over. This screen lasts a few Acknowledgements
seconds. The screen should say stop when complete.
The First author expresses sincere thanks Mr.
Table.1.Voltage outputs for pinch forces for various R.Sundaramurthy, Lecturer, Dept. of Electronics and
peoples Instrumentation Engineering, Pondicherry Engineering
College, Pondicherry and his Under graduate students
MEN WOMEN Mr. B. Balachandar Mr. R. Srikanth and Miss. T.
Thirumagal @ Vijayamala for their extensive effort for
AGE Hand Range in Range in the successful completion of this work.
Volts Volts
R 0.54 –1.13 0.36-0.91 References
30-34
L 0.45-1.22 0.32-0.77
R 0.54-1.22 0.36-0.86 [1] Mayr, W.T., and Rodriguez, M. Multiple
35-39
L 0.45-1.09 0.36-0.73 sclerosis update. Minnesota Medicine 85(6): pp
R 0.5-1.13 0.23-0.68 36- 92 - 2002.
40-44
L 0.54-1.13 0.27-0.77
R 0.54-1.36 0.41-0.86 [2] D.C.Laycock, ”Theory and application of
45-49 Magnetic Field therapy in Multiple Sclerosis”,2nd
L 0.54-1.27 0.32-0.82
International Conference on
R 0.5-1.1 0.41-0.82
50-54 Bioelectromagnetism, pp105-106,1998.
L 0.54-1.18 0.32-0.73
R 0.5-1.1 0.41-0.73 [3] B.T.Fay,M.L.Boninger, Aspects of fatigue in
55-59 Multiple Sclerosis during manual wheelchair
L 0.45-1.18 0.36-0.59
propulsion,proceedings of the 25th annual
international conference of the IEEE EMBS, pp
1606-1608 ,2003.
[4] Ernest O.Doeblin, “Measurement Systems
Application and Design”, McGraw Hill
International Edition, 2000.
[5] Robert.B.Northrop, Introduction to
Instrumentation and Measurements, CRC press,
New York, 1997.
[6] IC manual, Analog devices Inc, release, 2002.
Figure.7. Result displayed for a 51 years old patient [7] Analog products Data Book, National
Semiconductor –2004.
5. Conclusion
[8] IC Data Manual, Maxim Inc.,- 2003
The new client based low cost muscle strength [9] Wang, Wallace. Visual Basic 6 for Dummies.
testing device is designed and implemented for patients NY, NY Hungry Minds, Inc, 1998.
with neuroleptic diseases like multiple sclerosis,
paralysis etc., The developed hardware cum software [10] V. Mathiowetz,et al, Grip & Pinch Strength:
package with sequence of screens guide the patient Normative Data for Adults, Arch Phys Med
through the complete testing process. The highly Rehabilitant Vol-66, pp 69-72, 1995.
interactive screens make it easy for the patient to
navigate through the testing process. The experimental

174
MEDICAL STANDARDS
 OUR OWN HLX FOR TELEMEDICINE
Dr Shashi B Gogia, Dr Suman B Bhattacharya
Correspondence address: 28/31 Old Rajinder Nagar, New Delhi 110060
Tel 09810126883, E mail sbgogia@amlamed.com
Web www.amlamed.com

Abstract Availability of common record keeping processes will

allow applications to seamlessly exchange messages,
There is an urgent need to create inter thereby ensuring faster and pain-free adoption of EMR
connectivity standards for Telemedicine purposes. systems as well as Telemedicine.
With International Standards like HL7 being unsuitable
for Indian Purposes, Indian Standards have been With such an intent, a common Telemedicine structure is
proposed. A preliminary proposal for the basic being prepared to cover all centres in India by the Indian
database and proposed direction of information flow. Association for Medical Informatics (www.iamindia.org).
The authors were officially appointed for this preparation.
This paper is being presented more as an approach
Keywords document for consensus among our esteemed members.
HL7, Medical Standards, Telemedicine, Patient (Lot of work has already been done. This will be
Database, India presented partly or completely if enough time is allotted.)

Introduction Some Background Info

Western medical facilities had been computerized very 1. Medical record keeping in general is pathetic in most
early, mostly due to pressure from payers like the healthcare providing institutions, barring a very select
insurance companies. However, individual software few. This has led the Medical Council of India to
vendors (ISV) made their applications that were unable to issue particular guidelines making it mandatory to
exchange messages with each other due to lack of keep records (preferably on the computerv)
messaging standards. Consequently, this vendor-based
approach made data interchange very difficult and led to 2. Telemedicine has been found to be cost effective in
creation of standards. HL7i (www.hl7.org) is a very reducing healthcare costsvi, not only in developed
notable example of such an effort and is now recognized countries but also in the developing worldvii.
as the de facto worldwide standard, its version 3.0 having 3. Currently in India, EMR and Telemedicine are
been accepted as an ANSI standard in late-September traveling on two different and parallel streams. The
2004 by USA. ISOii certification is also expected in the occurrence of two separate and parallel societies to
very near future with CENiii having already adopted the cater to these is just a pointer.
same for its efforts in medical informatics.
4. Telemedicine is more successful in countries where
Unfortunately, HL7 has been found unsuitable for Indian healthcare initiatives are state-funded (like the
needs, mostly due to the requirement of extensive amount European Union, countries belonging to the Eastern
of use of the localization option. This has resulted in an Block and the erstwhile Soviet Union). In the US,
effort that is akin to “re-inventing the wheel” that is except for Teleradiology, attempts to incorporate
suitable for Indian conditions. telemedicine in day-to-day health care have been
impeded by lack of adequate insurance cover, as the
There have been voluminous discussions on these issues insurance companies do not reimburse for
in IAMI forum as well as others where a general consultations carried out using Telemedicine.
agreement has been reached on the following lines Although, recently BCBSviii has started some
supportix.
We need to kick start Electronic Medical Records
(EMR) for all 5. Tight integration of EMR with Telemedicine has
Standards for such software have to be established helped growth of both.
Our standards need to be different from those set by 6. In India Telemedicine has been successful
other countries.
In the creation of satellite units (Narayan Hridualaya
The need for standards has been identified as a crying is a notable example)
need of the hour. While the Ministry of Information Tele-Education (SGPGI Lucknow with Cuttack
Technology, Government of India, has already created an Medical College)
approach paperiv, we are trying to go further from there.

175
 Specialist’s interaction e.g. THE APPROACH AND DIRECTION
http://www.telepathologyindia.com,
A patient’s record can be broken into small discrete
plastic_surgery@yahoogroups.com
recognizable parts, most of which can be transmitted
7. So far in India, to the best of available knowledge,
electronically (See Figure I)
EMRs are not integrated with currently available
applications for Telemedicine, unless they happen to Figure I
be made by the same ISV. Even here, the data
exchange is made possible by the use of the same
data tables. Breaking the (Pt) record
8. A comprehensive EMR is lacking in most hospitals.
History
• History • Text
Computerization in most hospitals is so far restricted
to (in order of usage) • Examination • Images
Administration • Investigations
• Video Conference
Billing and Accounting – Lab
Diagnostic Reports and other Laboratory Services. – X Ray/US/CT/MR • Sounds
Inventory and drug management – Histo/Cytopath
• Video Clips
Patient tracking (reception services) – Others
Presentations (By individuals care providers – • Invasive Procedures • Wave Forms
doctors and nurses, not the hospital as a whole) 1

Though a large number of Medical Professionals own Breaking the (Pt) record
computers and have access to the Internet, only a small
percentage use it for professional purposes, with only 5% • History • Text
or less using any form of Software for EMRx. • Examination • Images
• Investigations
Besides the problems of the unwilling user, we have the – Lab
• Video Conference
additional problem of lack of good software that addresses – X Ray/US/CT/MR • Sounds
the various requirements that our doctors have. It is of – Histo/Cytopath
course only to be expected. Software manufacturers are • Video Clips
– Others
used to being paid in dollars. The average total cost of an • Wave Forms
• Invasive Procedures
appendix operation in India is between Rs 5000/- to
12000/- in a reasonable facility. In the US, a surgeon will
charge the same amount in dollars and the hospitalization
costs would be additional. Indian doctors are unable to
pay the expected high license fees that the ISVs demand,
One main lacuna in existing telemedicine applications
and we (according to the Software professionals) get what
has been the lack of means of sending the history,
we pay for. Additionally, most EMR and Telemedicine
software available in India so far has been buggy and full
findings, and prescriptions already provided (as a
of installation and implementation issues, so some of the review source for treatment given). We are trying to
braver institutions that went in for the same had to correct this anomaly.
discontinue, AIIMS being a notable example.
General Details
Telemedicine networking is being hindered by a number The following is the general suggested flow for the
of problems. Some of these are: proposed Telemedicine record-keeping mode
The general apathy of doctors towards Information
Technology All data to be maintained through a Relational
Non-availability of good (read user friendly) Database Management System (RDBMS)
applications for use by doctors Individual solutions can be created by multiple
The non-paying attitude of most doctors for purchase vendors using the SQLxi agreed to
of software. (Drug companies are known to subsidize A separate Telemedicine application to be run at
plenty of our needs – not only of medicines but also
all centres. This may be stand alone or as an
journals, books, travel, hotel stay and some
addition to their existing EMR or other patient
household goods too. So paying for anything,
especially what is a virtual product, does not make
record system – preferably the latter
sense to them) All users are required to use this database table
Many companies making medical software have been structure for Telemedicine, which can either run
known to fold up as a result of lack of sales and non- separately or as part of their existing tables with
viability. separate user qualifications
Of the companies that are providing solutions, the Data interchange to be carried out in XML format
approach has been individualistic and uncoordinated. As a Where the database does not support SQL, the
result, the developed applications are unable to exchange XML file containing the data will be used for data
data with each other. storage, display and exchange. However SQL
based creation will lead to better and more robust
applications

176
 In case the database table names clash with can be developed using any platform like Java (J2EE),
existing tables, alteration of the table names is ASP, C++, VB, PB, .NET, etc., the database must to be an
permitted but the structure and sequence of the RDBMSxiii that supports SQL. Not only the column names
columns shall be maintained at all times but also their data types will conform to the proposed
Store and forward method shall be available for format
these common tables
Data interchange will be by using the “Export” A good start would be the basic patient card. An example
functionality from the sender’s end and the of the same on which we need to build upon is the all
purpose smart card proposed by the government that
“Import” functionality at the receiver’s end
allows access for elections, driving license, ration card
The exact structure will require some discussion
etc. (TCS is making the same). Some items required for
between existing vendors but once agreed should Telemedicine needs can be added, e.g., Telephone No.,
be set as THE STANDARD Blood Group, Diagnosis, Allergies, Special Problems,
Images and other waveforms like ECG, EEG, etc. Current Medications, etc.
shall be transferred separately as linked files. The
URLxii of the linked file will however be The primary key for this table should be such that fresh
exchanged as part of the XML file (It is up to the records may be creatable at the doctors end without
individual vendors to decide as to how to store and knowledge of his State appointed code – whether ID or
later access the same) Election card, etc. Although this may result in data
duplication at various physicians’ ends, there will be
plenty of times when such data may not be accessible by
the treating physician, and the patients’ treatment cannot
Flow Diagram be made to wait till end. A unique set of National Care
Providerxiv Identifier like, e.g., a social security number,
will result in unique records for a patient being created all
Doctor's Local Database Telemedicine App over the country.
Facesheet Facesheet
Local transfer
Casesheet Casesheet
Prescription
For Referral
Prescription (More details of this paper have already been sent and
X Ray/US/Lab--- X Ray/US /Lab--- accepted for publishing as part of the IAMI Journal -
Comments
the issue carrying the same will be presented in the
XML/Text/ conference)
Dbf Format

Expert End Database Received in Expert End

Facesheet Facesheet Health Level 7
If patient 1
Casesheet
physically seen
Casesheet
Prescription
International Standards Organization
Prescription 1
X Ray/US /Lab/MRI/CT X Ray/US/Lab Comité Européen de Normalisation or European
Comments Committee for Standardization
1
http://mit.gov.in/health/telemedicine/index.pdf
1
Benefits of this sort of flow Hindustan Times, Delhi Edition, April 2002
1
• Maintains privacy and security of data Richard Wootton. Telemedicine, BMJ 2001; 323:557-
• Records exported only if required. 560 (8 September)
1
• Direct copy of record in Expert Location Vassallo DJ, Hoque F, Farquharson Roberts M,
• Simplifies record creation. Patterson V, Swinfen P, Swinfen R. An evaluation of
• Records co-located across various centres the first year's experience with a low cost telemedicine
• Prevents server overload link in Bangladesh. J Telemed Telecare 2001; 7: 125-
• Does not tax the communication network
138
• Works even if data links are down 1
Blue Cross and Blue Shield
1
Telemedicine WMJ
Data needs to be interchanged (i.e. Export /Import) 1
Rough figures based on information sourced from
in XML file format associates and friends
All data will be created locally in the user’s 1
Structured Query Language
own database. (Let us assume he is the 1
Uniform Resource Locator
Primary care Physician) 1
Relational Database Management System
On selection of a particular patient for 1
Individuals like doctors, dentists, or organizations like
referral, his/her data will be transferred as hospitals, nursing homes, etc.
depicted in figure 2 below

Figure 2

The ultimate aim is to create a common database structure

from which all applications will be able to interchange
data. Our recommendation is that while the applications

177
 IMPLEMENTATION OF HL7 STANDARDS IN TELEMEDICINE
ENVIRONMENT
Pramod Adiddam, Suman Kundu, J. Mukherjee
Department of Computer Science & Engineering, Indian Institute of Technology, Kharagpur.
e-mail:jay@cse.iitkgp.ernet.in

Abstract heterogeneous systems becomes too site-specific and

manual.
With the advances in computer and The HL7 is an application protocol for electronic
communication technology, health care industries have data exchange of information in healthcare
naturally given considerable attention to the environments. HL7 standardizes the protocols and
development and maintenance of medical information structure for exchanging key sets of health data at the
systems. Hospital information systems (HIS) maintain application level. XML [3] is the latest upcoming
medical information related to patient in a suitable technology to be utilized to the maximum extent by
database. Along with the development of HIS, health care services. The main advantage of XML is its
telemedicine has ushered a new way of delivering ease of availability of parsers and its efficiency in
health care service from distant locations. The distributing and aggregating all kinds of content in the
Telemedicine system needs to integrate patient HL7 messages. A message in XML means that each
information from different departments of a hospital, piece of data - a list of observations, locations or
such as radiology, pathology, Diagnostics, admissions descriptions is coded separately. In this way, an
etc. Telemedicine involves very high level of diversity available parser will know exactly what to gather and
of knowledge. Different Telemedicine systems process. Another standard widely popular is the Digital
maintain there own database schema for storing data Imaging and Communications in Medicine (DICOM)
and follow there own protocol for exchanging message. that is used for diagnostic images, waveforms and
Therefore exchange of information between diagnostic reports.
Telemedicine systems, developed independently by
diverse set of people need standardization of medical We have developed a HL7 message generating
terminologies and a common message exchange module and integrated it with our existing
protocol. To support interoperability in health care Telemedicine system. We have taken a client and
environment Health Level Seven (HL7) messaging server approach where servers and clients are located
standard has been proposed. This paper presents our in different health care environment and only the
approach in the development of a solution for the servers are exposed to the outside net. Servers generate
interoperability of Telemedicine systems based on HL7 messages and send it to the clients connected to it.
HL7. The clients receive the HL7 message and convert it into
its own message format. If the message is successfully
Keywords: converted the client stores it in the local database
Telemedicine, HL7, Medical Messaging Tele- otherwise it convert and keep the message in XML
consultation, Hospital Management, Database. document format. Aim of this paper is to explain the
integration of HL7 standard with Telemedicine system.

1 Introduction 2 System Architecture

Telemedicine is the use of telecommunication The existing telemedicine system have a medical
technologies to exchange health information and database which provides patient administration data
provide health care services across geographic, time, and extensive observations data used by hospital
social, and cultural barriers. Telemedicine system administrative systems, clinical laboratory systems,
stores patients’ personal and medical information in a diagnostics systems, and imaging systems. Tables in
database in the form of text, image, graphics, audio and the database store patient information acquired from
video. All the personal and clinical informations are various departments of a hospital.
stored in structured form in different tables in the
database. Each of the form data contains fields whose We have developed a system which generates
naming or referential dependencies are different from HL7 messages taking data from the existing database
HL7 standards but are independent of other forms. and sends it to another telemedicine system. A client
Diverse systems have their own output format, server based approach is taken to implement this
terminology, database layouts and application system. The server provides query based service and
interfaces. Hence data exchange between unsolicited message service to the clients form other

178
telemedicine system. The client gets the HL7 message server about the transaction. Server picks up the
and stores it in its own database by converting the HL7 message and creates a HL7 notification, i.e., an
message to its own format. The client stores the data in unsolicited message and sent it to target clients that
XML format if it can not be inserted into the database. have or will make a connection to that server. Target
systems receive messages online or in bulk when they
connect Messages are filtered according to the target
systems classification.

Fig 1. The Flow of messages in the network.

Fig 2.Unsolicited Message Flow.

2.1. The Client-Server Model

Hospital environments are by nature distributed 2.3 Query/ Response Processing

where the total workload is distributed among different
departments. In most of the cases, these departments A client can generate HL7 query using an interface and
are physically separated. We have divided the system send it to the server it is connected with. The server
into five broad departments as above: admin, labs responds with a result set matching the query criteria or
(pathology, haematology etc.), imaging (MRI) and an error message. It server do not have the desired data
radiology, diagnostics and finance. Each healthcare then the query is passed along the network by running
node in the network hosts one or more servers Query Hopper which passes on the query to the clients
depending on the number of departments it has. A which are connected to the particular server. For
server in the system ‘serves’ a single database and example in a hospital if we have three machines, one
accepts database Trigger message and ADT Message with admin data, another with diagnostic data and a
from client side. The server delivers messages third with financial data residing on them, then a
depending on whether it is in Diagnostics, Imaging, finance query sent to the admin server needs to be
Admin and Labs. For this purpose messaging is done redirected within the hospital itself to the finance
on a single ‘channel’ specific to departmental server. This is done by query hopper. Query hopper
messages. Client cannot be always connected. Hence a assumes that a client on the finance machine is already
facility for bulk delivery of messages has been connected to the admin server. Now since all three
provided for when the client connects. machines are in the same hospital, there will be no
extra cost for arranging a network where a client on the
diagnostic machine and a client on the finance machine
2.2 Unsolicited Messages are connected to the admin server. Also occasionally
finance data references to admin data. Hence we can
Each healthcare node also supports a host of justify the network thus setup in the hospital and
clients. Client connected to a pre specified port of a leverage upon it for queries by using query hopper. The
server receives unsolicited updates and are server would respond to client in the HL7 format if the
acknowledged. Whenever an insert, update or delete query matches with its local conformance data format.
operation takes place in a database from any interface a
trigger event runs a client named ‘Unsolicited Message
Client’, which picks up that event and notifies the

179
3. Message Conversion 2. A query message received by a server from a client
3. An unsolicited message received by a client from a
The proposed system supports any HL7 2.4 document server.
[1], integrated to the existing database. Each server’s
knowledge base contains a subset of the ontology On reception of such a trigger the translator
mappings (from local to HL7 fields and vice versa) invokes the message and data libraries to generate a
library relevant to the local database and the types of message associated with that particular trigger.
messages it processes. We have developed an Ontology Whenever a data entry takes place in the database a
server that parses and serializes HL7 messages into Trigger of type-1 is fired which passes a token to the
segments, fields, repeated fields, components, and server. The server generates unsolicited message and
subcomponents. passes it to the client. Trigger type-2 generates a
response message and sends it to the client when client
query for some data. Trigger of type-3 generates an
acknowledgement message and sends it to the server
after receiving an unsolicited message from the server.
The corresponding message (in the form of a tree data
structure) received from the HL7 message libraries is
converted into a text HL7 message and sent across to
the respective nodes depending on the network
architecture. The HL7 messages generated by the
server follow a text-based standard and use three
primary concepts: segments, fields and components. A
message consists of one or more segments, each of
which contains some number of fields. Each field have
zero or more components.

An example of OBR and OBX segments [6] for

depicting a surgery pathology report from a lab
regarding normalcy of a gal bladder is as follows:

OBR|1||1234^LAB|88304&SURG PATH REPORT|...

OBX|1|CE|88304&ANT|1|T57000^GALLBLADDER^
SNM|...
Fig 3. Ontology Server.
OBX|2|TX|88304&GDT|1|THIS IS A NORMAL
GALLBLADDER|...
3.1 The Ontology Server OBX|3|TX|88304&MDT|1|MICROSCOPIC EXAM
SHOWS HISTOLOGICALLYNORMAL
The representation of the constituent databases is
GALLBLADDER TISSUE|...
stored in the ontology server. The ontology server is
comprised of the following components:
The second task (search) of the translator is to search
• HL7 Translator through the data libraries for necessary data. This
• HL7 Message Library is required in two situations
• Data Mapping Library 1. An unsolicited response message is received
from a server by a client
3.1.1 Hl7 Message Translator 2. A query message is received by the server
from the client
The HL7 translator is responsible for the syntactic When the type-1 messages are received, the data
operations (required for generation of messages from in the message is to be stored in the local DB. The
the tree data from message/data libraries) and also for search technique is used to find out the corresponding
search operations (required for extracting data in the schema tables to use for storage is detailed in the data
local database format from received messages and for library section. When type-2 messages are received the
generating responses to queries). It also converts all same search technique as above is used but in this case,
received messages into a long name format HL7 XML it is used to determine the schema table and the
representation, provided their data can’t be converted message to be generated in response.
into local database schema.
The third task (syntactic) of the translator is to
The first task (syntactic) of the translator is to receive convert the type-1 messages from above which
triggers. These triggers are of three kinds. couldn’t be stored in any schema table, into XML file
format. This is done easily by converting the message
1. A token received by a server from the local into a tree format and then converting that tree into a
database due to a SQL trigger getting fired. DOM tree with appropriate DOM node names looked

180
up from three files containing HL7 segment, field, 2 respectively. All the three words (1,2,3 or 1,5,3 or
datatype names. The system converts all HL7 message 4,2,3 or 4,5,3)(3 is optional) should be present in OBR
to XML format for the easy readability of the message or OBX descriptors(OBR-4-2 and OBX-3-2). For
using helpful tags for the elements. It generates XML example if the vector is (past, history, NULL, NULL,
schemas to all HL7 messages, segments, fields, and NULL) then messages for "EYE PAST
data types and validates with XML DTD. A DTD is HISTORY"(i00) and "GENERAL PAST
used for accessing definitions of the implemented HISTORY"(h04) will match creating confusion. Hence
fields, data types etc. It takes care of the relationship here the vectors should be (eye, past, history, NULL,
between message types, event types, and structure of a NULL, NULL) for i00 and similarly for h04 the
message. Also it provides data for default values and vectors are (general, past, history, NULL, NULL).
allows user-defined variations to any part of HL7 – for When a new application is added to the interface, all
example, making a required field optional. Hence only the key fields in its associated database and their HL7
required fields are sent and the fields not present in the mappings are added as its base set of mappings in the
local database are stored as XML files. library.

3.1.2 Hl7 Message Library 4. Conclusion

The HL7 message library is based on the HL7 The obstacles faced for heterogeneous data
message structure. Information about all the HL7 exchange are that there are no standards. Applications
messages up to the field level - data type, alternate have to use complex protocols or put in high effort to
database field names, data patterns, if any, - are build one-of-a-kind interfaces. Hence the challenges
represented in the HL7 message library. The library is faced in healthcare integration is that it is currently
a collection of classes. Each class builds a particular expensive, time-consuming, risky, intra-provider and
message. The structure of the message is built into the not inter-provider and dependent on private networks.
class. Each message class has the objects in it, each
corresponding to a different sub-structure of the The solutions to these challenges cannot solely be
message. Some of the message segments are web based as it is highly unstructured, insecure, and
dynamically constructed, i.e., their number in the human-centric. Hence the simple credo of the internet
message and their individual contents are determined plus next-generation integration technologies such as
dynamically using support tables in the database. A XML for message structure and HL7 for message
particular table in the database named ‘formfields’ content is the solution for clinical information systems
maintains the mapping between local database interoperability
terminology of each table and HL7 terminology with
full support of segment number, code, unit and data
5. Results
type of each field. For example, an entry in
‘formfields’ table for a field of Routine blood test & There are two centres participating in the network
grouping is (ioo, wbc_cnt, LEUKOCYTES, 3001, • An administrative system
6690-2, 10*3/ml, varchar (50)) The ‘formOBRS’ • A diagnostic system
table keeps record the OBR segments with their value The systems in this trial run exchange the indicated
for all database tables. For example Complete blood messages as shown
count form is divided into two segments other blood
count and blood count report having segment no 3022
and 3023 respectively. Any change in the local
database is hence reflected easily without changing the
code of the application by just altering the database
entries in the support tables. Some message segments
can’t be changed without affecting a lot of front-end
applications. These can be coded into the application.

3.1.3 Data Mapping Library

The data mapping library contains mapping

information about HL7 and database fields as well as
information on key fields. Mappings can be added,
deleted or modified by a user holding the appropriate
privileges in the database table ‘formfields’. In
‘OBXVectors’ table the vectors for data not supported
by the database is to be entered. There are five
keywords in the vector. 1-3 are independent of each
other, 3 is optional and the rest 4-5 are synonyms of 1- Fig 4. Patient BCR01042003000 is Admitted

181
 converted into HL7 Reference Information Model
(RIM) which is completely based on XML. Hence
further avenues for non-redundant information transfer
open up with this model.

References
1. HL7 Version 2.4 Documentation and HL7
Mailing Lists. http://www.hl7.org/about/
2. Bergman, Dale W. Telehealth Interoperability
Laboratory. IEEE 1073, Salt Lake, October,
2001.
3. The Apache XML Project.
http://xml.apache.org/xerces-c/
4. Ramesh, V.; Canfield, K.; Quirlogico, S. and
Silva, M. “An Intelligent Agent-based
Architecture for Interoperability among
Fig 5. Diagnostic Receives A01 Message Heterogenous Medical Databases”. The Americas
Conference on Information Systems. 1996.
5. Vargas, Bill. Ray Pradeep. “Interoperability of
6. Future Work Hospital Information Systems: A Case Study”.
IEEE. 2003.
The present system only supports textual information
exchange. We are presently working on the transfer of 6. Clay, E, Willams. “A Language for Generating
DICOM. The existing XML representation can HL7 Reformatting Program”.
be

182
 HEALTHCARE DATA INTERCHANGE STANDARD
- HEALTH LEVEL SEVEN (HL7)
Sudhir Agarwal & Ravi Saksena
Space Applications Centre, Indian Space Research Organization (ISRO)
Ahmedabad, India
sudhir@sac.isro.org & rsaksena@sac.isro.org

Abstract in a diverse system environment to communicate with

Over the past 10 to 15 years, many data each other.
interchange (interface) standards have emerged to
address the growing need of healthcare industry. The 2. HL7 Background
purpose of HL7 standard is to facilitate communication HL7 is a Standards Developing Organization
in healthcare environment. HL7 is a Standards accredited by ANSI to write standards representing a
Developing Organization, accredited by ANSI consensus of various entities in the healthcare area.
(American National Standards Institute) to write
standards, representing a consensus of various entities The HL7 standard attempts to integrate a variety
in the healthcare industry. of technical and administrative environments. It
supports IT environments ranging from a fully OSI-
The HL7 Standard operates at seventh layer of compliant network stack, to older point-to-point
ISO standard and defines the messages for exchange interconnections. It also supports data transfer in
among application entities and the procedures used to various media, including floppy disk and tape. The
exchange them so as to make them interoperable. In HL7 Standard was developed in close cooperation with
India, ISRO has specified HL7 as interface standard for other standards organizations including ACR/NEMA,
Satellite based Telemedicine network. This paper DICOM, ASC X12, ASTM, IEEE/MEDIX, NCPDP
describes HL7 Background, Communication, HL7 and others.
message and its relationship with other protocols. Originally created as an intra-hospital
communications standard, HL7 has grown up and has
KEYWORDS: Telemedicine, healthcare, HL7, ADT become the undisputed, de-facto standard for the entire
healthcare industry. The HL7 standard is primarily
1. Introduction focused on the issues that occur within the seventh, or
Over the past 10 to 15 years, many healthcare application level. The HL7 standard defines methods
data interchange standards have emerged to address the for the exchange of clinical, financial, and
growing need to cut cost, improve the efficiency and administrative data among healthcare oriented
delivery of healthcare at the patient/clinical level and to computer systems. HL7 is based on conceptual idea of
improve the effectiveness of healthcare administration. what is called a "trigger event," a real word event that
Established in 1987, Health Level Seven (HL7) has causes the need for patient data to be exchanged
emerged as a predominant standard. between systems.

The purpose of HL7 is to facilitate HL7 is a messaging syntax that defines the
communication in healthcare environment. The messages which different systems will send in order to
primary goal is to provide standards for the exchange communicate with each other. The standard specifies
of data among healthcare computer applications and types of messages that relate to various functions found
making them interoperable, eliminating the custom in a clinical environment. For example, the message
interface programming. type ADT (Admit/Discharge/Transfer) is used to
communicate admissions data about patients. The
The HL7 standard currently defines the interfaces messages themselves are composed of segments that
among various systems that send or receive patient are in turn composed of fields. The definition of each
admissions/registration, discharge or transfer (ADT) type of message specifies which segments it contains
data, queries, resource and patient scheduling, orders, and in which order they will occur. The message
results, clinical observations, billing, medical records, definition for an ADT message will specify that it may
scheduling, patient referral, and patient care. It does contain certain segments, such as a PID (Patient
not try to assume a particular architecture with respect Identification) segment. The order of the segments and
to the placement of data within applications but is fields is specified, and also the rules for repetition or
designed to support a central patient care system as optionality of a segment or field. At the data level,
well as distributed environment where data resides at several data types, such as address, telephone number,
number of places. HL7 serves as a way for inherently patient name, and coded entry are defined which may
dissimilar applications and data architectures operating be used in any message.

183
Trigger Events catheterization application, there may be a trigger event
a procedure is scheduled for a patient who is not
The HL7 standard assumes that an event in the
registered in the cardiac catheterization application’s
real world of healthcare causes the data to flow among
database. The application may send a request message
systems. The real-world event is called the trigger
containing the patient’s ID number to the Patient
event. When a patient is admitted, a trigger event is
Administration (ADT) system and receive a response
generated, which may cause the need, for data about
containing the necessary data to permit processing of
that patient to be sent to a number of other systems.
the order. This requesting transaction is a query. The
Similarly when an observation for a patient is
information that flows between the systems is
available, a trigger event is generated, and may cause
contained in the response.
the need for that observation to be sent to a number of
other systems. When the transfer of information is
In all cases, the HL7 Standard consists of a
initiated by the application system the transaction is
simple exchange of messages between a pair of
termed an unsolicited update.
applications: the unsolicited update and its
acknowledgement or the query and its response. An
HL7 allows the use of trigger events at several
application interfaces with another application using an
different levels of inter-relationships. For example,
event code that identifies the transaction. The other
most Patient Administration (ADT) trigger events are
application responds with a message that includes data
concerned with single objects (such as an admit event
or an error indication. The initiating application may
about a single person and/or account). Other ADT
receive a reject status from the other application
trigger events are concerned with relationships between
indicating that its message was not received correctly.
more than one object (e.g., the merge events, which
specify patient or account merges). Some ADT trigger
2.1 HL7 Communications
events pertain to a collection of objects that may have
no significant inter-relationships (e.g., a record- The HL7 Standard is primarily concerned with the data
oriented location-based query, whose response contains content, interrelationship of messages and application-
data about patients who are related by local level error conditions.
geography).
Since the OSI protocols are not universally
Acknowledgements: Original Mode implemented, the HL7 provides for communicating
This acknowledgement mode specifies that when health data among systems operating in
the unsolicited update is sent from one system to communications environments that provide a high level
another it should be acknowledged at the application of functionality, but use protocols other than ISO OSI.
level. Because, it is not sufficient to know that the It supports following:
underlying communications system guaranteed
delivery of the message, but it is also necessary to Ad hoc environments that do not provide even
know that the receiving application processed the data basic transport reliability like point-to-point RS-
successfully at a logical application level. The 232 links.
acknowledgement may contain data of interest to the Environments that support a robust transport
originating system. For example, if a patient care level, but do not meet the high level requirements
system has processed the trigger event ‘a lab test like TCP/IP, DECNET, and SNA.
ordered for a patient’, it may send an unsolicited ISO and proprietary networks that implement
update to a lab application identifying the patient, the up to presentation and other high level services.
test ordered, and other information about the order. Two or more applications running on the same
physical and/or logical machine. In such
Acknowledgements: Enhanced Mode environments, the messaging capabilities may be
provided by inter-process communications
The HL7 acknowledgement paradigm has been
services like Pipes in a UNIX System.
extended to distinguish both acceptation and
application acknowledgements, and the conditions
The HL7 Standard assumes the communications
under which each is required. With a positive
environment will provide the following capabilities:
acceptation acknowledgement, the receiving system
Error free transmission Applications can
commits the message to safe storage so that the sending
assume that they correctly received all of the
system need not resend the message. After the message
transmitted bytes in the correct order that they were
has been processed by the receiving system, an
sent. This implies that error checking is done at a
application acknowledgement may be used to return
lower level.
the resultant data to the sending system.
Character conversion If the two machines
exchanging data use different representations of the
Queries
same character set, the communications environment
A different data exchange occurs when one will convert the data from one representation to the
system sends a query to another. In a cardiac other.

184
 Message length HL7 sets no limits on the component of another data type that itself contains
maximum size of HL7 messages. The Standard multiple components. Some segments allow fields to
assumes that the communications environment can repeat. Where this occurs, repetitions of data are
transport messages of any length that might be delimited by the repetition separator. Because
necessary. repeating fields might be of data types that have
multiple components, a system receiving an HL7
message should separate each field into its repetitions
2.2 HL7 Message
before breaking the individual repetitions down further
HL7 Standard contains rules that are used by into their components and subcomponents. An example
originating systems to build messages and by of an HL7 message appears below.
destination systems to interpret the message data they
receive. The rules deal with the formation and MSH|^~\&|Urology|CivilHosp|UroloDept|AIIMS|2004
communication of messages in general. These rules are 0529090131||ADT^A01|01052801|P|2.3.1
applied in forming the message definitions, query and EVN|A01|200405290900||||200405290900
acknowledgement messages. Examples of an event PID|||567837445^^^UAReg^PI~999855750^^^INSA^
would include a pharmacy order, a laboratory results SS||GUPTA^BINDU^M^||19670307|M|||10 NEEL
report, generation of a charge for services, a master file KARMA
update or a query. FLAT^SHAHIBAUG^AHMEDABAD^^^^|||||||0105I3
00
The message is a collection of segments or PV1||I|W^389^1^UABH^^^^3||||12345^SAKSENA^M
“lines” of information. Data in each segment are OHAN^J^^^MD^^^A^L||67890^SHARMA^RAJU^X^
delimited which allows for elements of varying lengths. ^^MD^^^A^L|MED|||||A0||13579^DAS^KAPIL^T^^^
HL7 defines six delimiters for use in standard message MD^^^A^L
formation: OBX|1|NM|HT^HEIGHT^99LOC1|HEIGHT|71|in^inc
hes^ANSI+|||||F
- the field separator (generally vertical bar, | ) OBX|2|NM|WT^WEIGHT^99LOC1|WEIGHT|175|lb^
- the component separator (generally caret, ^ ) pounds^ANSI+|||||F
- the subcomponent separator(generally AL1|1|DA|ASP^ASPIRIN^99LOC2|MO|GI
ampersand, &) DISTRESS
- the escape character (generally backslash, \ )
- the repetition separator (generally tilde, ~ ) The example shows the format of the message.
- the segment terminator (always ASCII hex The segment terminator at the end of each segment has
0D, carriage return) been converted into a carriage return/line feed
combination. This message contains seven segments:
Except for the segment terminator, which must
one message header (MSH) segment – every
always be ASCII hex 0D, systems may use any
message without exception has exactly one MSH
printable characters as delimiters. The delimiters used
segment at its beginning
in a message are defined in each message instance.
one event type (EVN) segment, which is used
Each segment of a message is composed of one or
in administrative message to communicate event
more fields which are separated by the field separator
type information
delimiter. Field lengths usually may vary up to a
one patient identification (PID) segment
defined maximum. The definition of the segment and
one patient visit (PV1) segment, carrying
of its fields remains the same in any message in which
information relevant to this particular encounter
the segment is used. Each field is of a defined data
two observation/result (OBX) segments, used
type. Data types define how information elements may
here to communicate vital signs collected at the time
be associated and interpreted. Many data types have
of the event
multiple components, which are delimited by the
one patient allergy information (AL1)
component separator.
segment
A segment is referred by the three-letter segment
Components allow for the expression of
ID that appears at its beginning e.g. “the MSH
individual elements of a composite unit of information,
segment,” “the PID segment,” and so on. The MSH
such as name (last, first, middle, etc.) or address (street,
segment contains important information that the
city, etc.). In many data types, components are also
destination system will use to separate the message into
essential for carrying validating or coding system
fields and process its data. The first critical piece of
information, such as the name of the enterprise that
information is the field separator, the fourth character
assigned a particular patient identifying number. Data
of the MSH segment. This message example uses the
types also may have subcomponents, which are
customary HL7 field separator, vertical bar ( | ).
delimited by the subcomponent separator.

Subcomponents are components within

components. These occur when a data type includes a

185
 Once the field separator is extracted (which HL7 b) The encoding rules of these protocols depend on the
defines to be the first field of the MSH segment), the assumption that lower level protocols provide
MSH segment can be broken into its remaining fields. transparency (i.e., all character codes can be
The second field of the MSH segment, which falls transmitted without being changed by the lower levels).
between the first and second occurrences of the field
separator, contains the message’s remaining delimiters Whenever HL7 is applied in a networking
(which HL7 calls encoding characters). A field is environment, addressing is an issue, as the Standard
referred by its segment ID and position within the does not specify how this addressing will occur. Fields
segment. The four characters of MSH-2 contain the such as MSH-5-receiving application, MSH-6-
following four delimiters, always in the order: the receiving facility, and MSH-11-processing ID, located
component separator, the repetition separator, the in the header of all HL7 messages may be used for this
escape character, the subcomponent separator. purpose.
Here, the customary values have been used for
Other Applications Protocols
these delimiters: caret, tilde, backslash, ampersand.
However, another message instance might use different a) ACR/NEMA DICOM : The HL7 Working Group
characters. Therefore, the receiving system must maintains an on-going liaison with the ACR/NEMA
extract and use the field separator and encoding DICOM working group.
character values that are carried in each individual
b) ASC X12 Standards for Electronic Document
message instance. MSH segment can be broken into its
Interchange : ASC X12 is a family of standards that
fields as follows:
provides both general and specific descriptions for data
interchange within a number of industries. The HL7
Field Name Value
Encoding Rules are modeled on the X12 standards,
1. Field Separator |
although there are differences.
2. Encoding Characters ^~\&
3. Sending Application Urology c) ASTM 1238.94 Laboratory Data Reporting : An
4. Sending Facility CivilHosp active liaison effort between the ASTM committee and
5. Receiving Application UroloDept the Working group has resulted in minor changes in the
6. Receiving Facility AIIMS ASTM and HL7 specification to enhance
7. Date/Time of Message 20040529090131 compatibility.
8. Security Empty
d) IEEE P1157 (“MEDIX”) : The MEDIX committee
9. Message Type ADT^A01
10. Message Control ID 01052801 is defining an application-level protocol similar in
11. Processing ID P scope to HL7 but built strictly on the ISO protocol
12. Version ID 2.3.1 stack. Despite the difference in approaches, the HL7
Working Group has regular liaison with the MEDIX
committee.
Drawing inferences from the data in this example
message: This message was generated by the Urology
system at Civil Hospital. It was sent to the Urology 2.4 Indian scenario
Department at the AIIMS. It was generated on 29 May In India, ISRO has taken up the task of
2004 at 9:01.31 a.m. Its message type is ADT establishing Satellite based Telemedicine network.
(Admit/Discharge/Transfer). Its trigger event is A01 This network presently has around 100 terminals and is
(Admit a Patient). It is a production system message likely to grow up to few thousands. HL7 is specified as
(as opposed to a development or training system an interface standard for this telemedicine network, in
message). Its elements are defined in Version 2.3.1 of order to ensure interpretability among different
the HL7. vendors. ISRO provides common platform along with
Ministry of Communication & Information Technology
2.3 Relationship to other Protocols (MCIT) and vendors to evolve components from HL7
standard for interchange.
A great deal of consideration is given to the
relationship between the HL7 Standard protocol and Telemedicine nodes, in this network, operate in
other protocols. client-server based model and patient data is
Lower Layer Protocols transferred between servers using HL7 format using
The HL7 Encoding Rules are substantially CD media or TCP/IP communication. Selected fields
different from the ASN.1 Basic Encoding Rules of patient demographics, hospital administration,
documented in CCITT X.409 and X.209 and ISO 8825. referral and allergy information are exchanged.
This is because: Provision is kept to incorporate 40 fields specific to
local hospital. It is possible to export patient data to
a) By definition, the HL7 encoding rules will be
applied where the environment does not include international healthcare entity also.
software to do so.

186
3. Conclusion
The HL7 standard defines methods for the exchange of References
clinical, financial, and administrative data among
healthcare oriented computer systems. The HL7 has
become de-facto standard for exchange of healthcare [1] HL7 Standards, Health Level Seven, Inc.,
information among clinical applications world wide. In http://www.HL7.org
India, for Satellite based network this standard is [2] Framework for Information Technology
specified for interface among healthcare professionals. Infrastructure for Health in India Volume I,
http://www.mit.gov.in/telemedicine/home.asp/
4. Acknowledgements
Authors are thankful to Sh M K Sharma, Head DCTD;
Sh K Bandyopadhyay, GD, SGSTG, Sh A R Dasgupta,
DD, SITTA for their support and guidance.

187
ABSTRACTS AND SHORT PAPERS

188
 STANDARDIZATION IN TELEMEDICINE IN INDIA:
INITIATIVES AND POLICY ISSUES

Dr B S Bedi
Senior Director, DIT, Ministry of Communication & IT
New Delhi

Abstract standards for practice of telemedicine in India has been

drawn up. The recommendations provide for setting up
In view of the present geographical and demographic a telemedicine center based on standards for systems,
spread of population and skewed distribution of telemedicine software, connectivity, security and
medical specialists Telemedicine assumes special privacy issues, data exchange, etc. It also covers
significance for India. Whereas two third of the telemedicine process guidelines for conducting
population lives in rural areas and about 90% of telemedicine interactions.
available Secondary and tertiary health care facility are
currently city based. Having the technical and medical
expertise capability to offer quality healthcare to The recommendations have been disseminated among
remote areas, a number of telemedicine centres have various stakeholders who have evinced keen interest.
been set up and are also being planned across the In addition to bringing awareness among various
country with initiatives from public and private sector. stakeholders and policy makers, it is also serving as a
To streamline establishment of telemedicine centres ready reckoner for judicious adaptation of the
and services, need for a set of standards and guidelines Information & Communication Technology for the
was felt. practice of telemedicine at different levels of existing
landscape of health care delivery system prevailing in
India. Steps to take forward the present initiative are
Under efforts of a specialized Technical Working underway of standardisation activity in Telemedicine
Group, set up by Department of IT, Govt. of India, a in India based on the suggested nucleus standard and
nucleus framework of recommended guidelines and guidelines are underway.

190
 TAKING TELEMEDICINE TO INDIAN VILLAGES THE WIMAX WAY

Dr D Lavanian
MBBS, MD Domain Expert – Telemedicine, Department of Telemedicine, Apollo Health Street Ltd,
Apollo Hospitals, Hyderabad, India

Abstract haphazard way. Today there are near two million care
providers in the country, consisting of a motley groups
Last mile connectivity has always been the of homeopaths, ayurvedics, electropaths, licentiate
Achilles heel of Telemedicine in India. This paper talks practitioners, health workers, … and various other
of a new emerging technology WiMAX that could quacks. They treat illnesses of the majority - mostly
bridge this vital gap. This emerging technology has the poor people - with little scientific training and
potential to lower Total Cost of Ownership (TCO) and accountability (1).
permit business plans with a higher level of viability
and confidence. With over 700 million across 600,000 villages in
India, it is no small measure to upgrade quality of life
(and healthcare) for so many. The per capita income of
Keywords
India's villages is perhaps no more than Rs 12-18,000
Telemedicine, WiMAX, IEEE 802.16, India, (USD 240-360) per annum, as compared to the national
TCO,Rural-urban divide average of Rs 25,000 (USD 500). (2)
With 75 to 80% of the doctors residing in urban
Introduction settings and not wanting to move to the village,
The dream of taking quality healthcare at an Telemedicine has come as a boon for the villages of
affordable cost to 600,000 villages of India is an India who can now access quality healthcare services
unenviable task that successive governments have remotely.
grappled with, with varying degree of success. One of the factors however, coming in the way
Telemedicine is now a proven technology that is of quick spread of this technology is the poor
capable of taking this dream closer to reality. However availability of connectivity in villages.
since it is heavily dependent on communication
technology there is need to drive technologies that As revealed by the DoT statistics itself, while the
make connectivity fast, dependable, accessible and low teledensity in the urban areas has reached 15.16, that in
cost. the rural areas continues to be low at just around
1.49. (3)
Discussion Even after 10 years of NTP 94, which allowed
private capital to come into telecom, there are 70
For every national health system of the world, thousand villages without telephones. The waiting list,
primary care is an essential foundation. The barefoot concentrated almost entirely in smaller semi-urban and
doctor in China, the General Practitioner in the United mofussil towns is still of the order of 4 million. (4)
Kingdom or the nurse, practitioners in several
advanced countries provide the basic first contact Various technologies like VSAT, wireless and
health care for people. They need to be accessible and store and forward technology have been tried to bridge
capable of giving common necessary care. For every this divide. However all these technologies suffer from
national health system, they are the gateway to health one issue or the other. Either they are too expensive,
care. Dealing with the bulk of common health care the bandwidth is not adequate or the technology is not
needs, optimize referrals to hospitals, offer economical available at the villages.
care, reduce access costs, increase comfort and hasten VSATS have adequate bandwidth and can be
recovery due to early treatment. They can also offer installed anywhere, they are however very expensive
palliative and follow up care for chronic illnesses. Any (bandwidth could be as high as Rs 70 per minute for
national health system without proper primary or first 256 KBPS).
contact care is therefore disadvantaged in several ways.
China is a good example of how primary care serves a Prof. Jhunjhunwala's interesting corDECT
national health system of a developing-nation. technology provides the last mile wireless connection,
which is less expensive but provides only 35 kbps
India however has faltered on this front, relying speeds to kiosks. Audio is compressed to 8-9 kbps and
on hospitals more than primary care. The first contact video to 20 kbps. But the trade off does not end with
or primary care in India developed later and in a very image shrinking - the voice and image do not stream

190
seamlessly. (5) This is not compatible with main stream Conclusions
telemedicine which requires at least 128KBPS up and
down stream (ideally 384 KBPS upward) with no Setting up of Telemedicine clinics in villages is
sharing of bandwidth. promising technology that can be adopted to bridge the
Rural-Urban healthcare divide.
WiMAX is a new wireless communication To be cost effective on a countrywide level the
technology that is like WiFi on steroids. It is a fast- last mile connectivity costs have to be addressed.
emerging wide-area wireless broadband technology
that shows great promise as a ‘last mile’ solution for WiMAX appears to be a promising technology
bringing high-speed broadband Internet access to that could address this issue by bringing down
village clusters. While the more familiar WiFi— connectivity costs.
802.11b and g—handles local areas within offices or
hotspots, WiMax- 802.16x covers wider metropolitan References
or rural areas. WiMax is perfectly positioned to solve
1.http://www.doctorndtv.com/feature/detailarchivefeat
the last-mile problem in developing countries like India
ure.asp?id=56 quoting Dr. Shyam Ashtekar The author
which telcos and cellular providers are facing—the
of ‘Health & Healing:
inability to quickly provide service in areas that are
A Manual of Primary Health Care’
hard for wired infrastructure to reach.
In India, where the telecom infrastructure is poor 2.http://www.emergic.org/collections/tech_talk_transfo
and last-mile connections are typically through copper rming_rural_india.html
cable, DSL and fibre optic, installation costs are high
as it requires ripping up streets to lay cables. The 3.http://www.blonnet.com/2003/05/26/stories/2003052
ability to provide these connections wirelessly, without 601710100.htm
laying wire or cable in the ground, greatly lowers the
cost of providing these services, which include Internet 4. http://www.delhiscienceforum.org/tele38.html
services, to many different parts of the country. WiMax
could thus bring broadband access into the homes and 5.http://www.hinduonnet.com/thehindu/thscrip/print.pl
businesses of millions of people in rural and ?file=2003010200010200.htm&date=2003/01/02/&prd
developing markets. =seta&
How does this happen? WiMax broadcasts its
6.http://www.expresscomputeronline.com/20040920/ne
signal over many more channels than WiFi, and those
wsanalysis02.shtml
channels are less cluttered. WiMax signals are also less
susceptible to interference. This helps a WiMax signal
travel as far as 30 miles, though on the average a Address for correspondence
WiMax base-station installation will probably cover Dept of Telemedicine, Apollo Health Street,
between three to five miles at speeds of up to 75 Mbps. 1st Floor, Life Sciences Building,
(6)
In the context of village connectivity WiMAX could Apollo Hospitals Complex,
possibly give coverage over a diameter of as much as Jubilee Hills, Hyderabad – 500033
50 to 60 KMs with bandwidths as high as 1 to 2 MBPS. India
Though we are in the early days of WiMAX Tel:
technology it is possible that this kind of last mile +91-40-23554350
connectivity could be even less expensive than ISDN Mobile:
connectivity, which is the main stay of Telemedicine +91-9885023504
connectivity today. +91-9849159419
Email:
lavanian_d@apollolife.com
lavanian@vsnl.net

191
 TELERADIOLOGY - CLINICAL EXPERIENCE AND EDUCATION

Gupta Archna, Gujral RB

Sanjay Gandhi Postgraduate Institute of Medical Sciences
Lucknow

Abstract upon the cases and discussion was held on weekly

bases at mutually convenient pre-fixed schedule.
Introduction
Teleradiology is the electronic transmission of Result
radiological images from one location to another for Different topics were selected for conventional &
the purpose of interpretation and consultation. The cross sectional imaging including interventional
focus of this information and communication radiology. The main emphasis was given to those
technology is the extension of specialist services to the imaging modalities, which were not available in the
rural domain of the general practitioners with increased radiology set up there. In few sessions - difficult cases
access to specialists. were sent from Cuttack Medical College for discussion
& opinion.
Material & Methods
Number of Sessions: Once in a week. In each
From September’ 2001 till January’ 2004, distal session 6 to 7 cases were discussed.
education on the subject of Radio diagnosis was
provided by the department of Radiodiagnosis and Participants - Radiologist, Residents and PG
imaging, SGPGIMS, Lucknow to the faculty & students, with consultants from other specialty.
residents of Radiology department of SCB Medical Only few sessions were cancelled because of
College, Cuttack, Orissa and other radiological setups communication failure due to interrupted power supply
of Orissa, which is 1500 Km away. Few sessions were / ISDN line.
conducted with Radiodiagnosis department of
PGIMER, Chandigarh located 700 Km. Teleradiology Response from radiology fraternity for Orrisa in
tool used was desktop video-conferencing system with term of teleradiology was very encouraging and was
PTZ camera interphased with film digitizer at both helpful to the students & faculty.
ends.
Conclusion
Communication link was provided with 128 Kbps Teleradiology is an ideal cost-effective tool to
ISDN telephonic lines for video-conferencing. achieve distal medical education and provide quality
Radiological images were transferred prior to main diagnostic service, training and emergency consultation
consultation and conference. Each session was to radiologist posted at peripheral radiological setups.
conducted for one and half to two hours, depending

192
 TELEMEDICINE FOR DISASTER MITIGATION IN HEALTH
A. Ayyagari, Ujjala Ghoshal
Dept. of Microbiology, SGPGIMS
Lucknow (UP), India.

Abstract Net Working

INTRODUCTION AND EPIEDIMOLOGY OF Direct (Land ) lines, metallic twisted pair,
HEALTH RELATED PROBLEMS wireless transmission links, satellites are used.
Telemedicine is a process in which expert
medical advice from afar is provided using electronic
signals to transfer the medical data from one site to Sgpgims Experience
another. It aims to provide quality health services to To asses the efficacy of telemedicine in
person whose access is otherwise restricted by developing countries like India- a pilot project was
geography or environment. initiated. A telemedicine center was setup at main
Mass disaster or local events that produce hospital of Mahakumbh Mela- at Allahabad India.
thousands of casualties, often over whelming the usual The daily reporting of the in-patient and out
capacity of health care delivery system. There is threat patient cases at the fair revealed a surge of diarrhoea
of widespread delivery of anthrax, small pox, HIV and cases among the pilgrims at the fair. This information
other agents of illness. was communicated to the referral center at SGPGIMS,
Lucknow conducted microbiological examination of
Role of Technology stool samples and rectal swabs of patients alongwith
various water & food samples.
The national aeronautics and space administration
(NASA) first used telecommunication technology to Vibra cholerae was isolated in 22.6% (7/31) of
furnish disaster aid following the devastating 1985 the samples. This information was immediately
earth quake in Mexico city. The advance technology – relayed to the health authorities who took strict and
3 (ATS-3) communication satellite provided voice prompt measures to improve hygiene. Subsequently
communication support for international rescue and the number of diarrhoea cases decreased considerably
relief efforts of the American Red Cross and Pan in a matter of few days, and thus an epidemic disaster
American Health organization The human factors was averted
research programme at the university of Maryland is
testing several application of technology during Conclusion
disaster. Voice communication is currently available Providing telemedicine facilities to local disaster
through a microwave network covering 97% of surface sites have been costly. Satellite plays a significant role
area of the state of Maryland. in providing mobility and land lines independence for
Another project known as “Mobidoc” makes use telemedicine. Soon new communication technologies
of next generation wireless technology to create an and miniaturization of computers and biosensor will
entirely mobile telecommunication syste. This enable a far great variety of users to engage in field-
communication kits, which is the size of briefcase, level use of telemedicine during disasters. In
contains eight cell phones and wireless data acquisition complicated disaster scenarios, internet- based
devices that are connected to cell phones. consultation system may be ideal.

193
 TELECONSULTATION / FOLLOW UP SCREENING IN
SPECIALITY MEDICINE

Dr Bidyut Kumar Das, MD, PhD

Associate Professor
Department of Medicine
Division of Clinical Immunology/Rheumatology
SCB Medical College, Cuttack. Orissa.

Abstract management outlined. If patients required further

evaluation for tests not available at the local centre,
Background they were instructed to visit Lucknow on a specific
date.
Rheumatology is not a well understood field in
Tele follow up was provided to patients on a
India although the prevalence of rheumatological
weekly basis by RM for those patients on long term
disorders is over 10 %. There are very few centres
treatment to assess suitability of therapy, drug toxicity
providing specialised care and few trained consultants
and to avoid the cost of travelling to Lucknow.
in this field. The care providers are mostly located in
metropolitan cities resulting in unsatisfactory screening
and treatment of patients with rheumatological Results
disorders in most states of the country. Telemedicine is
an important avenue in correcting this problem through A total of 34 educational and consultation
interaction with rheumatologists based in a sessions were held along will 12 tele-follow ups
superspeciality centre. SGPGIMS, Lucknow, is linked during this period. 180 post graduate students and
to a peripherally situated tertiary care centre, SCB interns attended these interactive courses. A didactic
Medical College in Cuttack, Orissa, which caters to a lecture supported by illustrations and radio-imaging
large population base within and from neighbouring preceded case presentation, analysis and management
states, has demonstrated the enormous benefits of this outline.
link up in the teaching of rheumatology and
management of complicated cases. 57 patients with complicated connective tissue
disorder were discussed. All patients were given a final
Objective diagnosis , management outlined and asked to follow
up periodically. The patients are doing well and did not
Role of telemedicine in screening and follow up have to go to higher centers for diagnosis.
of cases in speciality medicine like rheumatology The details of the results will be discussed.
through interaction between specialists in a centrally
located super-speciality centre and a peripherally 20 patients benefited from the 12 follow up
located tertiary care medical college. sessions. Logistically and financially it was beneficial
for the chronic patients who had difficulty in travelling
to distant places.
Material and Method
Patients attending the Rheumatology clinic at
SCB Medical College, Cuttack, Orissa between Conclusion
September 2003 and January 2005, were the subjects
who benefited from tele-consultation and subsequent Telemedicine is a vital link in providing
follow up. Cases were assigned dates for presentation speciality service in areas like rheumatology which has
and discussed with the consultants( RM, AA, VA ) of a dearth of consultants in the country. Besides
SGPGI MS, Lucknow. There were weekly sessions of providing training to students and physicians, it helps
one hour duration, three days in a week. Cases were in the diagnosis and management of a large number of
discussed in detail, investigations analysed and complicated cases.

194
 TELE-CME FOR POSTGRADUATE MEDICAL STUDENTS

Biswanarayan Mohanty, S. S. Mishra, D. Patel, S. K. Mishra

Nodal Officers Telemedicine, SCB Medical College, Cuttack; MKCG Medical College, Berhampur; VSS Medical
College, Burla; SGPGIMS, Lucknow.

Abstract telemedicine facility made available through Orissa

Telemedicine Network.Medical experts at SGPGIMS,
Postgraduate Medical Students need to learn Lucknow offered teaching, illustrative & operative
practical skills and in-depth understanding of subject demonstrations having significant educational value to
matter. The training period shapes them as future students at the user’s end .
specialists. Hence such training necessitates
availability of adequate study material and expert Analysis of data at the end of three years revealed
teachers – the requirements that are lacking in most remarkable benefit of students in terms of Subject
peripheral medical colleges of India. understanding, Gain of new knowledge, Skill updation
and Radiological image/ Pathological slide reporting.
An attempt was made in Orissa to achieve the
goal through carefully designed need based Tele-CME We feel that Tele-CME should be an essential
programs for postgraduate students of all three Govt. component of postgraduate medical teaching
Medical Colleges utilizing realtime interactive armamentarium wherever Telemedicine facility exists.

195
BIOMEDICAL APPLICATIONS OF ELECTROPORATION ON HUMAN RED
BLOOD CELLS

S. H. SANGHVI1 , K. P. MISHRA2
1
Mahatma Gandhi Mission’s College of Engineering and Technology, Navi Mumbai
2
Radiation Biology and Health Sciences Division, Bhabha Atomic Research Centre,
Trombay, Mumbai - 400 085

Abstract subjected to exponential and square wave electric field

Exposure of biological cells to high intensity under similar conditions resulted higher hemolysis in
and short duration electrical pulses ( a few Kv /cm) square wave then exponential. This result can be
beyond a threshold value induces dramatic transient attributed to longer effective pulse length in square
increase in the permeability of plasma membrane wave, which are in the order of milli sec compared to
which allows entry of otherwise impermeant exponential pulse in the range of micro second. The
exogenous molecules into the cell. This phenomenon elctroporative drug and gene delivery to target cell has
called electroporation, has opened many new provided novel prospect for biomedical application
possibilities of cellular manipulation which has so namely gene therapy, hybridoma technology, and drug
promise to develop a variety of new application in delivery. More recent studies on tumor cells have
biology in medicine. Electroporation allows to gain shown enhance cytotoxic effects of ionizing radiations
direct entry into the cell interior without seriously and anticancer drugs in combination with
compromising cellular viability. Studies have been electroporation pointing to new approach to improve
carried out on electroporation on red blood cells cancer therapy and is popularly known as
(RBC). It has been shown that human erythrocyte electrochemotherapy.

196
 ESTABLISHING HEALTHCARE BRIDGES IN THE INTERNATIONAL
ARENA: AN INFORMATICS COMPANY PERSPECTIVE
Dr. Rini Verma
CAL2CAL Corporation

Abstract

Questions in the healthcare information such as Public Health Nursing. We have developed
technology domain are many. We are increasingly a Epidemiological Data gathering tools for
global community and nations around the world are Communicable disease tracking and management, like
interested in determining how information technology SARS and other emerging diseases. In the clinical
can improve the access, quality, safety and efficiency domain we are building comprehensive World
in their healthcare systems. India, like many other
Informatics Information Systems for specific Cancer or
countries, continues to struggle with delivery of quality
Disease management. In the assistive and social issues
healthcare. Oxfam reports there are 51 physicians per
domain we are providing alternative solutions that
100,000 people in India. Tools to enable better
integrate non-traditional IT services and products with
management and tracking of healthcare are essential.
CAL2CAL Corporation’s healthcare initiatives explore available administrative services and infrastructure, to
the development of tools that can be applied equally provide aid to poor populations living in remote areas
within industrialized and developing nations to of developing countries. Public and Private healthcare
facilitate the creation of an electronic health issues were never boundary determined. Health
information infrastructure. In the public health domain standards and global interoperability have to be
is our Case and Patient Information Management tool adopted across such enforced boundaries.

197
 Telemedicine in India - ISRO's Initiatives
∗
L.S.Satyamurthy & R.L.N.Murthy

connected to 20 super speciality hospitals across the

The pride of the citizens of a nation is enhanced when country including a Telemedicine network for armed
the investments in Science and Technology reap social forces. The hall mark of ISRO’s Telemedicine system
benefits and more appropriately when the results are configuration and its adaptation has been the
visible by the common man at the grass roots. Indian progressive evolution from point to point system to
Space Agency’s endeavour in this direction is one such multipoint to multipoint system with Browser based
example of integration of technology for efficient client/server architecture with specific emphasis on
health care delivery to the masses. cost reduction.

The three tier Indian health care system mostly The present Telehealth system offering Telecardiology,
governed by respective State Governments is lagging Teleradiology, Telepathology and Teleophthalmology
behind the international standards in terms of with an appropriate video conferencing facility is
beds/population, health expenditure as a share of GDP, realized through a VSAT based Satcom connectivity
infant mortality and medical consultants etc. Further provided through a dedicated transponder of the Indian
this is compounded by the fact that majority of Indian National Satellite (INSAT) System. This initiative
population being rural, living in distant and received good acceptability as observed from the
inaccessible areas while the majority of medical utilization survey and the increasing number of
practitioners are predominantly concentrated in the requests for extending the service. More than 17000
urban areas. patients have been provided teleconsultation and
treatment during two year period.
Noting this important dichotomy, especially the urban-
rural health divide, the Satcom based Telemedicine This network is also being used as a communication
programme pioneered by the Indian Space Agency, backbone for CME (Continuing Medical Education)
ISRO (Indian Space Research Organisation) has been connecting some of the medical institutions and
tailored to the specific needs of the society vis-à-vis the hospitals across the country.
development of technology for the health care delivery
system in the country. This Telehealth project is aimed ISRO in association with the Departments of Health
at speciality health care delivery to the distant and and Information Technology has evolved Guidelines
needy sections of the society, which was earlier limited and Standards for the practice of Telemedicine in India.
to a privileged few. While the legal and ethical aspects of Telemedicine
practice is being addressed by the health
ISRO’s Telemedicine programme was initiated in the administration, it is planned to set the focus on
year 2001 integrating the experience gained in the development of operational and Business models to
previous two decades of Satcom based projects in the make the system viable and sustainable.
utilization of communication and Information
technologies with innovative approaches in the The first step towards this endeavour is the
development and improvement of solutions needed for involvement of the state governments , Non
education and health care. Initiated as pilot project at Governmental Trust Agencies and speciality hospitals
selected locations, and crossing the initial phase of providing the initial investment on infrastructure
proof of concept, technology demonstration and including the Telemedicine and Communication
followed by few operational endeavors, ISRO’s systems while ISRO providing the satellite bandwidth
Telemedicine programme is set to enter operational subsidized by the federal government for the cause of
arena for sustainable business model. ANTRIX societal development. There has been an encouraging
Corporation, the commercial and marketing arm of response to this proposition by the state governments
ISRO/Department of Space (DOS) is poised to and speciality hospitals.
facilitate a Public-Private Partnership endeavour
towards this business initiative. In the Southern Indian State of Karnataka, which
adopted the concept of Telemedicine in the early years,
During the past two and half years, ISRO’s initiatives the state government in collaboration with an insurance
alone have seen the networking of more than 75 company has come out with an innovative self-funding
hospitals - 55 remote/rural patient end hospitals health insurance scheme called YESHASVINI

∗
Indian Space Research Organisation (ISRO), Bangalore

198
(meaning the vehicle of success) for the benefit of
farmers in rural areas. Under this scheme a nominal
premium of 10 US cents per month per head will
enable the farmers and their families quality
healthcare that includes cost of critical surgeries. The
members get free outpatinet consultation in recognized
hospitals. Presently there are about 1.7 Million farmers
in this scheme, which has been growing significantly.
With its dynamic structure that generates funds to run
itself, this scheme forms a working model for more
such initiatives across the country. Thus it promises to
be a turning point in the business model of
Telehealthcare with more and more members joining
the scheme and thereby increasing the prospects and
services of health care centers.

While this kind of insurance schemes will increase the

volume of consultations/business, further steps in the
direction of converting the Indian Telemedicine
services as attractive business ventures are being
pursued along with efforts towards cost reduction of
the systems - Hardware, Software and Diagnostic
instruments and defining the optimum configuration
required at the three tiers of Primary, Secondary and
Tertiary Health Care system. ISRO is studying
different system configurations with the aim of
reducing the cost as per the medical needs of the
beneficiaries making it more affordable to the
stakeholders.

Health care in India being one of the important

obligations & responsibilities under the charter of the
state governments, the initial investments in
infrastructure are generally provided by the state
governments. The aim of the business model is to
break even adequately with regards to the operation
and maintenance costs.

The addition of multiple facilities at the same

Telehealth centre is also expected to add to the revenue
of the centre and the bottom line is “ Increased usage
by reduced costs and the reduced costs by increased
usage”.

ISRO envisages the development of an exclusive

HEALTHSAT for meeting the enhanced health care
needs of the country in view of the changing morbidity
patterns of diseases and the genetic susceptibility of the
population even at the rural areas.

199
 Sponsors

CAL2CAL Corporation

HCL Infosystems Limited

IBM India Limited

Indian Space Research Organization

Ministry of Communication & Information Technology,

Govt. of India

WEBEL

200

View publication stats