Advanced Terminology Systems

Maryelle Grace Joy M. Samson Angeliness Ildefonso Catherine Mangulabnan Rowel Ian Guillermo Paulo Gabrielle Malijan

Primary motivation: the need for valid, comparable data that can be used across information system applications to support clinical decision-making and the evaluation of processes and outcomes of care.

Vocabulary problem
failure to achieve a single, integrated terminology with broad coverage of the healthcare domain

Reasons for Vocabulary Problem:

1) Multiple specialized terminologies has resulted to overlapping content, areas of which no content exists, and large numbers of codes and terms.

2) Existing terminologies are primarily intended for human interpretation, with computer interpretation as only a secondary role.

Concept Orientation
In order to appreciate the significance of concept-oriented approaches, it is important to first understand the definitions of and relationships among objects, concepts and the terms we use.

Semiotic Triangle

ISO 1087-1:2000
Concept unit of knowledge created by a unique combination of characteristics. *characteristic is an abstraction of a property of an object of a set of objects.

 Object anything perceivable or conceivable. Term verbal designation of a general concept in a specific subject field. *general concept corresponds to two or more objects which form a group by reason of common properties.

Evaluation Criteria related to Concept-oriented Approaches

 Atomic-based concepts must be separable into constituent components.

Compositionality ability to combine simple concepts into composed concepts

 Concept permanence once a concept is defined it should not be deleted from a terminology.

Language independence support for multiple linguistic expressions.

 Multiple hierarchy accessibility of concepts through all reasonable hierarchal paths with consistency of views. Nonambiguity explicit definition for each term.

 Nonredundancy one preferred way of representing a concept or idea Synonymy support for synonyms and consistent mapping of synonyms within and among terminologies

A single concept may be associated with multiple terms, but a term should represent only one concept.

Components of Advanced Terminology Systems

Terminology Model Representation Language Computer-Based Tools

Terminology Model
a concept-based representation of a collection of domain-specific terms that is optimized for the management of terminological definitions. It encompasses both schemata and type definitions.

Schemata
incorporate domain-specific knowledge about the typical constellations of entities, attributes, and events in the real world and, as such, reflect plausible combinations of concepts.
Ex. dyspnea + severe = severe dyspnea

Type Definitions
obligatory conditions that state only the essential properties of a concept. Ex. A nursing activity must have a recipient, an action, and a target.

Representation Language
GALEN Representation and Integration Language (GRAIL) Knowledge Representation Specification Syntax (KRSS) Web Ontology Language (OWL)

Ontology Language
Represents classes and their properties Able to support the formal definition of concepts in terms of their relationships with other concepts, and facilitate reasoning about those concepts

Computer-Based Tools
A representation language may be implemented using description logic within a software system or by a suite of software tools.

Classifications of Terminology Systems

First-generation terminology systems Second-generation terminology systems Third-generation systems

First-generation terminology systems

consist of a list of enumerated terms, possibly arranged as a single hierarchy. Serve a single purpose or a group of closely related purposes and allow minimal computer processing NANDA, Nursing Interventions Classification (NIC)

Nursing Interventions Classification (NIC)

a comprehensive, standardized system to classify treatments performed by nurses.

North American Nursing Diagnosis Association (NANDA)

professional organization of nurses standardized nursing termin ology that develops, researches, disseminates and refines the nomenclature, criteria, and taxonomy of nursing diagnoses.

Second-generation terminology systems

include an abstract terminology model or terminology model schema that describes the organization of the main categories used in a particular terminology or set of terminologies.

 Can be used for a range of purposes, but they allow only limited computer processing; automatic classification of composed concepts is not possible Beta 2 version of the International Classification for Nursing Practice (ICNP)

Abstract terminology model

complemented by a thesaurus of elementary descriptors (terms) and templates or rules (grammar)

Third-generation systems
support sufficient formalisms to enable computer-based processing Include grammar that defines the rules for automated generation and classification of new concepts.

Advantages of Advanced Terminology Systems

Allow much greater granularity through controlled composition, while avoiding a combinatorial explosion of precoordinated terms. Facilitate two important facets of knowledge representation for computer-based systems that support clinical care

Two important facets:

Describing concepts Manipulating and reasoning about those concepts using computerbased tools

Describing concepts
Nonambiguous representation of concepts.
Facilitation of data abstraction or deabstraction without loss of original data.

 Nonambiguous mapping of terminologies Data reuse in different contexts

Manipulating and reasoning about those concepts using computer-based tools

Automated classification of new concepts Ability to support multiple inheritance of defining characteristics

Advanced Terminological Approaches in Nursing

ISO 18104:2003 GALEN SNOMED RT

ISO 18104:2003
Developed by ISO Technical Committee 215 (Health Informatics) Working Group 3 (Health Concept Representation) under the collaborative leadership of the International Medical Informatics Association Nursing Special Interest Group (IMIA-NI) and the International Council of Nurses

 Approved in 2003 Covers reference terminology models for nursing diagnoses and nursing actions The model built on work origination within the European Committee for Standardization

 Development was partly motivated by a desire to harmonize the plethora of nursing terminologies in use around the world Intended to be consistent with the goals and objectives of other specific health terminology models in order to provide a more unified reference health model

Potential uses:
Facilitate the representation of nursing diagnosis and nursing action concepts and their relationships in a manner suitable for computer processing Provide a framework for the generation of compositional expressions from atomic concepts within a reference terminology

 Facilitate the mapping among nursing diagnosis and nursing action concepts from various terminologies Enable the systematic evaluation of terminologies and associated terminology models for purpose of harmonization

 Provide a language to describe the structure of nursing diagnosis and nursing action concepts to enable appropriate integration with information models

GALEN
A concept-oriented approach developed within the GALEN Program Used in a range of ways, from directly supporting clinical applications to supporting the authoring, maintenance, and quality assurance of other kinds of terminologies

 GRAIL is an ontology language for representing concepts and their interrelationships the source material for construction of terminology models.

Two sets of tools used in development of GRAIL Model:

Computer-based modeling environment Terminology server

Computer-based modeling environment

Facilitates the collaborative formulation of models Allows authoring of clinical knowledge at different levels of abstraction

Terminology server
a software system that implements GRAIL A major motivation for applying GALEN to nursing was the desire to meet the requirements of users of clinical applications, and the need to provide a reusable and extensible model of nursing terminology

GALEN advocates five fundamental paradigm shifts:

user interface structure establishing standards presentation delivery

In the user interface

to shift from selecting codes to describing conditions allow a central concept to be described through simple forms.

In the structure
to shift from enumerated codes to composite descriptions terminologies are internally analogously to a dictionary and a grammar

 traditional coding systems are more like a phrase book; each sentence must be listed separately.

In establishing standards
to shift from a standard coding system to a standard reference model.

GALEN Common Reference Model

provides a common means of representing coding and classification systems so that they can be interrelated - a common dictionary and grammar.

In delivery
to shift from static coding systems as data to dynamic terminology services as software. GALEN originated the idea of a terminology server and is participating actively in the CorbaMed effort at standardizing the software interface.

In presentation
to shift from translations of monolingual terminologies to multilingual terminologies.

Functions of GALEN:
Internally managing and representing the mode Testing the validity of combinations of concepts Constructing valid composed concepts

 Transforming composed concepts into canonical form Automatically classifying composed concepts into the hierarchy Deliver the model for use by clinical applications and other kinds of authoring environments

SNOMED Reference Technology (SNOMED RT)

An alternative concept-oriented approach; developed through the collaboration between the College of American Pathologists and Kaiser Permanente, based on SNOMED International.

 Is a reference terminology optimized for clinical data retrieval and analysis Concepts and relationships are represented using modified KRSS rather than GRAIL

Functions of SNOMED RT:

Acronym resolution, word completion, term completion, spelling correction, display of the authoritative form of the term entered by the user, and decomposition of unrecognized input Automated classification Conflict management, detection, and resolution

SNOMED Clinical Terms (SNOMED CT)

Developed by College of American Pathologists and UK National Health Service Possesses both reference terminology properties and user interface terms.

Emerging Approaches

Web Ontology Language (OWL)

intended for use where applications, not humans, are to process information.

OWL builds on existing recommendations such as:

eXtensible Markup Language (XML) surface syntax for structured documents

Resource Description Framework (RDF) a data model for resources

 RDF Schema a vocabulary for describing the properties and classes of resources

IMPLICATIONS FOR NURSING

Provide for nonambiguous concept definitions Facilitate composition of complex concepts from more primitive concepts Support mapping among terminologies.

Benefits of clinical approach:

Facilitation of evidence-based practice Matching of potential research subjects to research protocols for which they are potentially eligible. Detection of and prevention of potential adverse drug effects

 Linking online information resources Increased reliability and validity of data for quality evaluation Data mining for purposes such as clinical research, health services research, or knowledge discovery.

END