ICEH Research Webinar Clare Sept 21 2023 v293 Read Only

Principles of good research for
eye health
Clare Gilbert
iceh.lshtm.ac.uk/
Inverse-research law
• ~90% of people with vision impairment live in low- and middle- income countries
• Only 30% of published research was undertaken in these regions
1%
5%
11%
Central/East Europe, Central Asia

High-income
4% Location of primary
Latin America and Caribbean 6% research 2000-2019
North Africa and Middle East
South Asia 2%
Southeast Asia, East Asia, Oceania
Sub-Saharan Africa
70%
Ramke et al. Lancet Commission on Global Eye Health, Supplement

Hierarchy of scientific evidence
Synthesis of data from all randomized controlled trials/studies a

addressing the same topic
Gold standard for assessing a new drug / surgical technique
Incidence and risk factors for a condition
Risk factors for a condition
Surveys: prevalence (and risk factors) for a condition
By The Logic Of Science - https://thelogicofscience.files.wordpress.com/2016/04/hierarchy-of-evidence-no-not1.png, CC BY-SA 4.0,

https://commons.wikimedia.org/w/index.php?curid=93799763
Principles of good research (for eye health)
Purpose
• To fill gaps in our knowledge by providing evidence that we can trust, use in decision
making and build on
• Without evidence decisions tend to be based on
• Experience – our own or other people’s
• Eminence – what the boss says!
• Eloquence – persuasive arguments
• Expectations – of patients, their families or health care workers
…all of the above are subjective, and are likely to be biased
• Local research, which addresses local needs and knowledge gaps, is essential
Where to start….
o Identify the problem/knowledge gap you want to address
o Decide who are you going to work with; more brains are always better than one!
o What have other researchers done and found? Find and read other studies
o Time spent on detailed, logical planning is very important: once data have been
collected it is usually too late to make changes
o Everything always takes longer than anticipated…….
o Observational / descriptive studies are more straightforward than studies which assess
an intervention (a new way of doing something, which may be context specific - a new
screening or diagnostic test; a new surgical technique or medication)
Planning a study
Detailed, logical planning is critical to success….
• Have an open mind about what you might find
• Use your imagination
• Think about assumptions you may have
• Beware of bias, which can lead to results which are difficult to
interpret
• A study with too few participants may not give valid results; too many
participants is a waste of their/your time and resources
Where to start…
Write down a good research question which addresses the gap identified
• Clear, specific, concise
Problem: Patients with open angle glaucoma are coming back blind because they have not
attended follow up
Question: What proportion of patients with open angle glaucoma drop out of regular follow up
after diagnosis, and which patients are less likely to attend regularly?
Problem: Children with congenital cataract are presenting too old for management to give good
outcomes
Question: Why do children with bilateral congenital cataract present after the age of two years to
the eye department?
Problem: The waiting list for adult cataract surgery has grown to one year
Question: What are the provider and patient factors which lead to the long waiting list?
Problem: During outreach most of the cataract blind individuals are elderly women
Question: What factors prevent elderly cataract-blind women from attending for
cataract surgery, and what are they willing to pay for surgery?
Problem: Many glaucoma patients say they have difficulty putting in their eye drops.
Question: What is the best way to teach patients how to put in their eye drops?
Planning an observational study
Aim and obje ctive s

- Aim: how the da ta collected will Me thods
Proble m / e vide nce Re s e arch que s tion Decide methods for ea ch objective
a ddres s the problem identified
gap to be Clea r, s pecific a nd - qua ntita tive methods : numbers
- Objectives : s pecific a ctivities
addre s s e d concis e - qua lita tive methods : na rra tive text
which need to be underta ken to
a ns wer the res ea rch ques tion - mix of the a bove
De cide data to be colle cte d

Study participants De ve lop and pilot te s t data
- extra ct da ta from
- pa tients / fa mily colle ction me thods Sample s ize
medica l/ s urgica l records
members - ques tionnna ires for Decide how ma ny
- ques tionnna ires for qua ntita tive
- community members qua ntita tive da ta pa rticipa nts a re needed
da ta
- eye hea lth profes s iona ls - interviews guides with to give relia ble res ults
- interviews guides with ques tions
- more tha n one group ques tions to a s k
to a s k for qua lita tive da ta
Data analys is Plan the logis tics Plan dis s e mination of Ethical approval
Decide how the da ta a re Timefra me; s ta ff tra ining; re s ults Budge t Not a ll s tudies need
going to be a na lys ed where da ta will be Pres enta tions , Bes t to do this once the ethica l a pprova l; s eek
Different methods a re collected a nd how (pa per; publica tions , reports ; methods a nd logis tics a dvice from a n ethics
needed for qua ntita tive electronic; a udio feedba ck to pa rticipa nts ha ve been worked out committee if you a re not
a nd qua lita tive methods recordings ) etc etc s ure
What next?
• Identify people with the relevant expertise (e.g., qualitative research,
statistics) and get their input
• Make revisions
• Seek an independent opinion from an experienced researcher
• Listen to what they say
• Revise the research plan (“protocol”)
Data analys is Plan the logis tics Plan dis s e mination of Ethical approval
Decide how the da ta a re Timefra me; s ta ff tra ining; re s ults Budge t Not a ll s tudies need
going to be a na lys ed where da ta will be Pres enta tions , Bes t to do this once the ethica l a pprova l; s eek
Different methods a re collected a nd how (pa per; publica tions , reports ; methods a nd logis tics a dvice from a n ethics
needed for qua ntita tive electronic; a udio feedba ck to pa rticipa nts ha ve been worked out committee if you a re not
a nd qua lita tive methods recordings ) etc etc s ure
Implementing an observational study
Unde rtaking the s tudy Che cking and analys ing the Inte rpre t the findings
Da ta collection m us t be data To wha t extent ha s the
rigorous a nd of high qua lity. Check com pletenes s of da ta , res ea rch ques tion been
Needs good da ta collection a nd follow the pla n for a na lys is a ns wered? Wha t a re the Dis s e minate the findings
intrum ents ; well tra ined s ta ff; (qua ntia tive a nd qua lita tive). im plica tions of the findings ? according to the plans
s upervis ion; m onitoring of Ma y need to report progres s to Wha t a re the lim ita tions of the
com pletenes s a nd a ccura cy of the ethics com m ittee. s tudy (a ll s tudies ha ve
da ta collection etc lim ita tions !)?
No research is perfect…..
• But there are poorer or better ways of conducting studies
• Some studies can be “fatally flawed” which means they cannot be published e.g.,
major omission in the data collected; lack of a comparison group
• It is much better to seek advice early rather than waiting until all the data have
been collected before asking someone to help with the analysis or writing up

ICEH Research Webinar Clare Sept 21 2023 v293 Read Only

Uploaded by

Copyright:

Available Formats

ICEH Research Webinar Clare Sept 21 2023 v293 Read Only

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

ICEH Research Webinar Clare Sept 21 2023 v293 Read Only

Uploaded by

Copyright:

Available Formats

Principles of good research for

Central/East Europe, Central Asia

Ramke et al. Lancet Commission on Global Eye Health, Supplement

Synthesis of data from all randomized controlled trials/studies a

Gold standard for assessing a new drug / surgical technique

Incidence and risk factors for a condition

Risk factors for a condition

Surveys: prevalence (and risk factors) for a condition

By The Logic Of Science - https://thelogicofscience.files.wordpress.com/2016/04/hierarchy-of-evidence-no-not1.png, CC BY-SA 4.0,

Aim and obje ctive s

De cide data to be colle cte d

You might also like