2. kcochrane@kemri.go.ke / www.kenya.cochrane.org @CochraneKenya
Evidence is anything used to determine or demonstrate the truth of an assertion.
Scientific evidence: Evidence which serves to either support or counter a Scientific Theory or
hypothesis – Accumulated through observations of phenomena which occurs in the natural
world or laboratory experiments
5. kcochrane@kemri.go.ke / www.kenya.cochrane.org @CochraneKenya
Generation
of Primary
research
Synthesis of
relevant
evidence
Evidence
reaches
policy
makers
Policy
makers value
evidence
and use it in
decision
making
Policy
makers
implement
largescale
change
Flow of Evidence from Generation to Implementation
6. kcochrane@kemri.go.ke / www.kenya.cochrane.org @CochraneKenya
The evidence Ecosystem
Evidence generation
Evidence Synthesis Evidence Translation
Regulatory
Culture
Economy
Social
Politics
8. kcochrane@kemri.go.ke / www.kenya.cochrane.org @CochraneKenya
Evidence Synthesis
The use of techniques to combine multiple sources of data- Clinical or observational, to comprehensively
understand their findings.
What is it & Why do we need it
9. kcochrane@kemri.go.ke / www.kenya.cochrane.org @CochraneKenya
• Making sense of research
• Same question – different answers from studies
• Coping with information overload
• Millions of new articles, grey literature, internet – cope by
relying on literature reviews to stay up to date
• Justification of future research
• What gaps in knowledge the proposed research intends to fill
• Facilitating access to relevant research
• Avoiding publication biases
• Better accountability
• Many effective treatments (alternatives)
• Costs (avoiding waste)
Why is Evidence Synthesis Important
10. kcochrane@kemri.go.ke / www.kenya.cochrane.org @CochraneKenya
What is a review
• A review is a summary or synthesis of the results or conclusions of more than
one study
• Several types of reviews available, eg:
• Narrative reviews
• Scoping reviews
• Rapid reviews
• Systematic reviews
• Reviews differ by rationale, methodological rigor and risk of bias
• A good review should be a readable summary of ALL the evidence, unbiased,
transparent, replicable and up-to-date
11. kcochrane@kemri.go.ke / www.kenya.cochrane.org @CochraneKenya
Narrative Review
•Summaries of evidence on a given topic usually
written by an expert in the field.
•Typically, involve informal and subjective methods to
collect and interpret information
•Problems with narrative reviews
• Personal bias
• Publication bias
• Language bias
• Small sample size
12. kcochrane@kemri.go.ke / www.kenya.cochrane.org @CochraneKenya
Scoping Reviews
A type of knowledge synthesis that uses a systematic and
iterative approach to identify and synthesize an existing or
emerging body of literature on a given topic
Aim is to map the body of literature on a topic area.
13. kcochrane@kemri.go.ke / www.kenya.cochrane.org @CochraneKenya
Rapid Review
A form of knowledge synthesis that accelerates the process of conducting a
systematic review through streamlining or omitting specific methods to
produce evidence for stakeholders in a timely and resource-efficient manner
shorten the process by:
Limited scope of review question
Limiting search by years, databases, language and sources beyond electronic searches
Single rather than two reviewers for some review steps (eg title and abstract review, full
text review, methodological quality assessment and data extraction process)
Limited risk of bias assessment
Qualitative summary rather than meta-analysis
Fast-tracking of PROSPERO registration
14. kcochrane@kemri.go.ke / www.kenya.cochrane.org @CochraneKenya
Systematic Review
•“A review in which bias has been reduced by :
• the systematic identification
• Critical appraisal
• Synthesis, and,
• If relevant, statistical aggregation
• of all relevant studies on a specific topic according to a
predetermined and explicit method.”
15. kcochrane@kemri.go.ke / www.kenya.cochrane.org @CochraneKenya
Main elements of a systematic review
• Pre-specified protocol
• Clearly focussed question
• Pre-determined selection criteria
• Detailed description of search strategy
• Explanation of types of evidence included and excluded
• Critical examination of quality of included studies
• Transparent process for synthesising and interpreting findings
• Kept up-to-date
16. kcochrane@kemri.go.ke / www.kenya.cochrane.org @CochraneKenya
Preparing for a Systematic Review
• Any review project requires three components:
• Review question: defines your destination
• Review protocol: details your proposed route and activities
• Review team
• Will continually consult the review question and protocol in the
systematic review process
• Review question—clear, well-defined, appropriate, manageable
and relevant to outcomes
• Beware that both review question and protocol may take longer
than you think.
• A review team helps to minimize bias
17. kcochrane@kemri.go.ke / www.kenya.cochrane.org @CochraneKenya
Preparing for a Systematic Review
• Aim of review protocol:
• To describe current evidence
• To identify the review question
• To outline/pre-specify methods to be used to answer review question
• What studies will be included and excluded?
• How (exactly) will each stage be conducted?
• What are the primary and secondary outcomes of interest?
• Needs to be registered (to avoid duplication of efforts) - PROSPERO
• Journals often require a protocol registration number before publishing
systematic reviews
18. kcochrane@kemri.go.ke / www.kenya.cochrane.org @CochraneKenya
PROSPERO Registration
• Search for ongoing reviews
• Register planned review online (& update as work progresses)
• Avoids duplication of reviews
19. kcochrane@kemri.go.ke / www.kenya.cochrane.org @CochraneKenya
Systematic Review Process
• Review question
• Has 4 components (PICO)
• Search for studies
• Need a search strategy
• Search more than one database
• No restriction with language, publication status
• Consider contacting experts, reference lists
• Selecting studies
• At least two authors screen studies independently
• Use the inclusion criteria developed from the review question (PICO)
• In two stages (titles and abstracts, full text papers)
• Can be done electronically (eg Covidence)
• Risk of bias assessment
• Use a validated tool, eg Cochrane risk of bias tool
20. kcochrane@kemri.go.ke / www.kenya.cochrane.org
Systematic review process
• kcochrane@kemri.go.ke / www.kenya.cochrane.org
@CochraneKenya
• Data extraction
• Need a (piloted) data abstraction sheet
• Sone by at least 2 authors
• Be systematic
• Resolve disagreement
• Data analysis/synthesis
• Develop analysis plan at protocol stage
• If data is not combinable, describe qualitatively
• If data is combinable, consider meta-analysis
• Use software (eg Revman, Stata, Met-analyst)
• Establish presence of heterogeneity
• Summary of the findings (use GRADE methodology)
• Report writing
• Structured report
• Include a plain language summary