What Counts as Healthcare Evidence?

#Start with the question

Evidence is useful when it answers the question being asked. A comparison of treatments, a test's accuracy and people's experiences of care need different kinds of information. NICE describes evidence review as a systematic process that can examine experimental, observational and qualitative research.

Specify the people, setting, comparison and outcomes before deciding what to read. A study answering a different question may be informative but indirect. Neither a large dataset nor a familiar study label makes its result automatically relevant to the decision in front of you.

Evidence: NICE guideline manual: reviewing evidence / Cochrane Handbook, Chapter 14: certainty of evidence

#Choose evidence for the outcome

Randomised trials can estimate the effects of assigning people to different interventions. They do not necessarily report every important benefit or harm. Cochrane notes that rare or long-term adverse outcomes may require evidence from non-randomised studies, with attention to whether those studies concern comparable people and settings.

Qualitative research can address experiences, acceptability and barriers that an effect estimate does not describe. NICE assesses confidence in qualitative findings separately from certainty in intervention effects. One kind of finding should not be presented as if it had measured the other.

Evidence: Cochrane Handbook, Chapter 14: certainty of evidence / NICE guideline manual: reviewing evidence

#Study quality and certainty are different questions

Assess how each study was done, then ask how much confidence the body of evidence supports for the particular outcome. In the GRADE approach described by Cochrane, concerns about bias, inconsistent findings, indirectness, imprecision and missing publication can reduce certainty. The judgement is outcome-specific, not a single quality badge for everything in a paper.

A systematic review is not automatically strong evidence merely because it is called a review. Its methods, included studies and missing information matter. Look for explanations of the judgement, rather than relying only on a label such as high or low certainty.

Evidence: Cochrane Handbook, Chapter 14: certainty of evidence / NICE guideline manual: reviewing evidence

#Evidence informs a choice; it does not make it

A research result estimates what may happen under particular conditions. A recommendation also weighs the importance of benefits and harms, available resources, feasibility and people's priorities. NICE explicitly distinguishes interpretation and recommendation-making from the evidence summary.

Read the effect size and uncertainty alongside the outcome that was measured and the people studied. Separate an uncertain estimate from an unsupported claim, and separate a guideline's recommendation from a study's result. This page explains evidence reading; it does not assign a treatment or replace individual clinical advice.

Evidence: NICE guideline manual: interpreting evidence and recommendations / Cochrane Handbook, Chapter 14: certainty of evidence

Source note

The sections above were checked against the linked sources. No clinical review has been performed. This is general research education, not a clinical guideline.