Clinical Evidence Appraisal

Considering findings alongside their methods and limitations.

Clinical evidence appraisal is the work of establishing, through its method, how much weight the result of a published study can carry. A reading that starts at the results section usually makes a finding look more certain than it is.

Reading the method

The reliability of a study is examined under several headings:

• Randomisation and allocation concealment — whether groups are comparable in what was not measured as well as what was.
• Blinding — whether an effect is systematically inflated, particularly for assessor-dependent outcomes such as pain or quality of life.
• A pre-specified primary endpoint — the rising chance of a chance-significant result where many endpoints are measured.
• Handling of missing data — how participants who did not complete the study entered the analysis.

Effect size and uncertainty

To judge what a result means in practice, relative effect, absolute difference and confidence interval are read together. A p-value alone describes neither the size nor the clinical importance of an effect; it is a partial indicator of how compatible an observed difference is with chance.

Certainty of evidence is a separate question

Statistical significance and confidence in the evidence are not the same thing. The GRADE approach grades certainty across study design, risk of bias, consistency, directness, precision and publication bias. The distinction separates "was found significant" from "has been reliably demonstrated".

External validity

That a result is true does not mean it applies to every patient. Eligibility criteria, the comorbidities excluded and the healthcare setting in which a study ran determine to whom a finding can reasonably be generalised.

What reporting guidelines are for

CONSORT defines what should be reported for randomised trials and STROBE for observational studies. Missing elements do not show that a result is wrong; they show where an appraisal could not be carried out.