defining evidence

Last reviewed 03/2021

Evidence-based medicine attempts to identify evidence from clinical and health care research that can be applied by medical and health professionals.

A major tool in the identification of "evidence", especially relating to therapeutics, is the randomised controlled trial. Evidence may be available from individual randomised trials e.g. ISIS trials, or via meta-analysis of several trials e.g. the effectiveness of specialised stroke units published in the Cochrane Database of Systemic Reviews. Evidence relating to epidemiology can be identified from cohort studies such as the Nurses' Health Study.

The defining of evidence relating to a particular subject area must, as far as possible, be complete. Therefore the process may include both published and unpublished data.

This is not a textbook on evidence-based medicine and so there is no need to detail rigorous guidelines on how a study is included/excluded from an analysis. The authors of this textbook instead use the products of evidence-based medicine as one of the sources for maintaining the validity of the knowledge within the database.