A probability database for decision-analytic models of coronary revascularization procedures.

The time required to extract probabilities from medical literature is a primary reason decision analysis is not used more frequently for individual patient management decisions. Objective clinical trial information from the medical literature for one management decision was placed in a database which provided probabilities when queried. The database was tested with decision-analytic models of specific patient cases from the medical literature. Performance was assessed in terms of the ability to select trials which resembled the patients' individual characteristics, the number of trials providing probabilities for a given outcome, and the number of follow-up points available for that outcome. The timely assistance the database provides in expediting literature review and synthesis could enable the more common use of decision analysis in management decisions for individual patients.