Abstract
How should clinicians select the specific drug, operation, splint, exercise, or conversations that will best achieve the therapeutic goal for a given patient? This essay will examine three strategies for doing so: induction from our own, individual prior experiences; abdication to the authority of our teachers and those who write review articles; and deduction from the reports of randomized clinical trials. By means of examples from the recent past, the fallibility of the first two approaches will be illustrated. Then, strategies for critically appraising the published reports of randomized trials will be described. Finally, the reasons why the results of even the proper trials may not be accepted by frontline clinicians will be introduced.