Criteria for patient eligibility in a randomized clinical trial can be used to improve the design of observational case-control studies. The new strategy has been illustrated in a case-control investigation of whether lidocaine prophylaxis prevents death in patients with acute myocardial infarction. From among patients hospitalized with myocardial infarction in a special care unit during 1974 to 1978, there were 151 fatalities (cases) and 151 survivors (controls) selected for study. After removal of the ineligible cases and controls, lidocaine prophylaxis was found to have no effect on death from pump failure or nonarrhythmic causes, but was significantly protective against death from ventricular arrhythmias, particularly in patients with congestive heart failure. If confirmed by rigorously designed longitudinal studies, the improved case-control technique could serve as a useful "screening" device to assess the efficacy of medical or surgical treatments that cannot be tested with randomized clinical trials.