Two recent epidemiologic case-control studies suggested that fenoterol, a selective beta-adrenergic agonist, was associated with an increase in the risk of asthma death. The results of these studies were criticized because of methodologic problems in the choice and selection of control subjects; the different methods used to gather exposure data in cases and control subjects; and because of inadequate classification and adjustment for asthma severity. In response to this controversy, a new study is underway, the Saskatchewan Asthma Epidemiology Project. The SAEP includes two complementary studies, an historic cohort and a case-control analysis, that employ the computerized databases of the Saskatchewan Health Department. A unique aspect of the SAEP is the attempt to incorporate knowledge of asthma physiology and management into the design of the studies. Specifically, the study design recognized the role of antiinflammatory drugs in asthma treatment; the distinction between asthma death and near-fatal asthma; the severity of asthma; patterns of drug use; and the distinction between inadequate clinical care and disease severity. The strategies we employed in the SAEP may prove helpful to investigators whenever clinical and biologic processes create sources of potential bias requiring special procedures for the design and analysis of epidemiologic studies.