Epidemiology and biostatistics are difficult to teach in medical school. To determine the feasibility of applying a collaborative learning model, we randomly assigned one section of an epidemiology and biostatistics class to use this teaching model. Students in this section did not attend lectures or follow the regular curriculum but instead learned epidemiologic concepts by identifying clinical problems of interest to them and reviewing relevant journal articles, with the group leader functioning as facilitator. All students took an identical examination and anonymously evaluated the course. Examination performance was comparable between the study group (mean 72.1%) and the other students (mean 73.5%). However, the students' evaluation of the study group method was more favorable, especially in their perceived mastery of specific cognitive skills. The data suggest that this model can be successfully applied in teaching epidemiology and biostatistics to medical students.