Since its first implementation in a medical programme at McMaster University, Canada, problem-based learning (PBL) has become a well-established means of teaching and learning medicine. Extensive research has been conducted and a number of strengths of the method are well supported. Several items, however, remain unclear although there is evidence that no relevant difference exists in factual knowledge among students from PBL and traditional curricula, a controlled, randomized study has not been conducted to address this issue. The Medical Faculty of the University of Cologne is in the process of integrating elements of PBL into its curriculum. In the spring term of 1997, after seven semesters of experience with PBL supplementing the traditional course of basic pharmacology, we did for the first time use PBL instead of the lecture-based course (LBL) and conducted a controlled prospective study to determine the effects of this intervention. One-hundred and twenty-three students were randomly assigned to either PBL (n = 63), with tutorial groups of up to nine students, or to the traditional, lecture-based course (n = 60). Analysis of the results of both groups in the examination of basic pharmacology, consisting of multiple-choice and short-essay questions, revealed similar scores with a tendency favouring PBL students in the category of short-essay questions. Hence, it seems clear that PBL does not imply a disadvantage in terms of factual knowledge. Students considered PBL to be an effective learning method and favoured it over the lecture format. Furthermore, students reported positive effects of PBL in terms of use of additional learning resources, interdisciplinarity, team work and learning fun.