BACKGROUND Because of the great interest in developing ambulatory care components within medical school curricula, there is a need for outcome data concerning such changes. METHODS In 1991-92, 40 third-year students at the University of California, San Francisco, School of Medicine were assigned to the San Francisco Veterans Affairs Medical Center for their core medicine clerkships; ten of these students were exposed to ambulatory care by being assigned to medical residents who worked in a general medicine continuity clinic. The ten students in the continuity clinic and the 30 who were not were evaluated by means of their (1) final clerkship evaluation scores for overall competence, (2) performances on a standardized-patient exercise where the majority of the simulations were of ambulatory care problems, and (3) self-assessment of clinical skills and knowledge and attitudes about primary care. Student's t-test was used to determine whether there were statistically significant differences between the performances of the two groups. In addition, the students in the clinic kept diaries about their clinic experiences. RESULTS No significant difference was found between the two groups of students on the objective or subjective measures of evaluation, although the diary narratives of the clinic students revealed that they were extremely enthusiastic about the clinic experience. CONCLUSIONS Because there was no significant difference between the performances of the two student groups, the ambulatory care model in this study, despite its feasibility to implement, was not adequate to teach the students the different problem-solving and therapeutic skills necessary in the ambulatory care setting.