Purpose: To assess the variation in outpatient educational experiences among residents in a primary care internal medicine residency program. Method: The authors conducted an observational study within a primary care residency program in Boston, Massachusetts, involving eight primary care residents in 2001-02. A data management system was created that uses information on clinical experiences collected from an electronic medical record (EMR). The EMR records clinical information from patient encounters in resident continuity clinics, including patient demographics and diagnostic codes entered by the residents. Primary and secondary diagnosis from the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) were assigned to clinical categories (e.g., cardiology, orthopedics) and these assignments were programmed into a spreadsheet that could take the diagnostic data directly from the EMR downloads and produce reports using only the primary diagnosis, using all diagnoses, or using the assigned clinical categories. Results: The EMR download showed 2,468 patient encounters for the eight residents in the primary care program. The most common clinical encounters were for health maintenance (17%), cardiology (14%), and women's health (10%). In contrast, rheumatology (0.5%) and hematology/oncology (2%) encounters were less frequent. There were substantial variations among residents in terms of distribution of encounter diagnoses, and the age and gender of patients seen in continuity clinic. Conclusions: Abstracting data from an EMR represents a feasible method for assessing programmatic and individual learner experiences in the outpatient setting. Such information may help target curricular adjustments to ensure an appropriate diversity and depth of clinical training.