From 1985-2000, 58 Dutch family physicians (FPs) of the Transition Project collected ICPC-coded data on 47, 2451 episodes of care, first in paper records for direct encounters only, later with a complete electronic patient record (EPR) for all (direct and indirect) encounters. Based on these data, the effects of a long observation period and the inclusion of all encounters (both direct and indirect) in the EPR were studied. Long observation periods in EPRs appear to have important advantages for patient documentation and the assessment of the content of family practice. Comparison of data from a one-year versus a four-year observation period showed a shift in utilization rates. In a long observation period, visiting patients appeared to make less demand on care, while the content of the FP's care for selected chronic diseases was reflected more realistically. The inclusion of all indirect encounters in an EPR (as compared to the previous inclusion of direct encounters only) resulted in more than twice the number of coded entries per listed patient, and thus led to a major shift in perspective on the FP's involvement in patient care.