The objective of this cohort study was to discover how often hospital discharge summaries were available to physicians seeing patients for follow-up visits after hospitalization. Seven hundred and ninety-two patients discharged from an internal medicine service in a teaching hospital in Ottawa, Ontario, after treatment for acute illness were studied. The following data was obtained on these individuals, when and by which physician each patient was seen during the first 6 months after discharge; and, the date each patient's discharge summary was printed and the physicians to whom it was sent. Confirmation that summaries were received was obtained by a survey or by telephoning physicians' offices. Patients were observed for 6 months,or until they were readmitted to hospital. The main outcome measures were the proportion of follow-up visits to physicians for which discharge summaries were available. During the observation period, patients made 6619 visits (median six per patient, interquartile range [IQR] 2 to 9) to 914 different physicians (median three per patient, IQR 2 to 4). Discharge summaries were available for only 996 (15%) visits. Summaries were available for only 65 initial visits (8.2%); no summaries were available for any visit for 542 (68.4%) patients. Summaries were most commonly not available because they were not generated in time for follow-up visits (20.0%) or were not sent to follow-up physicians (50.8%). At this institution, discharge summaries often did not get to physicians seeing patients after discharge from hospital.