Objectives: Appointment-keeping after hospitalization is a poorly understood link between inpatient and outpatient care. We investigated how health care system and patient characteristics influence appointment-keeping after discharge from an acute care hospitalization. Design: Prospective cohort study. Setting: Urban public teaching hospital. Subjects: All 372 consecutive eligible patients admitted over a 15 week period to medicine wards. Methods and Measures: We interviewed patients during hospitalization and after discharge, searched the hospital's electronic databases, and reviewed charts. We measured medication compliance, health care access and use, health status (SF-36), previous appointment compliance, and physician recommended follow-up appointments. Main out-come was appointment adherence after discharge. Results: Patients were primarily African American (71%), uninsured (64%), female (53%), and had a mean age of 48 years; 64% of first appointments after discharge were kept. Adjusted odds ratios (95% confidence intervals) for appointment-keeping were 3.3 (1.7, 6.5) for receiving a written appointment at discharge, and 0.50 (0.27, 0.90) for previous difficulty with obtaining health care. Readmission rates were not associated with appointment adherence. Conclusion: Modifiable system, as well as patient, characteristics are associated with follow-up appointment-keeping. The practice of not giving patients written appointments at the time of discharge may constitute an implicit form of "rationing by inconvenience." Further studies should also evaluate potential associations between appointment-keeping and re-hospitalization.