BACKGROUND--The demand for open heart surgery has driven current practice towards early postoperative discharge and interhospital transfer to maximise patient throughput. The extent to which this redirects morbidity to other healthcare providers is unknown. OBJECTIVE--To define the incidence of inhospital and early postoperative morbidity within 6 weeks of primary hospital discharge after cardiac surgery. DESIGN--Prospective inhospital data for 322 consecutive adult patients undergoing cardiac surgery were compared with retrospective information obtained by postal questionnaire. RESULTS--Mean (SD) primary postoperative hospital stay was 8.3 (3.1) days. There were 13 inhospital deaths (4%), and three patients died within 6 weeks of primary discharge. Retrospective information was obtained from 297 patients (96%). Of these, 77% patients were discharged home directly, while 23% were transferred to other hospitals for continued medical care. Mean (SD) hospital stay after transfer was 12 (8.4) days and required 741 additional hospital bed days. Thirty nine patients (13%) were readmitted to hospital, requiring a further 275 hospital bed days. The readmission rate was lower in patients sent home directly (10%), than in those who were transferred (22%; P < 0.001). CONCLUSIONS--Inhospital audit underestimates early morbidity after cardiac surgery. The burden transferred to other healthcare providers is considerable and has important financial implications for purchasers.