Accepted for publication August 1990 The last two or three years have seen an explosion of interest medical audit, phrase hitherto barely used the UK except by few hardy pioneers (for references, see1 2). Then quick succession the last two years appeared the Confidential Enquiry on Perioperative Deaths,3 the Report of the Royal College of Physicians on Medical Audit,4 and the White Paper Working for patients.5 The White Paper provided helpful definition of audit: a systematic critical analysis of the quality of medical care, including the procedures used diagnosis and treatment, the use of resources, and the resulting outcome for the Since the publication of these documents, there have been several conferences about medical audit, and local health authorities the UK should by now have established District Audit Committees. It is perhaps useful time to take stock of the situation, and of the different approaches to medical audit, and attempt to predict the way which matters are likely to develop over the next few years. Conceptually, perhaps the simplest form of audit is review of the medical record. The lessons learnt from the 12 years' experience the Departments of Medicine and Clinical Pharmacology at Birmingham University are described recent article by Heath.1 The advantages of such an audit are that no complex information systems are required, and that review of records, best selected randomly, or of patients with single diagnosis or undergoing single procedure, can be accomplished friendly non-confrontational manner that is enjoyed by many participants. The physicians at Birmingham found that the recording of the initial symptoms of the illness and of the patient's subsequent progress rapidly improved, as did the record of what had been said to the patient. However, once the meetings were well established and the problem of poor quality notes tackled, difficulty arose due to the repetitive nature of the procedure. It became apparent that for audit to continue to be valuable it had to be supported by senior consultants who believed its benefit and who were regular attenders at audit meetings. The Royal College of Physicians now requires that when visiting team from the Training Office reviews post for suitability for General Professional or Higher Medical Training, random sample of records is drawn from the hospital by the visiting Regional Adviser, and the quality of the records inspected this way. The Regional Advisers have reported after the first year of this procedure that there are considerable deficiencies recording what is said to the patient, although general the quality of the initial history and the record of progress appear to be adequate. There seems little doubt from the Birmingham experience that such record review does improve the quality of record keeping, which is essential bearing mind the increasing fragmentation of care amongst different specialist teams, and the shortening of hours of junior doctors recent years, so that patients are often looked after for long periods by young doctors who are not on the team responsible for that patient's primary care. However, review of small sample of clinical records by senior physician, however distinguished, can only give general impression of the quality of care an institution. In the United States medical record (chart) review is more formal. Peer Review organisations are contracted to the Health Care Financing Administration to fulfil statutory obligation that care to Medicare patients will be provided economically and only to the extent medically necessary, will be of quality which meets professionally recognised standards of health care, and will be supported by evidence of medical necessity in such form and fashion and at such time as may reasonably be required by reviewing Peer Review organisation the exercise of its duties and responsibilities. As may be expected from this last clause, vast bureaucratic organisation has grown up to review medical records by registered chart reviewers. The Peer Review organisation for the State of Massachusetts, for example, draws sample of 3?0 of all Medicare discharges for chart review, employs full time medical director, four associate directors, 70 chart reviewers who are registered nurses, and has budget of $7m. The chart reviewers look at the medical record against so called generic quality screens, an example of which is published elsewhere.6 Many of the screening items are comparatively trivial, but the Peer Review organisation system does often reveal misuse of antibiotics, particularly lack of care prescribing aminoglycosides, failure to use antibiotics accord with laboratory determined sensitivities, and failure to seek specialist advice when appropriate. The Peer Review organisations have powerful legal sanctions, including witholding Medicare payments on occasion. However, many physicians the United States are not convinced that the system uses resources