OBJECTIVES: To audit the diagnostic yield and cost implications of the use of a 'liver screen' for inpatients with abnormal liver function tests. DESIGN: We performed a retrospective audit of inpatients with abnormal liver function tests. We analysed all investigations ordered including biochemistry, immunology, virology and radiology. The final diagnosis was ascertained in each case, and the diagnostic yield and cost per positive diagnosis for each investigation were calculated. SETTING: St Thomas' NHS Trust. PARTICIPANTS: All inpatients investigated for abnormal liver function tests over a 12-month period. MAIN OUTCOME MEASURES: We calculated the percentage of courses due to each diagnosis, the yield of each investigation and the cost per positive diagnosis for each investigation. RESULTS: A total of 308 patients were included, and a final diagnosis was made in 224 patients (73%) on the basis of both clinical data and investigations. There was considerable heterogeneity in the tests included in an acute liver screen. History and ultrasound yielded the most diagnoses (40% and 30%, respectively). The yield of autoimmune and metabolic screens was minimal. CONCLUSIONS: Our results demonstrate the low yield of unselected testing in patients with abnormal liver function tests. A thorough history, ultrasound and testing for blood-borne viruses are the cornerstones of diagnosis. Specialist input should be sought before further testing. Prospective studies to evaluate the yield and cost-effectiveness of different testing strategies are needed.