Computer generated discharge summaries and their use as a case mix sensitive audit engine. A tale of two cities.

Background: An 'audit engine' allows critical appraisal of clinical practice without necessitating cumbersome data input, editing or analysis. This is achieved by capturing data from an otherwise necessary task, in this case writing discharge summaries, and using standardised analyses to illustrate the effects of operational changes. ■ Design and setting: Retrospective analysis of clinical outcome of 1,829 sequential discharges from one consultant's team in two geriatric medicine departments. ■ Main outcome measures: Mortality, discharge destination and functional performance. ■ Results: Median length of stay in the two departments differed significantly (8 vs 13 days; p <0.0001), but patients in the latter department were more disabled, with almost twice as many needing domiciliary services after discharge and suffering impaired mobility or incontinence. Despite this disparity, a similar proportion of survivors was placed in institutional care (31/300 (10%) vs 94/1,100 (8%); NS). ■ Conclusions: This audit engine demonstrated that apparently worse performance indicators were explained by adverse case mix in one department, and the similar institutionalisation rates suggest superior care there.