Computers in audit: servants or sirens?

In Greek mythology the song of the Sirens was of such beauty and promise that no sailor could resist steering his ship towards them. Instead of the promised heaven, however, the route led to disaster upon the rocks. Audit has developed rapidly over the past three years in Britain from an activity carried out by only a few enthusiasts'4 to one in which all doctors are enjoined to take part.5 Prominent among the guidance to clinicians are documents on the choice of computer systems for audit,67 and several papers on audit have emphasised the role of microcomputers.2489 There are now many commercial audit packages that not only offer help in audit but will also write discharge letters, manage waiting lists and theatre lists, and enable the com-puterisation of medical records for research.' The attraction of being able to combine audit with all these other activities might appear irresistible. We argue that microcomputer based audit packages, far from offering solutions to audit problems, are more akin to the Sirens luring the unwary auditor to disaster. Fallacy of the common core of data Audit, patient administration systems, and clinical research may seem to have a common core of data as all collect basic demographic details and information on investigations, diagnosis, and treatment. If audit is to be performed anyway, why not collect a little extra data and achieve a great deal more? In reality, although there is a degree of overlap in the data sets collected, the differences among them are greater than the similarities. The essence of good audit is to focus on a specific topic and to collect only essential information on a sample of patients over a limited time.' A patient administration system extends the time of data collection to include all patients and embraces a wide range of topics on which it can collect only a limited amount of data. A system of computerised case notes also surveys all patients but collects detailed information on all topics. Clinical research is in fact closest to audit and differs only in the amount of detail in which the data are collected. These two activities, however, differ in their purposes: audit is concerned with improving the quality and efficiency of medical care whereas research focuses on generating or testing hypotheses. Audit seeks to ensure that current kn'owledge is properly and fully used; research entails the discovery of new knowledge. complexity of …