How accurate is a computerized surgical audit when resident medical staff collect the data?

Since July 1988 all eight general surgeons at Fremantle Hospital have used a computer-based surgical audit and discharge system. At the time of writing (September 1991) 10,919 computer-generated discharge letters have been produced by the system. This paper describes the system and reports a series of quality control assessments carried out between 1 July 1988 and 30 June 1990 during which 30 pre-registration surgical residents completed 5,716 data collection forms. It was found that: (1) data collection for 23 of 24 monthly surgical audits was at least 95% complete; (2) outstanding surgical discharge summaries were reduced by 89%; (3) the residents recorded 17/19 wound infections and identified 15 (79%) of these as a surgical complication; and (4) the residents tended to under-record complications in patients who had more than one complication during their hospital stay. It was concluded that the system was robust, and that resident staff collected data in such a way that good quality computer-generated discharge letters were produced in a timely manner. Closer attention to aspects of data collection will be required before the optimum surgical audits of the QX system can be generated.