Evaluation of clinical decision-aids, with an application to a system for dyspepsia.

The evaluation of clinical decision-aids is discussed both in general and in the particular case of a microcomputer-based system designed to help general practitioners in deciding whether to refer patients with dyspepsia for specialist care. Parallels are drawn with the established framework for evaluating a new therapy, including a description of four phases of evaluation and the contrast between pragmatic and explanatory trials. Past studies are reviewed, emphasizing the difficulties of establishing a credible control group, and including some relevant examples from the areas of evaluating new diagnostic technology, medical information systems and audit procedures.