The use of expert systems in the healthcare industry

The large, enormously complex expert systems (ES) such as MYCIN, ONCOCIN, and INTERNIST raises the expectations of many that artificial intelligence (AI) technology would have a significant impact on the everyday practice of medicine. A rather eclectic literature search reveals that while hundreds of applied expert systems have been prototyped, these have not led to a groundswell of interest among either practicing physicians or healthcare administrators. To account for this apparent failure of expert systems to deliver on their promised capabilities, three key factors are explored: physician disinterest; the low priority assigned to AI among healthcare information systems executives; and the relatively new competitive market pressures affecting health services delivery. A domain of the healthcare industry where ES and knowledge-based technology has taken hold, i.e. the managed care/utilization review arena, is described. This new application of ES technology has the potential to influence standards of medical practice through ‘medical necessity criteria’ embedded in the knowledge base, and to disseminate key findings from research on medical treatment outcomes. Ethical and legal issues surrounding the appropriate use of this technology remain as yet untested and unresolved, however.