No abstract

Summary The 1974 paediatiic certification process of the Royal College of Physicians and Surgeons of Canada included four CPMPs, in addition to the conventional MCQs and orals. The CPMPs were successfully administered to 160 candidates in eight cities across Canada on the same day. The examination was judged by the candidates to be a better test of clinical skill than MCQs. The CPMP examination did not penalize French‐speaking candidates, was economically feasible, and a good security risk. Correlations between MCQ, orals and CPMPs, indicated that each examination measured some aspects of paediatric competence not tested by the others.

No abstract

A feasibility study is described involving the use of a computer program in an antenatal clinic whose decisions were based on the observed policies of the group of obstetricians running that clinic. In 200 test cases the programme successfully detected all »at risk« cases at the first visit. Abnormalities appearing at later visits were all detected and appropriate investigations ordered by the programme. The programme ordered rather more investigations than the obstetrician and this difference has been investigated, in some detail with reference to the full blood count, glucose tolerance test and the ultrasonogram.Since all of the basic data in the proposed clinic would be collected by the clerical staff or midwives, a comparison was made between two midwives and an obstetrician in estimating fundal height. The small number of discrepancies found was not regarded as serious from the viewpoint of the proposed clinic.This preliminary study suggests that the technique described might play a useful part in some areas of ante-natal care by relieving the obstetricians of much of the workload associated with routine cases and enabling the midwife to play a larger role in ante-natal care.

Attempts have been made for almost two decades to use computers to make clinical diagnoses. Interest in such techniques persists despite little practical success in their application to routine patient care. The explanation of this lack of success is complex and includes (i) lack of understanding of the decisionmaking process itself and (ii) the need for a theoretical framework within which such systems can be developed. A major objective of this paper is to illustrate the contribution that the concepts and techniques associated with Decision Theory can make to both of these problems.After attempting to analyse the deficiencies of current approaches to computer-assisted diagnosis the paper discusses some of the basic concepts of decision theory. It then illustrates the role of decision theory in studies using different forms of clinical decision analysis.Finally the paper indicates the relevance of decision analysis to medical education and to the provision of health care.