Some points on the use of 'independent Bayes' to diagnose acute abdominal pain.

This paper outlines some of the difficulties encountered when using independent Bayes as a statistical decision aid for acute abdominal pain. Methods of reducing the resultant problems are suggested. Restriction of the number of facets in the system reduces violation of the simplifying assumption of symptom independence without adversely affecting efficiency. Introduction of a realistic utility structure is investigated as is the potential transportability of the decision aid. Generalizations of the results are discussed and potential uses of medical decision aids are considered.