Measuring the complexity of clinical problems.

Recent studies on clinical problem-solving show that a considerable proportion of the problem-solving strategy employed by doctors and medical students is "case specific" or "system specific." Such studies have failed to consider adequately the starting place of the problem-solving process, that is, the inherent complexity of the problem being considered. In presenting clinical problems to students and writing problem-solving skills assessments, educators have demonstrated no quantifiable ways of grading the complexity of the clinical problems they select for instruction or examination purposes. The author in this study reports on the development and application of a complexity index to 29 clinical problems used in a new graduate curriculum. The index uses five components of problem complexity: symptoms, physical signs and investigations, socioeconomic and behavioral factors, diagnoses, and management plans. A test has been made of its interrater reliability between two observers applying it independently in one application. Trials to establish the validity of the complexity index with various criterion groups of physicians and educators are discussed. Suggestions for further refinement of such an index and its application in research in the quality of care and in medical education are offered.