Assessing Disease Class-Specific Diagnostic Ability : A Practical Adaptive Test Approach / Frank J. Papa and Randall E. Schumacker.

Measures of the robustness of disease class-specific diagnostic concepts could play a central role in training programs designed to assure the development of diagnostic competence. In the pilot study, the authors used disease/sign-symptom conditional probability estimates, Monte Carlo procedures, and artificial intelligence (AI) tools to create test items (case vignettes) representing varying levels of typicality for the disease class known as myocardial infarction (heart attack). The typicality estimate assigned to each test item was converted to a Rasch logit scale value representing its difficulty level. Selected test items were then embedded within a paper-based examination and the performance of 628 first-year postgraduate residents-in-training determined for each item. The residents' performance was then simulated in the context of a practical adaptive testing (PAT) format. Results from residents for the actual paper-based and simulated PAT are compared and discussed. These two testing formats are also discussed in terms of their use to measure the robustness of disease-specific diagnostic concepts. An appendix explains a simulation procedure. (Contains 1 figure, 1 table, and 17 references.) (SLD)

Microfiche. [Washington D.C.]: ERIC Clearinghouse microfiches : positive.