This research examines the effect of microcomputer-based decision support systems (DSS) on the clinical reasoning of potential users, using QMR and ILIAD as test systems. Two issues are the primary focus: (1) the extent to which users modify their diagnostic opinions and workup plans as a result of interaction with the DSS and (2) the effects of clinical experience upon interaction with the DSS and the resulting consultations. Case materials will be selected and abstracted from those admitted for diagnostic evaluation on the Medicine services of three collaborating medical centers and will be presented to subjects for diagnostic assessment and workup planning, first without and then with the aid of a DSS. The study examines the effects of three independent variables (level of clinical experience, time of case work-up [pre- or post- consultation] and case cluster) on seven dependent variables: selection of relevant information for input into the DSS, user's diagnostic accuracy, confidence of the user in the diagnosis, cost and appropriateness of the planned next steps, diagnostic volatility, and system's diagnostic accuracy. 36 cases will be used, assigned to 4 clusters; all cases will have diagnoses in the knowledge base of two decision support systems. The subjects will be fourth-year medical students, second-year residents in internal medicine, and experienced general internists. By studying how consultations with computer-based DSSs are utilized by and affect the reasoning of clinicians with varying levels of clinical experience, this research contributes to understanding the role of decision support systems in improving diagnostic reasoning and clinical care, identifying how much clinical experience is needed to use them effectively, and assessing their potential for reducing unnecessary and expensive diagnostic procedures.
