This research represents a paradigm shift in studies of Clinical Decision Making by moving the focus from three earlier generations of work to a new """"""""fourth generation"""""""" study of cognitive (and reasoning) processes. This theoretically based study builds directly (and cost efficiently) on successfully completed studies which used experimental (factorial) designs and demonstrate that patient attributes (gender, age, race and SES), provider characteristics (gender, medical specialty and years of clinical experience) and health care system characteristics independently influence (most p<.001) the diagnosis and management of Coronary Heart Disease (CHD). Clinical decision-making (CDM) with respect to CHD is shaped as much by who the patient is, who the provider is and the setting in which care is provided as it is by what the patient actually presents (the signs/symptoms of CHD). The study will describe how and explain why (and not simply that) different physicians, in different practice settings evidence variations in a broad-range of CHD care with equivalent patients. This project has 4 aims. (1) To estimate the independent effects of two cognitive experimental manipulations: (a) cognitive """"""""priming"""""""" of physicians as to CHD possibility; and (b) systematic substitution of patient attributes purportedly associated with CHD; on the following primary outcomes: probability of a CHD diagnosis, the number and type of diagnostic features recalled, and the timing of CHD diagnosis. The impact of cognitive intervention (a) on intervention (b) will provide estimated effects of analytic vs. non- analytic reasoning (i.e., intentional vs. non-intentional discounting) as measured by the primary outcomes. (2) To understand how patient attributes intrude on physicians' cognitive reasoning processes to produce the observed CHD variations. (3) To explain, using cognitive analysis, how provider characteristics contribute to the documented CHD variability. (4) To understand how organizational influences also intrude on CHD decision-making. Primary care providers (who encounter most CHD) from New York State will be randomly sampled (n=256) and invited to view a clinically authentic videotaped presentation of a """"""""patient"""""""" presenting with signs/symptoms of CHD. Information concerning the primary outcome (probability of CHD diagnosis) and a range of well-justified secondary outcomes will be elicited through structured interviewing. The factorial design permits estimation of independent (unconfounded) main effects and all two-way interactions. Results from theproposed research could reduce or eliminate health care variation in at least two types of interventions: (a) Organizational and reimbursement policies couldbe developedand evaluated which steer providers away from thepatterns of decision-making which produce healthcare variations; (b) Interventions during medical education, when practice styles are not yet firmly established.
