This presentation is part of: Oral Concurrent Session A - Clinical Strategies and Guidelines
EVALUATING CLINICAL GUIDELINES AND PREDICTIVE INSTRUMENTS FOR ACUTE CARDIAC ISCHEMIA
Britain Mills, BS1, Steven Estrada, BS1, Valerie F. Reyna, PhD1, Jayni Parikh, BS2, Michael Pham, BS2, and Kristin Poirier, BA3. (1) University of Texas at Arlington, Psychology, Arlington, TX, (2) University of Arizona, Informatics and Decision Making Laboratory, Tucson, AZ, (3) University of Arizona, Surgery, Tucson, AZ
Purpose: To compare clinical guidelines and predictive instruments for acute cardiac ischemia. Methods: Two samples of emergency-room patients presenting with non-traumatic chest pain or pressure were classified as low, intermediate, or high risk according to American Heart Association/American College of Cardiology unstable angina guidelines, Agency for Health Care Policy and Research (AHCPR) guidelines, Acute Cardiac Ischemia Time Insensitive Predictive Instrument (ACI-TIPI), and physicians’ triage decisions (discharge, ward or monitored bed, or cardiac intensive care). Sample 1 consisted of every fifth consecutive patient eligible for enrollment (N=1004) and Sample 2 consisted of patients for whom physicians had returned a detailed questionnaire (N=1028); overlap was 19%. Ethnic breakdown was 72% White-Non-Hispanic, 18% White-Hispanic, and 10% other. Of the 1028, 333 patients returned to the hospital for a cardiac procedure (e.g., percutaneous transluminal coronary angioplasty or coronary artery bypass graft) and/or received a cardiac diagnosis (e.g., acute myocardial infarction, unstable angina, or stable angina) within a year of their initial visit. Of the 1004, 200 patients returned within a year. Sequential logistic regression analyses with occurrence of a cardiac procedure as a criterion variable were conducted, inputting coronary artery disease (CAD) risk level for the older and newer guidelines, acute myocardial infarction (AMI) risk level for the older and newer guidelines, as well as level of triage for initial visit and ACI-TIPI scores as predictors. Cardiac diagnosis was similarly analyzed. Results: Occurrence of a cardiac diagnosis was predicted by the older AHCPR guidelines but not by the newer guidelines. Physicians’ initial triage decisions, ACI-TIPI scores, AHCPR risk levels, and newer guidelines’ AMI risk level predicted procedures—and each contributed unique variance. Examining procedures and diagnoses separately, the older guidelines consistently outperformed the newer guidelines in assigning higher levels of risk to cardiac outcomes. Conclusions: Although the newer guidelines were designed to improve on the older guidelines, empirical comparisons revealed that the newer guidelines had lower predictive validity for the occurrence of subsequent cardiac diagnoses. Paradoxically, although the newer guidelines increased levels of risk, lack of resolution contributed to difficulty in predicting cardiac outcomes. These results indicate that changes in guidelines should be empirically evaluated before widespread implementation. They also indicate that the use of multiple, related guidelines in concert with physicians’ judgments can improve prediction of cardiac risk.
