Abstract 19417: Method of Risk Assessment has a Substantial Impact on Appropriateness Level Among Symtomatic Patients Referred for Cardiac CT
Introduction: While the current appropriate use criteria (AUC) for cardiac CT (CCT) recommends the Diamond and Forrester method (DF) for assessment of pretest probability of CAD in symptomatic patients, the Duke Clinical Score (DCS) uses a more robust predictive model that integrates multiple clinical variables. Given that pre-test probability is a major determinant of CCT appropriateness, we sought to evaluate how the DF and DCS compare in classifying appropriateness level for CCT in symptomatic patients.
Methods: 114 consecutive symptomatic patients who underwent CCT for evaluation of CAD in native coronaries were classified as having low, intermediate, or high pretest probability of obstructive CAD by DF and DCS. Using the AUC, exam indications were classified based on pre-test probability and prior testing. CCT Results were classified as revealing obstructive (≥70% stenosis), nonobstructive (<70%), or no CAD.
Results: When patients' risk was classified using DF, 18% were low, 65% intermediate, and 17% high risk. Subsequently, exam indications were appropriate for 50%, inappropriate for 4%, uncertain for 14%, and not included in AUC for 32% of exams. When the DCS was used, the pre-test probability of CAD was reclassified for 57% of patients, most of whom changed from intermediate to high risk (29% low, 16% intermediate, 55% high risk). As a result, exam appropriateness was reclassified for 24% of patients, and there were significantly more inappropriate exams and fewer appropriate and uncertain exams (41% appropriate, 17% inappropriate, 7% uncertain, 34% not included; p<.001). Of the 16 exams reclassified as inappropriate, nearly all (n=15) revealed CAD, although most often (n=13) nonobstructive.
Conclusions: When using the DCS instead of DF, more patients had a high pre-test probability of CAD and were thus reclassified as having an inappropriate indication for CCT. While the DCS may have a stronger predictive value for detection of CAD, it is unknown whether this reclassification represents an improved identification of high risk patients not suitable for CCT or an overestimation of disease. These results emphasize that the method of risk assessment used to define appropriateness has important implications for quality assurance programs.
- © 2010 by American Heart Association, Inc.