Abstract 227: Patient and Provider Generated Appropriateness Ratings for Myocardial Perfusion Imaging: Agreement With Published Criteria

Abstract

Background: Appropriate use criteria (AUC) for myocardial perfusion imaging (MPI) help doctors decide which patients may benefit from MPI. The AUC development process does not include patients. Ordering providers are included, but their opinions on AUC have not been more broadly studied. We conducted an anonymous survey of patients and providers on the appropriateness of MPI in various settings. We hypothesize that both groups will overestimate the appropriateness of MPI compared to AUC. Methods: We administered our survey to patients in primary care clinics, cardiology clinics, and nuclear medicine laboratories at our academic medical center and affiliated Veterans Affairs medical center. Participating providers were physicians, nurse practioners, and physician assistants within primary care, cardiology, and hospital medicine. The primary outcome was a comparison of patient and provider appropriateness ratings as compared to published AUC. Respondents used the same methodology as used to develop the AUC; five clinical scenarios were rated on a 1 (inappropriate) to 9 (appropriate) scale. Scenarios were based on common MPI indications and included appropriate, inappropriate, and uncertain examples (symptomatic with prior revascularization, symptomatic with low risk of heart disease, asymptomatic with low risk of heart disease, symptomatic with intermediate risk of heart disease, and asymptomatic with prior revascularization). We used subgroup analyses and logistic regression models for both groups to investigate associations between ratings and respondent characteristics. Results: The survey was completed by 449 respondents, 338 patients and 111 providers between June and August 2014. Patient sampling was distributed between cardiology clinic (32.2%), primary care clinic (36.0%), and nuclear medicine (31.9%) laboratory. Providers were cardiology (35.1%), primary care (37.7%), hospital medicine (22.8%), and other (4.4%); 86.8% physicians. Of the five scenarios, the patient and provider ratings differed from AUC for three. In two cases, the patient generated rating was higher than the provider or published rating (5 versus 3.5 and 3, p<0.0001; 5 versus 2 and 1, p<0.0001). In the third case, the provider generated rating was lower than both the patient and published rating (2 versus 6 and 6, p<0.0001). Subgroup comparisons and logistic regression did not reveal any significant correlations between ratings and respondent characteristics. Conclusion: In some clinical scenarios, patients and healthcare providers have differences of opinion about the appropriateness of nuclear MPI as compared to published AUC. Patients tended to overvalue MPI while healthcare providers undervalued one of the published AUC ratings. Evidence of differences in opinion raise the question as to what role patients should play in future iterations of AUC.