eFigure. Distribution of Appropriate Use Criteria Categories
eTable. Common Overall and Inappropriate Indications
Winchester DE, Meral R, Ryals S, Beyth RJ, Shaw LJ. Appropriate Use of Myocardial Perfusion Imaging in a Veteran PopulationProfit Motives and Professional Liability Concerns. JAMA Intern Med. 2013;173(14):1381–1383. doi:10.1001/jamainternmed.2013.953
Myocardial perfusion imaging (MPI) is performed millions of times annually in the United States to assess patients for coronary ischemia. Some have expressed concern that MPI is being used inappropriately, possibly because of self-referral profit motives and professional liability fears.1 To inform clinicians about situations in which patients are likely to benefit from MPI testing, appropriate use criteria (AUCs) for MPI were developed, last revised in 2009.2 Prior investigations have applied AUCs to describe the magnitude and patterns of inappropriate testing. Rates of inappropriate testing have ranged from 7% to 24%.3 We hypothesized that the single-payer environment of the Veterans Affairs (VA) health system, which eliminates self-referral profit motive and limits liability concern, will result in less inappropriate use of MPI.
We conducted a retrospective cross-sectional investigation of MPI ordered in a single VA medical center. Our institutional review board reviewed the protocol and waived the requirement for informed consent. Participants were identified from records of MPI performed using the VA Computerized Patient Record System, beginning December 2010. Demographic information, medical history, and information regarding symptoms were gathered to establish associations with inappropriate MPI ordering. Appropriateness categorization was performed on the basis of the 2009 criteria2 and using a data collection instrument endorsed by professional societies and the Imaging in FOCUS initiative for assessment of appropriateness. Patient symptoms that prompted ordering of MPI were attributed to ischemia, unless specifically indicated otherwise. The primary outcomes were the proportion of inappropriate MPI tests and univariate associations between patient characteristics and the likelihood of an inappropriate MPI. Statistical analysis was performed using SPSS, version 20 (IBM). We used the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE)4 statement in designing our investigation.
We identified 332 MPI studies performed between December 1, 2010, and April 11, 2011. The population of patients was predominantly male with high prevalence of obesity, diabetes mellitus, hypertension, and hyperlipidemia (Table). Coronary artery disease and abnormal electrocardiogram results were also common findings. Chest pain was the most commonly reported individual symptom (49.4%); however, typical angina was rare (0.6%). Positron-emission tomography was used for 304 patients (92%) and single-photon–emission computed tomography for the remaining 28 patients (8%).
For all but 4 patients (1%), an indication from the 2009 AUCs could be identified. Study indications were 78% (n = 259) appropriate, 13% (n = 42) inappropriate, and 8% (n = 27) uncertain (eFigure in Supplement). The most common inappropriate MPI indications included testing of patients with low pretest probability who could have undergone treadmill electrocardiogram testing (7 patients [16.7% of total inappropriate MPI]) and asymptomatic patients with low coronary heart disease risk (7 patients [16.7% of total inappropriate MPI]) (eTable in Supplement). Of the 9 preoperative MPI tests reviewed in this investigation, 6 were inappropriate and 3 were appropriate.
Patient characteristics were compared with MPI indications to identify factors associated with inappropriate MPI ordering (Table). The absence of symptoms was associated with a nearly 5-fold higher likelihood of inappropriate testing (odds ratio [OR], 4.80 [95% CI, 2.39-9.66]; P < .001). Both chest pain (OR, 0.07 [95% CI, 0.02-0.20]; P < .001) and diabetes mellitus (OR, 0.37 [95% CI, 0.17-0.80]; P = .01) were associated with a lower likelihood of inappropriate testing.
In this retrospective cross-sectional investigation regarding the appropriate use of MPI in a VA health care setting, we observed that a substantial portion of MPI tests were ordered for inappropriate indications. The findings are in contrast to our initial hypothesis but are similar to those of another VA-based investigation, the results of which were published during our investigation.5
Our hypothesis was based on unique characteristics of the VA patient care environment. First, no self-referral or profit motives exist. Second, whereas the Federal Tort Claims Act permits medical malpractice lawsuits against federally employed physicians, the substantial majority of claims are resolved through administrative processes. Only 3 judgments against the US government were recorded for tort claims in 2010.6 We did not detect a significant reduction in inappropriate testing in the VA environment, which suggests a lesser role of defensive medicine and self-referral in the inappropriate use of MPI. A recent survey of cardiologists did not find any association between ownership of diagnostic equipment (MPI, catheterization laboratory, computed tomography scanner, or echocardiography laboratory) and concordance with appropriateness categories.7
The magnitude and pattern of inappropriate testing in our investigation are similar to those in prior reports.3 Hendel et al8 observed similar common inappropriate MPI indications, as well as significant associations between asymptomatic patients and inappropriate testing (OR, 22.5 [95% CI, 15.2-33.2]) and lower likelihood of inappropriate testing in patients with diabetes mellitus (OR, 0.4 [95% CI, 0.4-0.5]).
Reasons for the observed patterns of ordering MPI are unclear. Conceivably, commonalities in medical training, independent of postgraduate practice environment, could contribute to an exaggerated perception of benefit of MPI in asymptomatic patients and those at low risk of coronary heart disease. This exaggerated perception of the benefit would also seem to hold true for preoperative risk assessment, with the majority of preoperative MPI in our study having been inappropriately ordered.
Corresponding Author: David E. Winchester, MD, MS, Medical Service, Malcom Randall VA Medical Center, 1601 SW Archer Rd 111D, Gainesville, FL 32608 (firstname.lastname@example.org).
Published Online: June 10, 2013. doi:10.1001/jamainternmed.2013.953.
Author Contributions: Dr Winchester had full access to the data and accepts responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Winchester.
Acquisition of data: Winchester, Meral, Ryals.
Analysis and interpretation of data: All authors.
Drafting of the manuscript: Winchester, Shaw.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Winchester.
Obtained funding: Winchester.
Administrative, technical, and material support: Meral, Ryals.
Study supervision: Winchester, Beyth, Shaw.
Conflict of Interest Disclosures: None reported.
Funding/Support: This investigation was supported by NIH T35 Training Grant T35-HL007489-28. This material is based on work supported by the Department of Veterans Affairs, Veterans Health Administration.
Role of the Sponsor: The views expressed in this article are those of the authors and do not necessarily reflect the position or policy of the Department of Veterans Affairs or the United States government.
Previous Presentation: This study was presented at the American College of Cardiology Scientific Session; March 11, 2013; San Francisco, California.