Appropriate Use of Myocardial Perfusion Imaging in a Veteran Population: Profit Motives and Professional Liability Concerns
What is the frequency of inappropriate stress nuclear myocardial perfusion imaging (MPI) in a single payer environment?
The authors retrospectively reviewed MPI studies performed during a 4-month period at a single US Veterans Administration (VA) hospital. Studies were characterized as appropriate or inappropriate based on the 2009 ACCF/ASNC/ACR/AHA/ASE/SCCT/SCMR/SNM appropriate use criteria (AUC) for radionuclide imaging. Univariate predictors of inappropriate testing were identified using logistic regression.
Out of 322 MPI examinations performed during the study period, positron emission tomography (PET) was performed in 304 (92%) patients. The remainder were studied using single-photon emission computed tomography (SPECT). Indications could be classified per the 2009 AUC in 318 of 322 patients (99%). Studies were classified as appropriate in 259 (78%) patients, inappropriate in 42 (13%), and uncertain in 27 (8%). The most common reasons for inappropriateness were: low pretest probability, which could have been evaluated with exercise testing without imaging (n = 7); and asymptomatic status with low coronary heart disease risk (n = 7). The primary predictors of inappropriate testing were asymptomatic status (odds ratio [OR], 4.8; 95% confidence interval, 2.4-9.7; p < 0.001). Predictors associated with lower odds of inappropriate testing included diabetes mellitus (OR, 0.4; 95% CI, 0.2-0.8; p = 0.01) and chest pain (OR, 0.07; 95% CI, 0.02-0.20; p < 0.001).
Even in an environment where self-referral profit motives and liability concerns are limited, inappropriate MPI is not uncommon and occurs at similar rates to other settings.
This work builds on several prior studies, which were carried out with similar study designs in non-VA hospital settings. Rates of inappropriate testing in these prior studies ranged from 7-24%, comparable to the rates seen in this study. Taken together, these studies suggest that although inappropriate testing is at most moderately prevalent, it remains a significant area for improvement. The observation that rates of inappropriate testing in the VA environment were similar to prior non-VA studies raises concerns that self-referral profit motives and liability concerns may not be the primary drivers of inappropriate testing. However, a few caveats should be noted. First, most prior studies were performed in academic rather than community settings. It is possible that profit and liability concerns may have a lesser influence on referring provider behavior than in the community. Second, many providers in VA environments also practice in non-VA environments and may not compartmentalize their practice patterns based on differences in potential liability. Nonetheless, these results do raise doubt about the potential for policy and regulatory changes to reduce inappropriate testing. Further research into education and training interventions aimed at reducing inappropriate testing may be more fruitful.
Another concern in this particular study is the extremely high rate of PET stress testing. Although PET testing has many advantages over SPECT, exercise stress testing with PET remains extremely challenging and is limited to highly specialized centers. The additional diagnostic and prognostic value of exercise parameters is unavailable with pharmacologic PET testing. The authors acknowledge that seven subjects could have been evaluated with exercise testing without imaging. However, they do not indicate how many subjects could have been evaluated with exercise SPECT testing rather than pharmacologic PET.
Keywords: Uncertainty, Myocardial Perfusion Imaging, Odds Ratio, Tomography, Emission-Computed, Single-Photon, Referral and Consultation, Exercise, Fatty Acids, Coronary Disease, Health Facilities, Proprietary, Positron-Emission Tomography, Prevalence, Veterans, Chest Pain, Confidence Intervals, Logistic Models, United States, Diabetes Mellitus, Exercise Test
< Back to Listings