Appropriateness of Percutaneous Coronary Interventions in Patients With Stable Coronary Artery Disease in US Department of Veterans Affairs Hospitals From 2013 to 2015

IMPORTANCE: In hospitals outside of the US Department of Veterans Affairs (VA) system, 1 in 10 percutaneous coronary interventions (PCIs) for stable coronary artery disease is considered rarely appropriate by the appropriate use criteria, with variation across hospitals. The appropriateness of PCIs in VA hospitals has not been documented. OBJECTIVE: To characterize the appropriateness of PCIs in VA hospitals. DESIGN, SETTING, AND PARTICIPANTS: This retrospective cohort study included patients with stable coronary artery disease undergoing elective PCI from November 1, 2013, to October 31, 2015, within the VA Clinical Assessment, Reporting, and Tracking Program, an operational program that includes 59 VA hospitals. Data were analyzed from March 1, 2019, to August 8, 2019. EXPOSURES: Elective PCI at a VA hospital. MAIN OUTCOMES AND MEASURES: Proportion of PCIs classified as appropriate, may be appropriate, or rarely appropriate; extent of hospital-level variation in rarely appropriate PCIs using criteria issued by cardiovascular professional societies in 2012. The extent of hospital-level variation in rates of rarely appropriate PCI was characterized using hospital proportions and random-effect logistic regression. RESULTS: Among 2611 patients undergoing elective PCI (mean [SD] age, 66.3 [7.6] years; 2577 [98.7%] men) at 59 hospitals, a total of 778 PCIs (29.8%) were classified as appropriate, 1561 PCIs (59.8%) were classified as may be appropriate, and 272 PCIs (10.4%) were classified as rarely appropriate. Rarely appropriate PCIs were more commonly performed in patients who had low-risk stress test findings (220 patients [89.1%]), who were taking no (100 patients [36.8%]) or 1 (167 patients [61.4%]) antianginal medication, or who had 1 coronary artery stenosis (185 patients [68.0%]). The unadjusted hospital-level rates of rarely appropriate PCIs ranged from 0% to 28.6%, with a median (interquartile range) of 9.7% (6.3%-13.9%). Random-effect models yielded an estimated median (interquartile range) rate of rarely appropriate PCI of 10.4% (8.7%-12.3%). CONCLUSIONS AND RELEVANCE: These findings suggest that in VA practice, most PCIs for stable coronary artery disease were classified as appropriate or may be appropriate. However, 1 in 10 PCIs was classified as rarely appropriate, with variation across VA hospitals. Efforts to improve patient selection are needed.