Variation in treatment strategy for non-ST segment elevation myocardial infarction: A multilevel methodological approach.

Variations by hospital and region in the selection of an early invasive strategy (EIS) after non–ST-segment elevation myocardial infarction (NSTEMI) in patients with high-risk criteria are unknown. We evaluated the data of 7037 patients with NSTEMI from 20 hospitals of 3 regions from the Korean Acute Myocardial Infarction Registry–National Institute of Health database. We used hierarchical generalized linear mixed-models to estimate region- and hospital-level variation in the selection of an EIS after adjusting for patient-level high-risk criteria. We explored the variation using the median rate ratio (MRR), which estimates the relative difference in the risk ratios of two hypothetically identical patients at two different sites. An EIS was selected in 84.4% of patients. At the hospital level, the median selection rate was 80.4%. At the region level, the median selection rate was 74.9% in the east region, 81.3% in the north region, and 83.9% in the west region, respectively. After adjusting for patient-level covariates, we found significant hospital- (MRR 2.19, 95% confidence interval [CI]: 1.74–3.03) and region-level (MRR 1.88, 95%CI: 1.26–5.44) variation in the selection of an EIS. Among patient-level factors, male sex, ongoing chest pain, history of coronary artery disease or acute heart failure, and GRACE risk score > 140 were independently associated with the selection of an EIS. We observed significant hospital- and region-level variation in the selection of an EIS after NSTEMI in high-risk patients. Quality improvement efforts are required to standardize decision making and to improve clinical outcomes. • There is significant hospital- and region-level variation in early invasive strategy selection. • An early conservative strategy is often chosen when an early invasive strategy would provide more benefit. • Site-level variation reflects that there is significant undertreatment for high-risk patients that could worsen clinical outcome. • Quality improvement efforts are required to standardize decision making and improve clinical outcomes. [ABSTRACT FROM AUTHOR]