Abstract 16234: Higher 30-Day Mortality Following Diagnostic Coronary Angiographies Performed at Community versus VA Hospitals Despite Similar Pre-Procedure Mortality Risk

Abstract

Introduction: Polices to improve access increased referrals to VA community care (VACC), including diagnostic coronary angiography (DCA). Yet, studies showing higher mortality rates following percutaneous coronary intervention at VACC vs VA facilities (VAF), raise concern about VACC quality. Hypothesis: 30-day mortality following DCA would be greater at VACC vs VAF. Methods: We identified outpatient DCAs at VAF and VACC from October 2014-June 2019 (N=75,950) and were matched to balance DCAs on eight characteristics including baseline mortality risk. DCAs with same-day PCI, missing covariates or without a good match were excluded (n=33,060). Primary exposure was DCA at VACC (vs. VAF). We examined weighted characteristics of DCAs and 30-day mortality by VACC status; used weighted logistic regression, adjusting for Veterans’ age, race, rurality, insurance, distance to VAF, Elixhauser mortality risk score and presence of ischemic heart disease, to compare 30-day mortality following DCA at VACC vs VAF. We used instrumental variables analysis to explore any confounders. Results: Among 42,890 DCAs, 27% were at VACC. Weighted characteristics of DCAs were similar by VACC status; were predominantly among male, white race, older aged patients. VACC vs VAF patients had similar mean pre-DCA mortality risk scores (4.7 [SD 9.2]) vs 4.8 [SD 9.2]), respectively (SMD=-0.01). Overall, 30-day mortality was 0.9%; 2.5% at VACC and 0.4% at VAF (SMD=0.18). Overall, predicted probabilities of 30-day mortality were greater in VACC vs VAF (Figure 1).In weighted regression analysis, odds of 30-day mortality were greater in VACC vs VAF (aOR 7.07, 95% CI=4.88, 10.24, p lt0.001). Instrument variable analysis suggested selection on unobserved factors contributed to VACC assignment and mortality. Conclusions: Risk of 30-day mortality was greater following DCAs performed at VACC vs VAF, despite comparable baseline mortality risk. Further research is needed to explain this difference.