Comparison of 30-Day Mortality After Diagnostic Coronary Angiography at VA and Community Hospitals
摘要
The US Department of Veterans Affairs (VA) delivers care in its own facilities (VAF) and purchases community care (VACC) for some services including diagnostic coronary angiography (DCA). Whether patients experience comparable outcomes for DCA through VAF and VACC is unknown.
ObjectiveTo compare 30-day all-cause and cardiac-related mortality associated with DCA in VAF and VACC.
DesignRetrospective cohort study with pooled cross-sectional analysis.
ParticipantsVA patients who lived in the USA and underwent outpatient DCA in VAF or VACC between October 2015 and June 2019 (N = 31,640) were matched on nine baseline characteristics: age, race, rurality, drive distance to nearest VA primary care site, history of ischemic heart disease, history of valvular disease, procedure fiscal year, insurance status, and Elixhauser comorbidity index. Analysis was conducted from 2023 to 2025.
Main MeasuresThirty-day all-cause and cardiac-related mortality. Average treatment outcomes of VACC were estimated using weighted regression adjustment to account for selection into VACC. Sub-analyses included regression models stratified by diagnosis of valvular disease and heart failure.
Key ResultsAmong 31,640 DCAs, 28.06% were performed through VACC. VACC and VAF DCAs were well-balanced. Overall, 30-day all-cause mortality was 0.96% (2.49% in VACC and 0.37% in VAF [SMD = 0.182]), and most were cardiac-related. Compared with those who underwent DCA at VAF, those who had DCA at VACC had greater odds of 30-day all-cause (aOR 6.918, 95% CI = 4.521, 10.586), cardiac-related (aOR 6.112, 95% CI 3.715, 10.055), and non-cardiac related (aOR 13.45, 95% CI 5.352, 33.801) mortality. In stratified models, VACC DCA was associated with greater odds of 30-day mortality for all subgroups.
ConclusionsResults highlight the need for further research regarding strategies to improve follow-up and outcomes for VA patients undergoing DCA, particularly in VACC settings.