Abstract 3996: Does Treatment of Acute Coronary Syndrome Patients in Invasive Hospitals Improve Survival Compared to Treatment in Community Hospitals? Propensity score analysis of the Minnesota Heart Survey 2001–2002
Background: In acute coronary syndrome (ACS) patients, an invasive strategy with percutaneous coronary intervention (PCI) is associated with improved survival and reduced cardiac morbidity. Observational studies suggest patients remaining in community hospitals (CH), where PCI is not available, have worse outcomes than those transferred to PCI hospitals. Aim: To determine whether the outcomes of ACS patients presenting to CH would have been different if they would have presented or been transferred to a PCI-hospital.
Methods: ACS Cases were identified from the Minnesota Heart Survey (MHS), an ongoing longitudinal cross-sectional survey of patients treated with ACS during 2001– 02. Transfers were excluded. A propensity score (PS), reflecting the probability of presenting to a CH, was estimated from 86 abstracted data elements (demographics, medical history, comorbidities, presentation, exam findings, and laboratory data). Patients presenting and treated at CH were compared to PS-matched subjects presenting to PCI-hospital. The primary endpoint was mortality, measured at 6 months and 1 year.
Results: We identified 353 CH and 2036 PS-matched PCI-hospital patients. Baseline variables were statistically different in 52/86 variables included in the PS model. The PS analysis achieved a significant 87% reduction in bias after matching with an excellent discriminating value (c-statistic=0.83). The crude and PS-estimated mortality of both groups is presented in the table⇓. Coronary angiography and revascularization were used in only 61% and 40% respectively of the PCI-hospital matched cohort.
Conclusions: ACS patients treated at CH had twice the mortality of those presenting to PCI hospitals. Their worse outcome was explained by their increased comorbidity. The low rate of revascularization may account for the absence of an expected survival benefit in the matched cohort presenting to the PCI hospitals. Future research is warranted to address these questions.