Hospital-level variation in outcomes after in-hospital cardiac arrest in Denmark.

Abstract

We investigated hospital-level variation in outcomes after in-hospital cardiac arrest (IHCA) in Denmark, and assessed whether variation in outcomes could be explained by differences in patient characteristics. Adult patients (≥18 years old) with IHCA in 2017 and 2018 were included from the Danish IHCA Registry (DANARREST). Data on patient characteristics and outcomes were obtained from population-based registries. Predicted probabilities, likelihood ratio tests, intraclass correlation coefficients (ICCs), and median odds ratios (ORs) were calculated for return of spontaneous circulation (ROSC), survival to 30 days, and survival to 1 year. A total of 3340 patients with IHCA from 24 hospitals were included. We found that hospital-level variation in outcomes after IHCA existed across all measures of variation. The unadjusted median OR for ROSC, survival to 30 days, and survival to 1 year were 1.28 (95% confidence interval [CI]: 1.24, 1.45), 1.38 (95% CI: 1.33, 1.60), and 1.44 (95% CI: 1.39, 1.70), respectively. The unadjusted ICC suggest that 2.0% (95%: 1.6%, 4.4%), 3.3% (95%: 2.7%, 6.8%), and 4.3% (95%: 3.5%, 8.6%) of the total individual variation in ROSC, survival to 30 days, and survival to 1 year was attributable to hospital-level variation. These results decreased but persisted in the analyses adjusted for select patient characteristics. In this study, we found that outcomes after IHCA varied across hospitals in Denmark. However, only about 2%-4% of the total individual variation in outcomes after IHCA was attributable to differences between hospitals, suggesting that most of the individual variation in outcomes was attributable to patient-level variation.