Sepsis-Associated Mortality, Resource Use, and Healthcare Costs: A Propensity-Matched Cohort Study*

Objectives: To examine long-term mortality, resource utilization, and healthcare costs in sepsis patients compared to hospitalized nonsepsis controls. Design: Propensity-matched population-based cohort study using administrative data. Setting: Ontario, Canada. PATIENTS: We identified a cohort of adults (≥ 18) admitted to hospitals in Ontario between April 1, 2012, and March 31, 2016, with follow-up to March 31, 2017. Sepsis patients were flagged using a validated International Classification of Diseases , 10th Revision-coded algorithm (Sepsis-2 definition), including cases with organ dysfunction (severe sepsis) and without (nonsevere). Remaining hospitalized patients were potential controls. Cases and controls were matched 1:1 on propensity score, age, sex, admission type, and admission date. INTERVENTIONS: None. Measurements and Main Results: Differences in mortality, rehospitalization, hospital length of stay, and healthcare costs were estimated, adjusting for remaining confounders using Cox regression and generalized estimating equations. Of 270,669 sepsis cases, 196,922 (73%) were successfully matched: 64,204 had severe and 132,718 nonsevere sepsis (infection without organ dysfunction). Over follow-up (median 2.0 yr), severe sepsis patients had higher mortality rates than controls (hazard ratio, 1.66; 95% CI, 1.63–1.68). Both severe and nonsevere sepsis patients had higher rehospitalization rates than controls (hazard ratio, 1.53; 95% CI, 1.50–1.55 and hazard ratio, 1.41; 95% CI, 1.40–1.43, respectively). Incremental costs (Canadian dollar 2018) in sepsis cases versus controls at 1-year were: $29,238 (95% CI, $28,568–$29,913) for severe and $9,475 (95% CI, $9,150–$9,727) for nonsevere sepsis. Conclusions: Severe sepsis was associated with substantially higher long-term risk of death, rehospitalization, and healthcare costs, highlighting the need for effective postdischarge care for sepsis survivors.