Decreased racial disparities in sepsis mortality after an order set-driven initiative: An analysis of 8151 patients

Abstract

Background: Sepsis is a leading cause of hospital mortality and there is evidence that outcomes vary by patient demographics including race and gender. Our objectives were to determine whether the introduction of a standardized sepsis order set was associated with (1) changes in overall mortality or early antibiotic administration or (2) changes in outcome disparities based on race or gender. Methods: Patients seen in the emergency department and admitted to the hospital with a diagnosis code of sepsis were identified and divided into a preintervention co- hort seen during the 18 months prior to the initiation of a new sepsis order set and an intervention cohort seen during the 18 months after a quality initiative driven by in- troducing the order set. Associations between time period, race, gender, and mortality were assessed using univariate and multivariate logistic regression models. Other outcomes included early antibiotic administration (<3 h from arrival). Results: Overall mortality was unchanged during the intervention period (7.8% vs. 7.2%) in both univariate (relative risk [RR] 1.08, 95% confidence interval [CI] 0.93–1.26) and multivariate logistic regression (RR 1.11, 95% CI 0.93–1.28) models. Although male gender tended to have higher mortality, there was no statistically significant association between gender and mortality in either cohort. In the multivariable model, Black race was associated with increased risk of death in the preintervention period (RR 1.41, 95% CI 1.02–1.94), but this association was not present in the intervention period. Patients of color also saw significantly more improvement in early antibiotic administration during the intervention period than White patients. Conclusions: An order set–driven sepsis initiative was not associated with overall improved mortality but was associated with decreased racial disparities in sepsis mortality and early antibiotics.