Dependence of All-Cause Standardized In-Hospital Mortality on Sepsis Mortality Between 2005 and 2010

Sepsis is the 11th leading cause of death in the United States. The authors evaluated the dependence of all-cause in-hospital mortality on sepsis mortality. A retrospective observational cohort design and All Patient Refined Diagnosis Related Groups were used to evaluate 150 410 patients (>17 years of age) over 6 years. The hospital standardized mortality ratio (HSMR) decreased from 0.80 ± 0.04 to 0.65 ± 0.04 (observed/expected [O/E]). The HSMR was correlated with the standardized sepsis mortality ratio (SSMR), accounting for 61% of the variation (P < .0001). The mortality rate for sepsis declined from 14.54 ± 1.60% to 8.57 ± 1.88% (P < .002), and the SSMR decreased from 0.82 ± 0.09 to 0.48 ± 0.07 (O/E; P < .002). The months of sepsis mortality matched the months of hospital mortality for the lowest of each (odds ratio = 0.30, 95% confidence interval = 0.14-0.65; P < .003) and the highest of each (odds ratio = 1.47, 95% confidence interval = 1.18-1.84; P < .0007). Monitoring monthly mortality of sepsis can provide insight for timely intervention into variation of total in-hospital mortality.