Patient and hospital characteristics associated with inpatient severe sepsis mortality in California, 2005-2010

Objectives: The primary objective of this study was to identify which patient demographic, patient health, and hospital characteristics were associated with in-hospital mortality. A secondary objective was to determine the relative influence of these characteristics on mortality. Design and Setting: Public-use data for 2005-2010 were used in this retrospective, cross-sectional analysis of discharges from nonfederal, general acute hospitals in California. A staged logistic regression approach was used to examine the relative influence of variables associated with in-hospital mortality. Patients: A total of 1,213,219 patient discharges for adults (aged >= 18 yrs) having International Classification of Diseases-9 diagnosis and procedure codes indicating severe sepsis. Intervention: None. Measurements and Main Results: Patient demographics (age, gender, race, ethnicity, and payer category), patient health status (acute transfer, Charlson-Deyo comorbidity index, and organ failures), and hospital characteristics (ownership type, teaching status, bed size, annual patient days, acute discharges, emergency department visits, inpatient surgeries, severe sepsis as a percentage of all discharges, and year) were obtained from the California Office of Statewide Health Planning and Development. Overall, in-hospital mortality was 17.8%. There was a steady annual increase in the number of sepsis discharges, but a decrease in mortality throughout the study period. Mortality increased with age and was associated with white race, and Medicaid (Medi-Cal) and private insurance. Patient health status additionally explained inpatient mortality. Hospital volume measures were statistically significant in regression analysis, whereas static structural measures were not. There were modest associations between measures of annual treatment volume and likelihood of inpatient mortality, notably decreasing likelihood with more acute discharges and with greater severe sepsis volume. Conclusions: Although patient demographics and health status are the most important predictors of in-hospital mortality of patients with severe sepsis, hospital characteristics do play a substantial role. Findings regarding hospital volume can be used to improve processes and improve patient outcomes. (Crit Care Med 2012; 40: 2960-2966)