Sepsis-Associated 30-Day Risk-Standardized Readmissions: Analysis of a Nationwide Medicare Sample*
To determine national readmission rates among sepsis survivors, variations in rates between hospitals, and determine whether measures of quality correlate with performance on sepsis readmissions.Design:
Cross-sectional study of sepsis readmissions between 2008 and 2011 in the Medicare fee-for-service database.Setting:
Acute care, Medicare participating hospitals from 2008 to 2011.Patients:
Septic patients as identified by International Classification of Diseases, Ninth Revision codes using the Angus method.Interventions:
None.Measurements and Main Results:
We generated hospital-level, risk-standardized, 30-day readmission rates among survivors of sepsis and compared rates across region, ownership, teaching status, sepsis volume, hospital size, and proportion of underserved patients. We examined the relationship between risk-standardized readmission rates and hospital-level composite measures of quality and mortality. From 633,407 hospitalizations among 3,315 hospitals from 2008 to 2011, median risk-standardized readmission rates was 28.7% (interquartile range, 26.1–31.9). There were differences in risk-standardized readmission rates by region (Northeast, 30.4%; South, 29.6%; Midwest, 28.8%; and West, 27.7%; p < 0.001), teaching versus nonteaching status (31.1% vs 29.0%; p < 0.001), and hospitals serving the highest proportion of underserved patients (30.6% vs 28.7%; p < 0.001). The best performing hospitals on a composite quality measure had highest risk-standardized readmission rates compared with the lowest (32.0% vs 27.5%; p < 0.001). Risk-standardized readmission rates was lower in the highest mortality hospitals compared with those in the lowest (28.7% vs 30.7%; p < 0.001).Conclusions:
One third of sepsis survivors were readmitted and wide variation exists between hospitals. Several demographic and structural factors are associated with this variation. Measures of higher quality in-hospital care were correlated with higher readmission rates. Several potential explanations are possible including poor risk standardization, more research is needed.