Claims-Based Estimates Not Accurate for Sepsis Analysis

binder that says claims
binder that says claims
The study includes a large number of diverse hospitals which account for approximately 10% of all acute care hospitalizations in the United States in 2014.

Claims-based estimates of sepsis would have us believe that sepsis and related mortality and are on the decline. However, a recent retrospective cohort study published in the Journal of the American Medical Association provides a more objective estimate.

The study, conducted by researchers at Harvard Medical School, Brigham and Women’s Hospital, the Centers for Disease Control and Prevention, and other institutions around the country used detailed clinical data from 7,801,624 adults admitted to 409 United States hospitals between 2009 and 2014. Sepsis incidence from 2009 through 2014 was 6% using clinical criteria and 10.3% per claims data. The clinical data showed that although in-hospital mortality decreased during that period (3.3% per year), there was no significant change in incidence of sepsis, death, or discharge to hospice. 

According to claims, the incidence of sepsis increased 10.3% per year, mortality declined 7% per year, and death or discharge to hospice declined 4.5% per year. The clinical criteria were 69.7% sensitive in identifying sepsis (95% CI, 52.9%-92%) vs 32.3% for claims data (95% CI, 24.4%-43%; P =<.001). The comparable positive predictive value is 70.4% (95% CI, 64%-76.8%) vs 75.2% (95% CI, 69.8%-80.6%), respectively (P =.23).

These findings not only indicate that sepsis and related mortality or discharge to hospice seem to be remaining stable, but they also suggest that clinical data provide a more reliable estimate of sepsis in US hospitals than does claims-based data.

Related Articles


Rhee C, Dantes R, Epstein L, et al. Incidence of trends of sepsis in US hospitals using clinical vs claims data, 2009-2014. JAMA. 2017;318:1241-1249.