Mortality among Veterans with Severe Sepsis Declines, but Significant Variation across VA Hospitals Persists
Severe sepsis is a leading cause of hospitalizations throughout the developed world, accounting for more than one million hospitalizations each year in the U.S. alone. Despite recent improvements, mortality remains unacceptably high. One in five patients with severe sepsis – and one in two of those with septic shock, die during hospitalization. This large, retrospective cohort study sought to determine the extent to which variation in short-term mortality following severe sepsis is explained by the VA hospital and regional VA healthcare network where a Veteran receives care. Investigators identified Veterans' first hospitalization involving severe sepsis at any of 114 VA hospitals in 2008 (n=33,095) and in 2012 (n=43,733), and then compared 30-day mortality after adjusting for severity of illness, hospital and region, as well as the number of severe sepsis hospitalizations at each location.
- Unadjusted 30-day mortality among hospitalized Veterans with severe sepsis declined from 18% in 2008 to 15% in 2012, despite very similar severity of illness between years.
- After severity-of-illness and case-mix adjustment, variation persisted in 30-day mortality across hospitals – and to a lesser extent, across regions. For example, the median hospital in the worst quintile of performers had a risk-adjusted 30-day mortality rate of 17% in 2012 compared with the best quintile with a rate of 13%, suggesting a 20% greater risk of death (4% absolute mortality difference) when treated at a hospital in the bottom versus the top quintile.
- Veterans in this study were predominately white (73%) and male (97%), with a mean age of 70.
- To adjust for case-mix and severity of illness, investigators used the VA ICU severity score, which has not been formally validated in patients treated outside the intensive care unit.
- It is possible that unmeasured case-mix factors may still influence sepsis-related mortality.
- While investigators show that hospitals and regions impact short-term sepsis-related mortality, they did not test which institutional or organizational characteristics explain this variation.
CCM also published an Editorial on this article. The Editorial (Higgins TL) notes that "The Veterans Affairs (VA) health system, despite a skewed gender distribution, offers a unique opportunity to study hospital and regional variation in a system well represented for size, teaching status and geography, having system-wide policies, and a shared electronic database." In addition, Dr. Higgins states that "Despite media headlines of steady deterioration in American healthcare, this study confirms that critical care outcomes are indeed improving."
This study was partly funded by HSR&D. Mr. Kepreos and Drs. Wiitala and Iwashyna are part of HSR&D's Center for Clinical Management Research in Ann Arbor, MI.
Prescott H, Kepreos K, Wiitala W, and Iwashyna T. Temporal Changes in the Influence of Hospitals and Regional Healthcare Networks on Severe Sepsis Mortality. Critical Care Medicine. July 2015;43(7):1368-1374.