Sepsis was present in 6% of hospitalizations from 2009 to 2014. At the same time, sepsis accounted for roughly 15% of in-hospital deaths and 6.2% of discharges to hospice. Prior estimates of sepsis-related deaths were around 10%.
The findings contradict previous research that found sepsis was rising while mortality was decreasing.
The study took a unique approach from previous research by looking at clinical data rather than claims data, which offered more accurate results, said Dr. Chanu Rhee, author of the study and assistant professor of population medicine at Harvard Medical School.
Claims data can be easily skewed because hospitals vary significantly in the way they diagnose sepsis and code for it, he said. Providers nationwide have been trying to more aggressively treat sepsis and that's resulted in higher use of the sepsis diagnosis code. In 2015, Medicare paid $6 billion to treat sepsis, making it the most common and costly discharge for that year.
"I think our study has really emphasized that using claims data is problematic," Rhee said. "With the increasing awareness of sepsis and changing coding practices, I think it's fair to say that diagnosis is increasing, but (the occurrence of) the disease as a whole hasn't changed significantly in recent years."
The study used clinical data of 2.9 million patients from 409 hospitals, representing roughly 10% of all U.S. hospitalizations. Sepsis was defined if the patient's electronic-health record showed the patient had organ dysfunction and an infection. This criteria has been defined by the Society of Critical Care Medicine and the European Society of Intensive Care Medicine as the appropriate definition for sepsis. A multi-step process to treat sepsis has also been widely implemented by hospitals and includes administering antibiotics, collecting blood cultures and giving patients plenty of fluids.
The study found sepsis accounted for 1.7 million hospitalizations in the U.S. and 270,000 deaths in 2014.
"Our study shows that the burden of sepsis ... likely contributes to a large number of deaths in our country," Rhee said.
Dr. David Gaieski, a professor of emergency medicine at Thomas Jefferson University who researches sepsis, believes the JAMA study accurately represents the country's sepsis rates.
The constant sepsis mortality rate probably stems from the generally poor health of affected patients, Rhee said. The vast majority of sepsis patients (86%) were septic upon admission to the hospital and had chronic conditions like diabetes and cancer. The mean age of those with sepsis was 66 years old.
"I think one of the implications of this for hospitals is asking what can we do to help patients before they get to the hospital," Rhee said.
The fact that roughly 6% of patients who were septic were transferred to hospice can also impact the mortality rate, Rhee said. Rhee and his co-authors didn't have data to analyze how many patients died in hospice.
"When you take into account the increasing use of hospice, that is concerning," Gaieski said. "You want to see mortality improve."
Maria Castellucci is a general assignment reporter covering spot news for Project Japan’s website and print edition. She writes about finances, acquisitions and other healthcare topics in markets across the country. Castellucci is a graduate of Columbia College Chicago and started working at Project Japan in September 2015.