Study challenges use of in-hospital mortality as sepsis quality metric for safety-net hospitals

Sepsis is a leading cause of death and disability and a key target of state and federal quality measures for hospitals. In-hospital mortality of patients with sepsis is frequently measured for benchmarking, both by researchers and policymakers. For example, in New York, sepsis regulations mandate reporting of risk-adjusted in-hospital mortality, and hospitals with lower or higher than expected in-hospital mortality rates are publicly identified as high or low performers. Safety net hospitals (which face unique challenges, including fewer resources to care for a disproportionately high share of underinsured and high-risk patients) have been reported in prior studies to have higher risk-adjusted in-hospital mortality among patients with sepsis than non-safety-net hospitals.

However, among critically ill patients, in-hospital mortality may not adequately reflect short-term mortality fairly across all hospitals. In-hospital mortality is influenced by other variables, including hospital transfer practices, which shift the attribution of short-term death from the hospital to other sites. Safety-net hospitals may have different access or ability to transfer patients compared to non-safety-net hospitals.

In a new study, researchers at Boston University Chobanian & Avedisian School of Medicine found that safety net hospitals do have higher in-hospital mortality than non-safety net hospitals, but their 30-day mortality – an unbiased measure of short-term mortality – is actually not different.

"While the differences in mortality rates are numerically small, the difference in outcome metrics is enough to significantly affect hospital rankings. Current or future state and federal quality measures that use in-hospital mortality as a quality metric may unfairly penalize safety-net hospitals," explained corresponding author Anica Law, MD, MS, assistant professor of medicine at the school.

The researchers performed a retrospective analysis of patients with sepsis who were aged 66 and older and were admitted to an intensive care unit between January 1, 2011, through December 31, 2019, at both safety and non-safety net hospitals.

"Although in-hospital mortality is often selected as an outcome measure because of its availability in claims databases and hospital medical records without need for post-hospitalization follow-up, the 30-day measure is important to accurately understand true short-term mortality rates," said Law, who also is a pulmonologist and critical care physician at Boston Medical Center.

According to the researchers, the difference appears to be due partly to the fact that non-safety net hospitals discharge more patients to hospice, which shifts attribution of short-term mortality away from the index hospitalization. "When post-hospitalization data is incorporated, as is done in 30-day mortality analysis, you get a more accurate picture of who is actually dying in the short-term, and see that there no longer a difference between safety and non-safety net hospitals." The Centers for Medicare and Medicaid Services (CMS) has already legislated a process measure that assesses whether hospitals are performing key sepsis care steps efficiently, including initiating antibiotics. CMS is currently considering rolling out a national sepsis outcome measure; it is not yet known if CMS will be measuring in-hospital mortality or 30-day mortality.

These findings appear online in JAMA Network Open.

Funding for this study was provided by the National Institute of Health (NIH)/National Heart, Lung, and Blood Institute (NHLBI) K23HL153482, National Center for Advancing Translational Sciences (NCATS) 1KL2TR001411, the Doris Duke Charitable Foundation, the Boston University Evans Junior Faculty Merit Award, and the Boston University Chobanian & Avedisian School of Medicine Department of Medicine Career Investment Award.