

Using data to drive sepsis care
A researcher discusses successful statewide efforts to improve sepsis treatment–and setting the bar for change at the national level

Sepsis–the body’s extreme, life-threatening response to infection–is a leading cause of hospitalization.
Around 1.7 million people are hospitalized with sepsis each year.
Of those, 350,000 will die of the condition, making it the most common cause death in the hospital.
When it comes to treating sepsis, time is of the essence.
With rapid recognition and deployment of the proper antibiotics, fluids and other supportive measures, patients are more likely to recover.
In the effort to drive care forward, Prescott leads one of the most successful sepsis registries in the country.
Funded by Blue Cross Blue Shield of Michigan, the collaborative quality initiative for sepsis brings together 69 hospitals from across the state participating in the Michigan Hospital Medicine Safety Consortium.
Tracking of sepsis outcomes is made more complicated by the fact that the definition of sepsis continues to evolve.
According to Hallie Prescott, M.D., an internationally renowned sepsis researcher with the Department of Pulmonary and Critical Care Medicine at U-M Health and the VA Medical Center, sepsis shouldn't be considered just one disease, but rather a family of diseases, much like cancer.
But developing a full understanding of the subtypes of sepsis and related critical illnesses, including how to quickly recognize and treat them, will require a lot of good data, she says.
The federal government recently launched an NIH-funded national study called the APS Consortium (ARDS, pneumonia, and sepsis)–a coordinated effort to collect health-related data and biospecimens from 4,000 adults hospitalized with critical illness across the country, from six primary centers including the University of Michigan.
“It's very challenging to monitor sepsis epidemiology and outcomes because who gets labeled as having sepsis is highly variable at both the physician and hospital level,” said Prescott.
“The more clinical data we get, the more insight we have into what improves which subset of sepsis.”
“In Michigan, we are able to identify high performing hospitals who have made big improvements in sepsis care and learn from them,” said Prescott.
“We’ve also made site visits to hospitals with challenges to walk through their care processes and provide feedback.”
The model was recently highlighted by the United States Agency for Healthcare Research and Quality as an example of a successful sepsis improvement model and the data collected informed a set of guidelines called the Hospital Sepsis Program Core Elements published with the Centers for Disease Control and Prevention.
“Our program is very much improving care in the state, but then it's also providing a lens into real world care in a way that is informing national sepsis policy activities,” said Prescott.
A key strength of the Michigan sepsis registry is the rigorous approach to identifying patients with sepsis.
Rather than simply relying on diagnostic codes, professional abstractors at each hospital review charts for patients with infection to identify those meeting clinical criteria for sepsis.
“If you just pull in hospitalizations that are called sepsis, you are comparing apples and oranges.”
Prescott and co-author Nuala Meyer, M.D., recently penned a review in the New England Journal of Medicine about sepsis, summarizing the body of literature and evolving definitions of the condition and best practices for addressing it.
Additionally, Prescott and her colleagues recently published a mortality risk-adjustment model for the state of Michigan using patients’ data from the registry, including chronic health conditions, age, recent hospitalizations, and physiologic data from just the first six hours of a hospital visit.
“We really want to tell hospitals ‘what do your patients look like when they arrive to you? And given what they looked like, how well did you perform with regards to sepsis care?’”
This type of benchmarking–using risk-adjusted outcome measures–is just beginning at the national level, putting Michigan ahead in its efforts, said Prescott.
Paper cited: "Sepsis and Septic Shock", New England Journal of Medicine. DOI: 10.1056/NEJMra2403213
Funding/disclosures: Support for HMS is provided by Blue Cross and Blue Shield of Michigan and Blue Care Network as part of the BCBSM Value Partnerships program.