Closing a Critical Measurement Gap in Pediatric Sepsis
Sepsis, a life-threatening response to infection that can lead to organ failure and death, is one of the most serious conditions affecting hospitalized children. Yet physicians and policymakers have long struggled to answer a crucial question: how common is pediatric sepsis, and how can we reliably measure it?
According to a new multi‑center study led by Harvard Pilgrim Health Care Institute’s Chanu Rhee and published in JAMA, the burden is substantial. By developing and applying a new surveillance method based on objective clinical data rather than variable billing codes, the researchers provide a more reliable picture of its burden and lay the groundwork for better measurement, benchmarking, and ultimately improving outcomes.
Uncovering a measurement gap: unclear definitions, strict clinical directives
Dr. Rhee’s clinical work has long focused on caring for patients with life-threatening infections, particularly when they progress to sepsis. He trained in internal medicine, followed by dual fellowships in infectious disease and critical care medicine.
His interest in studying sepsis took root during his critical care fellowship. Doctors were being given directives to identify septic patients quickly and treat them aggressively, often including early ICU admission. However, in practice, they faced ambiguity in determining which patients had sepsis and how severe each case was. This gray area, according to Dr. Rhee, was not an ideal building ground for policy and protocol. As he became more involved in infection control and hospital epidemiology during his infectious disease fellowship, he saw how healthcare-associated infections were tracked using rigorous, standardized surveillance methods. Sepsis, however—despite its severity—lacked a comparable approach.
“A central goal of our work is to improve how we measure sepsis,” says Dr. Rhee. “Without reliable measurement, it’s difficult to understand the true scope of the problem or to know where to focus efforts to improve outcomes.”
With the guidance of his mentors, including Michael Klompas (co-senior author of the study), and several collaborators, Dr. Rhee began by demonstrating that administrative claims data are an unreliable way to measure sepsis, as diagnosis and coding practices vary widely across hospitals and change over time.
A new surveillance framework based on objective clinical data
Building on this work, he and Dr. Klompas led a CDC-supported study in 2017 that established the first national estimates of sepsis burden in adults using objective clinical data from electronic health records rather than billing data. This study also directly informed CDC’s Adult Sepsis Event (ASE) surveillance framework, which has since been widely used to study sepsis epidemiology, treatment patterns, and the impact of quality improvement initiatives, and now serves as the foundation for a forthcoming national sepsis outcome measure.
From adults to children: extending EHR-based sepsis surveillance
This work sparked strong interest from the CDC and other partners in developing a similar, data-driven approach for pediatric sepsis. Although trained as an adult physician, Dr. Rhee notes that the impact of sepsis in children has resonated across the whole field, with several tragic, high-profile cases spurring public awareness campaigns and new policies. “Despite these efforts, we’ve still historically lacked reliable, standardized ways to define and track pediatric sepsis at a national level,” he says.
Providing concrete, objective data on the true burden of pediatric sepsis was a next step in our research.
These discussions led to a 2019 white paper in Pediatrics by Drs. Rhee, Klompas and colleagues, which outlined a roadmap for developing a pediatric sepsis surveillance framework, including key challenges, a preliminary definition, and next steps for refining and validating the approach.
When the pandemic hit, the team—and CDC—shifted focus to studying and responding to COVID-19. They returned to their pediatric sepsis research efforts in 2023, reuniting with many of the same collaborators, including leading pediatric sepsis experts. In 2024, the publication of the Phoenix criteria, a new consensus definition for pediatric sepsis, helped address one of their biggest challenges: the lack of a clear reference standard.
The Phoenix definition provided a strong, consensus-based clinical structure to anchor the team’s surveillance definition. Building on this, the researchers adapted and simplified the Phoenix criteria into an EHR-based surveillance framework, applying the same principles as the Adult Sepsis Event (ASE) approach to enable consistent and scalable retrospective identification of cases across adult and pediatric populations. The team also partnered directly with researchers leading the Phoenix effort, gaining not only additional expertise but the ability to evaluate their proposed approach against the same types of data used to develop the Phoenix criteria.
“A core principle in medicine is that kids are not just small adults,” Dr. Rhee says. “There are key differences in physiology, even between infants versus adolescents. Children are also evaluated and treated differently than adults. We had to carefully adapt our surveillance framework to account for age-specific vital signs, patterns of organ dysfunction, and differences in clinical practice.”
Large burdens revealed, and a necessary wake-up call
Armed with this framework, the team developed the new Pediatric Sepsis Event (PSE) definition and applied it to 3.9 million pediatric hospitalizations from 2016 through 2023 across hundreds of hospitals and health systems.
The results showed that sepsis is present in approximately 1 in every 75 pediatric hospitalizations, and that more than 1 in 10 affected children do not survive their hospital stay. Furthermore, nearly 1 in 5 deaths among all hospitalized children involved sepsis. Of additional concern, progress in preventing sepsis and improving outcomes has been stagnant: the nationwide rates of pediatric sepsis and associated deaths did not change from 2016 to 2022.
“By replacing billing-code–based estimates of sepsis cases and deaths with objective clinical data, we were able to create a much more reliable picture of pediatric sepsis in the US,” adds Dr. Klompas. “The stability of sepsis incidence and mortality over time underscores the urgency to do more to prevent sepsis and improve sepsis outcomes.”
Important distinctions, ongoing conversations
“Our findings show that sepsis remains a major contributor to death among hospitalized children,” says Dr. Rhee. “Better measurement is critical because it allows health systems to understand where the burden lies and to target interventions more effectively.”
Dr. Rhee also emphasizes what the PSE is not: a clinical definition. Rather, it’s meant to be the necessary basis to conduct accurate, standardized measurements that can be scaled within hospitals, across health systems, and even at a national level.
As with any new approach introduced into the health care field, he expects—and welcomes—pushback.
“One practical concern is that, even though we designed the Pediatric Sepsis Event framework to use routinely collected electronic health record data, and provided a toolkit to support its implementation, it still requires a meaningful investment in data infrastructure and IT expertise,” he says. “Some health systems may ask why simpler approaches, such as billing codes, aren’t good enough. Our view is that they’re not—and our findings support that.”
Looking ahead
The team is still actively discussing next steps but is optimistic about the opportunities the new framework and large, multi-center datasets afford. For example, they can now look more closely at how sepsis incidence, severity, and outcomes vary across different patient populations and hospital settings. They can also assess differences in treatment patterns and outcomes across hospitals and health systems to identify practices that may be associated with better outcomes.
We have both the responsibility and the opportunity to drive meaningful change in the fight against sepsis.
According to Dr. Rhee, the study has implications beyond hospital walls. The takeaway for parents and caregivers is awareness — understanding that sepsis can be serious and to act quickly when a child may have sepsis. For policymakers, it’s the importance of supporting the infrastructure and systems needed to measure and improve sepsis care at scale.
The work is also just one piece of the larger research mission Drs. Rhee and Klompas have set out to accomplish. The pair launched the Center for Sepsis Epidemiology and Prevention Studies (SEPSIS) in early 2025, partnering with Institute investigators, staff, and external collaborators. The center focuses on clinical and epidemiologic research using electronic health record data to improve disease surveillance, diagnosis, prevention, and treatment, bringing together multidisciplinary expertise to advance solutions in this area.
“Our past work has taught us the importance of establishing integrated research infrastructure to reduce sepsis incidence, inform more targeted and effective treatments, refine sepsis policies, and ultimately help drive better patient outcomes,” says Dr. Klompas. "We have both the responsibility and the opportunity to drive meaningful change in the fight against sepsis.”
