CHOP Researchers Co-leading CDC Project to Develop National Pediatric Sepsis Surveillance Tool

A surveillance algorithm created by a multidisciplinary Children's Hospital of Philadelphia team to study local pediatric sepsis trends will be validated and translated as a national surveillance tool, with new funding support from the Centers for Disease Control and Prevention (CDC).

Developing an objective, efficient, reliable surveillance tool to estimate the national incidence and outcomes of pediatric sepsis is an unmet need recognized by the CDC. Sepsis, a syndrome of life-threatening organ dysfunction resulting from a dysregulated response to infection, is the leading cause of non-accidental death and disability for children worldwide. It is challenging to measure the burden of pediatric sepsis because of complex and non-specific diagnostic criteria.

"I have worked with sepsis data for more than 10 years, and I know how difficult it is to identify patients with sepsis," said Marianne Chilutti, MS, director of Library Education and Data Services in the Department of Biomedical and Health Informatics at CHOP.

CHOP researchers from the Division of Emergency Medicine and Arcus are partnering with Harvard Pilgrim Health Care Inc. and others to develop and optimize a surveillance definition of pediatric sepsis, based on CHOP's previous work on pediatric sepsis surveillance. This work, supported by a CDC contract, will harness data from electronic health records (EHR) to define and track pediatric sepsis.

The study to identify the CHOP algorithm was the first Arcus-based research study, which launched in 2017. CHOP's Arcus repository links EHR data to clinical outcomes. Data from the CHOP Sepsis Registry and the then-new Arcus Archives led to the programming and validation of the sepsis algorithm. The algorithm was subsequently published in 2020 and recognized by professional societies and federal agencies.

"Translation is the outcome we strive for," said Jeffrey Pennington, MSCS, chief research informatics officer at CHOP Research Institute. "Linking research and clinical data together allowed this work to happen, which is the purpose of Arcus. The translation of CHOP's sepsis surveillance algorithm to a national surveillance tool is an important outcome for CHOP and Arcus."

Arcus projects are a team effort in which data scientists, information analysts, librarians, and more collaborate with biomedical researchers to conduct data-driven research that is highly innovative, computationally complex, and reproducible. Arcus projects link biological, clinical, research and environmental data about patients and research participants across CHOP while protecting patient privacy.

Surveillance Challenges

Over the last decade, researchers have started leveraging information from EHRs to develop methods to identify patients with sepsis in other ways than diagnostic codes. A team of Harvard University investigators — who are co-leading this CDC contract — developed an adult sepsis surveillance tool that now informs CDC's estimates of adult sepsis.

Frances Balamuth, MD, PhD, attending physician in the Division of Emergency Medicine and co-director of the Pediatric Sepsis Program, collaborated with Scott Weiss, MD, now division chief of critical care medicine at Nemours Children's Health, to develop the pediatric version of this tool that is being scaled nationally with the CDC contract.

"One of the challenges with sepsis is that it's complicated," said Dr. Balamuth, who is also an associate professor of Pediatrics at the University of Pennsylvania. "Diagnostic codes can easily identify patients with pneumonia or cellulitis, for example, but people have shown time and again that diagnostic codes are neither sensitive nor specific for identifying sepsis."

Tailoring Improvement Efforts

The first aim of the two-year project is to develop and validate an EHR-based pediatric sepsis event (PSE) surveillance definition by assembling a panel of experts to refine the CHOP PSE definition to ensure its usability across diverse settings. Once the definition is established in a testing cohort, the researchers will evaluate its performance in a validation cohort. These cohorts are established at three sites: CHOP, Children's Hospital Colorado, and Lurie's Children's Hospital of Chicago.

The project's second aim is to apply the validated PSE definition to geographically and demographically diverse datasets to estimate the burden of pediatric sepsis and mortality. This will involve using EHR data from hundreds of U.S. hospitals. The last aim is to develop a PSE toolkit that hospitals can use to facilitate local sepsis surveillance.

"I want to see outcomes for pediatric sepsis improve," Dr. Balamuth said. "And for them to improve, we need to be able to count it well and figure out where we're struggling. We hope this project will allow us to identify sepsis patterns in the United States so that we can tailor our future improvement efforts."