TOPLINE

Among 250 hospital-onset sepsis events reviewed across nine hospitals, 79% were attributed to infections, and out of 100 infection-related cases selected for preventability assessment, 31 were judged potentially preventable. Respiratory infections accounted for more than half of all infection-related sepsis events, with nonventilator hospital-acquired pneumonia (NV-HAP) representing 44% of cases.

METHODOLOGY

• Researchers conducted a retrospective study to determine the underlying causes and preventability of hospital-onset sepsis events among adults aged 18 years or older who were admitted to nine acute care hospitals between January 2021 and December 2023.
• Hospital-onset sepsis was defined as presumed serious infection and acute organ dysfunction beginning on or after the fourth day of hospitalization, as per the CDC's Adult Sepsis Event criteria. 
• A total of 5212 (0.9%) of 558,037 admissions had hospital-onset sepsis; 250 cases (median age, 72 years; 53% men) were randomly selected for detailed medical record review to ascertain the probability and source of infection.
• Of the infection-related cases, 100 were selected for preventability assessment using a 6-point Likert scale, with scores 1 to 3 categorized as potentially preventable and 4 to 6 as not likely preventable.

TAKEAWAY

• Of 250 hospital-onset sepsis events, 79% (n = 198) were attributed to infections, whereas 21% (n = 52) were linked to noninfectious conditions.
• Among 198 infection-related cases, respiratory infections were the leading source of sepsis, accounting for 56%, driven largely by NV-HAP at 44%, followed by bloodstream infections (17%) and intra-abdominal infections (15%). 
• Only 13% of infection-related cases (25 of 198) fit the definition of currently reportable healthcare-associated infections, and blood cultures confirmed a noncommensal organism in only 22% of cases (44 of 198).
• Among cases evaluated for preventability, 31% were classified as potentially preventable. The two most common gaps in care — each present in 38% of flagged events (22 of 58) — were inadequate treatment of an already-identified infection before it worsened to sepsis and failure to prevent NV-HAP. Common noninfectious triggers included bleeding, cardiac ischemia or arrhythmia, and hypovolemia.

IN PRACTICE

"Hospital-onset sepsis events identify serious complications of hospitalization that are largely missed by current HAI [healthcare-associated infection] surveillance programs yet are highly morbid and often preventable. Hospital-onset sepsis surveillance could provide a more comprehensive framework for identifying preventable harms and expanding opportunities for quality improvement," the authors of the study wrote.

SOURCE

The study was led by Fizza Manzoor, MD, Department of Population Medicine, Harvard Pilgrim Health Care Institute in Boston. It was published online on September 17, 2026, in JAMA Network Open.

LIMITATIONS

The study's small sample size may have limited generalizability. Preventability assessments were retrospective and subjective, carrying risk for cognitive bias.

DISCLOSURES

The study was supported by the Agency for Healthcare Research and Quality and the Thomas O. Pyle Fellowship in the Department of Population Medicine at Harvard Pilgrim Health Care Institute. Two of the authors reported receiving CDC grants and personal fees from UpToDate, with one of them also reporting fees from DynaMed.

This article was created using several editorial tools, including AI, as part of the process. Human editors reviewed this content before publication.