The source article did not report specific renal endpoints or detailed operational metrics beyond hospital length of stay and ICU admission. Investigators limited reported outcomes to four reference endpoints: shorter hospital length of stay (HLOS <72 hours), no ICU stay, survival to discharge, and absence of sepsis adjudication per institutional standards. Any inference about kidney-specific outcomes or other operational measures would exceed the material presented in the preprint.
The cohort comprised 9,443 emergency department visits from five hospitals in New York City, queried from electronic health records spanning 2022–2024. The cohort skewed toward younger adults: 45.9% were under 70 years of age. A majority were not febrile on presentation; 68.1% were described as normothermic. Most were not hypotensive at triage—64.7% were normotensive. Laboratory values captured at presentation showed that 32.0% had white blood cell counts in the normal range, and 17.6% had lactate values reported as normal. Fewer than one-quarter of pathway activations carried an ED diagnosis of sepsis.
Across the full cohort, the study team observed that 84.5% of patients survived to hospital discharge. Two-thirds avoided an ICU stay: 65.9% had no ICU admission during that encounter. A relatively small fraction experienced a short hospital admission by the chosen threshold—8.5% had HLOS under 72 hours. Finally, 28.0% of encounters were not adjudicated as sepsis according to the institution’s quality-of-care standards. The authors used these four outcomes as reference measures for identifying lower-risk patient features.
Investigators used multivariable logistic regression to evaluate associations between patient characteristics and the four favorable outcomes. Several features appeared repeatedly across models. Age emerged as a consistent predictor: younger age groups, particularly those 18–49 years, had the strongest association with shorter HLOS (<72 hours) and with survival to discharge. Temperature measurements mattered; patients with temperature >36°C were linked to multiple favorable endpoints. Hemodynamic stability at presentation—absence of hypotension—was associated with improved outcomes across some models. A white blood cell count between 4,000 and 15,000/µL was the sole variable associated with all four favorable outcomes, with adjusted odds ratios reported in the source. Lower lactate values (<2.0 mmol/L) were also connected to several favorable endpoints. The report emphasized that the set of features predictive of each favorable outcome was not identical: variables important for predicting no ICU stay were not always the same as those predicting HLOS <72 hours or no sepsis adjudication.
The study team classified primary ED diagnoses and compared outcome patterns across those categories. Patients with a primary ED diagnosis of infection other than pneumonia had more favorable outcomes than those diagnosed with sepsis or with pneumonia. Lack of a severe diagnosis in the ED had the strongest association with avoidance of ICU admission. Conversely, an ED diagnosis other than sepsis or infection had the strongest association with not being adjudicated as having sepsis, according to institutional review. These diagnostic signals influenced model performance and were among the most discriminating variables for specific endpoints.
This work was a retrospective EHR-based cohort study and the authors acknowledged limitations inherent to that design. The preprint does not report prospective validation of the identified features. Investigators noted that the predictive factors differed by the chosen outcome, limiting the immediate use of a single rule for de-escalation. The source did not publish granular timing, dosing, or protocolized elements of sepsis pathway care that might influence outcomes. The authors called for further research to prospectively evaluate how well these identified clinical features perform in real-time identification of patients most likely to experience favorable outcomes and who might therefore be candidates for sepsis care de-escalation.