By Jim Sianis, PharmD, MBA, Executive Director of Business Development, Prenosis
LinkedIn: Jim Sianis, B.S., PharmD, MBA
LinkedIn: Prenosis
Sepsis remains one of the most complex and deadly conditions treated in emergency departments (EDs) today. The SEP-1 bundle is a challenging measure for hospitals. The bundle mandates strict compliance with early resuscitation criteria based on limited and sometimes conflicting information about the patient. In 2024, CMS moved SEP-1 bundle compliance from pay-for-reporting to pay-for-performance, which will result in either incentives or penalties for providers in 2026.
A Unique Approach
A 2025 study published in JAMA indicates that an all-or-nothing approach with regard to SEP-1 bundle compliance may miss the bigger picture. Sepsis is a heterogenous condition; patients present differently. Real-world sepsis care is nuanced, which means a rigid adherence to SEP-1 may not suffice. The multicenter cohort study reviewed 590 ED patients with sepsis and found that SEP-1 noncompliance was frequently associated with more complex, atypical presentations such as no fever, altered mental status, bedside procedures, and overlapping noninfectious conditions (e.g., heart failure or cancer). While these factors delayed SEP-1 bundle completion, they did not necessarily reflect substandard clinical care. After adjusting for these complexities, SEP-1 compliance had no statistically significant impact on mortality.
Sepsis ImmunoScore™, the first FDA-authorized AI diagnostic for sepsis prediction, gives providers the ability to navigate the complex early sepsis presentation in real-time.
This helps to address the problem highlighted in the JAMA study: delayed initiation of SEP-1 bundle elements for the right patients.
Sepsis ImmunoScore shifts the paradigm of sepsis diagnosis by helping providers make individualized decisions based on the patient’s biology during the first critical hours of presentation. Results from the diagnostic tool can help determine which patients should be placed into the SEP-1 bundle, should go to the ICU, or could potentially avoid antibiotics if there is no evidence of infection.
How it works:
- Sepsis ImmunoScore embeds directly into the clinical workflow and utilizes 22 parameters, integrating vital signs, routine labs, and biomarkers – including CRP and procalcitonin – without requiring new equipment or specialized training.
- Patients are categorized into four clearly defined risk levels (e.g., “Very High Risk”, “High Risk”, “Medium Risk, “Low Risk”) for progression to sepsis within 24 hours, providing valuable insights to act early and initiate appropriate treatment.
- An intuitive, transparent display of the algorithm’s results allows providers to visualize what drives each patient’s individual risk of developing sepsis.
- Higher Sepsis ImmunoScore risk scores are highly correlated with the risk of clinical deterioration defined as need for ICU in 24 hours, need for vasopressors in 24 hours, need for mechanical ventilation in 24 hours, in-hospital mortality and length of stay.
Instead of providers trying to check boxes for every suspected sepsis case and potentially overutilizing resources such as antibiotics, nursing time and ICU beds, Sepsis ImmunoScore accurately risk stratifies patients to facilitate delivery of the appropriate level of care.
This is extremely valuable for:
- Patients without fever or with unclear infectious sources to avoid unnecessary use of antibiotics;
- Elderly or comorbid sepsis patients with atypical symptoms; or
- Resource-constrained settings where time and staff are limited.
The future of sepsis care must move beyond checkboxes. The study published in JAMA underscores how high-quality care can occur in complex cases yet still lead to noncompliance of the SEP-1 bundle. Sepsis ImmunoScore is designed precisely for this gray area, helping clinicians accurately diagnose, personalize interventions, and appropriately treat sepsis patients.
For more on this topic, read our article in NEJM AI on the use of Sepsis ImmunoScore to predict adverse outcomes.
This article was originally published on the Prenosis blog and is republished here with permission.