Features Partner Sites Information LinkXpress hp
Sign In
Advertise with Us
IBA-Radcal

Download Mobile App




Events

08 Dec 2026 - 10 Dec 2026
25 Jan 2027 - 28 Jan 2027

Deep Learning Model Predicts Sepsis Patients Likely to Benefit from Steroid Treatment

By HospiMedica International staff writers
Posted on 24 Nov 2025

Sepsis continues to be one of the toughest problems in critical care, with a chaotic immune response that can push patients into multi-organ failure within hours. More...

Even with modern intensive care, only about 60–70% of patients with septic shock survive the first 30 days. Corticosteroids remain one of the most hotly debated treatments — they help some patients but may harm others, and traditional clinical trials aren’t built to untangle that kind of individual variability. Now, researchers have used deep learning to identify patients with sepsis who are most likely to respond to corticosteroids, improving treatment precision.

A research team from the Department of Intensive Care Medicine, Amsterdam UMC, Netherlands) turned to causal deep learning to pinpoint which patients truly benefit. They built a predictive model using TARNet, focusing on 28-day mortality, and trained it on 2,920 Sepsis-3 patients in the AmsterdamUMCdb database. Each profile included 19 commonly collected clinical variables such as lactate, pH, and the PaO₂/FiO₂ ratio.

To test how well the model held up elsewhere, they validated it on the large American MIMIC-IV v2.2 dataset of 30,639 patients. In findings published in the Journal of Intensive Medicine, the model showed a strong performance: AUROC was 0.79 internally and 0.71 in the external cohort, with Brier scores of 0.14 in both sets. Calibration remained robust, and TARNet achieved near-perfect covariate balance, significantly outperforming conventional propensity score matching.

Researchers then grouped patients based on a clinically meaningful 10% shift in predicted 28-day mortality. This created three subgroups: responders (245 patients), non-responders (2,098 patients), and those likely to be harmed (577 patients). Individuals with severe metabolic acidosis and circulatory dysfunction, characterized by low pH, low bicarbonate, elevated lactate, and higher creatinine, were the most likely to benefit from corticosteroids. These findings align with current understanding of sepsis physiology.

By learning shared representations before estimating each patient’s potential outcome, TARNet could cleanly separate benefit from no benefit or harm, while its excellent covariate balancing reduced confounding. The dual-database approach also added weight to the findings, with the model performing consistently even though the MIMIC-IV population was less severely ill. Importantly, all 19 input variables are routinely collected worldwide, and the response threshold mirrors real-world decision-making, making the results highly practical for ICUs.

“Our study addresses a critical gap in current practice. Causal deep learning enables the estimation of individualized treatment effects, overcoming the limitations of traditional population level studies,” said Dr. Ameet Jagesar. “This is the first application of the TARNet model in the context of corticosteroid therapy for sepsis.”

Related Links:
Amsterdam UMC


Gold Member
NEW PRODUCT : SILICONE WASHING MACHINE TRAY COVER WITH VICOLAB SILICONE NET VICOLAB®
REGISTRED 682.9
Hand-held Hemoglobin Analyzer
DiaSpect Tm
Creatinine/eGFR Meter
StatSensor® Creatinine/eGFR Meter
Resorbable Bovine Collagen Membrane
GenDerm
Read the full article by registering today, it's FREE! It's Free!
Register now for FREE to HospiMedica.com and get access to news and events that shape the world of Hospital Medicine.
  • Free digital version edition of HospiMedica International sent by email on regular basis
  • Free print version of HospiMedica International magazine (available only outside USA and Canada).
  • Free and unlimited access to back issues of HospiMedica International in digital format
  • Free HospiMedica International Newsletter sent every week containing the latest news
  • Free breaking news sent via email
  • Free access to Events Calendar
  • Free access to LinkXpress new product services
  • REGISTRATION IS FREE AND EASY!
Click here to Register








Channels

Artificial Intelligence

view channel
Image Credit: Shutterstock

Explainable AI Examines Sepsis Mortality Risk and Emergency Care Delays

Preventing sepsis-related deaths in emergency departments (EDs) remains challenging because patient deterioration can result from multiple rapidly evolving and interconnected factors. These include triage... Read more

Patient Care

view channel
Image: A view inside the custom virtual reality environment used in the study: a calm room with a teddy bear, soft colors and large windows looking out over the sea. (Image Credit: Walaa S. Abu Rukbah et al., Virginia Commonwealth University; CC-BY 4.0, https://creativecommons.org/licenses/by/4.0/)

Virtual Reality Program May Reduce Anxiety Before Pediatric Surgery

Presurgery anxiety affects an estimated 50% to 75% of children undergoing surgery and can complicate anesthesia induction, increase postoperative pain, and prolong recovery. Sedation is often used to manage... Read more

Medical Imaging

view channel
Image: A new automated approach could help predict which patients with dystonia will benefit from deep brain stimulation before surgery (Image Credit: Adobe Stock)

AI Uses Routine MRI to Predict Deep Brain Stimulation Response in Dystonia

Dystonia is a debilitating neurological movement disorder characterized by involuntary muscle contractions that cause abnormal movements and postures. Deep brain stimulation (DBS) is one of the few treatments... Read more

Health IT

view channel
Image Credit: Shutterstock

Telemonitoring Program Improves Quality of Life After Heart Failure Discharge

Heart failure is a chronic condition associated with high mortality, frequent hospitalizations, and reduced quality of life. Patients face particularly high risks of readmission and death after discharge,... Read more

Point of Care

view channel
Image: Jessica Ramella-Roman, professor of biomedical engineering at Florida International University, holding the PPRIM (Portable Preterm Imaging) device designed to detect early signs of preterm labor risk. (Photo courtesy of Florida International University)

Handheld Imaging Device May Help Identify Preterm Birth Risk Earlier

Preterm birth, defined as delivery before 37 weeks of pregnancy, can lead to serious neonatal complications, including neurological, developmental, and respiratory problems. However, identifying pregnancies... Read more

Business

view channel
Image: Stryker has acquired ZuriMED, adding the FiberLocker System, a rotator cuff augmentation technology designed to strengthen repairs, to its Sports Medicine portfolio. (Photo courtesy of ZuriMED)

Stryker Completes Acquisition to Broaden Shoulder Surgery Offerings

Stryker (Portage, MI, USA) has completed its acquisition of ZuriMED (Zurich, Switzerland), a privately held developer of the FiberLocker System, expanding its Sports Medicine portfolio with a commercialized... Read more
Copyright © 2000-2026 Globetech Media. All rights reserved.