Pancreas
July 1, 2026 · via Europe PMC
AI-Driven Test-Free Prediction of ICU Admission, Insulin Dependence, and Exocrine Dysfunction After Acute Pancreatitis.
Chattopadhyay I, Onishchenko D, Kern PA, Conwell DL.
AI summary — for doctors, generated from the abstract only
A claims-based AI model predicts ICU admission, exocrine dysfunction and insulin dependence after acute pancreatitis with high accuracy, but needs external validation.
What was studied
Model development and validation on a US administrative claims database of over 164 million people, using only coded history to predict ICU admission, incident exocrine pancreatic dysfunction and incident insulin dependence.
Key findings
- AUC for same-day ICU admission was 0.986, 0.933 within 1 week and 0.927 within 2 weeks
- AUC for incident exocrine dysfunction was 0.913 in men and 0.901 in women
- AUC for insulin dependence was 0.861 in men and 0.884 in women
Clinical relevance
If validated, risk stratification at first presentation without extra tests could guide monitoring and early follow-up for late pancreatic sequelae.
Who should care
Limitations
Claims data coding may leak information about care, and the authors note external validation outside US systems is still needed.
This is not yet ready for clinical use; keep current assessment but watch for validated tools.
This is an AI-generated summary of the published abstract, not a clinical recommendation. It may not reflect the full text or subsequent updates — verify against the original source before acting on it.
Abstract
Objectives
Acute pancreatitis (AP) has heterogeneous trajectories: some patients deteriorate rapidly and require ICU care, whereas others develop delayed sequelae such as exocrine pancreatic dysfunction (EPD) and pancreatogenic diabetes. Existing scoring systems are burdensome, not available at first presentation, and poorly suited for forecasting longer-term outcomes.
Methods
Our AI platform operating exclusively on routinely collected EHR predicts 3 outcomes after a first recorded AP diagnosis: (1) ICU admission (same-day, within 1 wk, within 2 wk); (2) incident EPD; and (3) incident insulin dependence among patients without prior diabetes diagnosis or antihyperglycemic prescriptions. Models were trained and validated using a US administrative claims database comprising > 164 million individuals.
Results
We demonstrate area under the receiver operating curves (AUCs) of 0.986 (same-day ICU), 0.933 (ICU within 1 wk), and 0.927 (ICU within 2 wk). For longer-term outcomes, AUCs were 0.913 (male) and 0.901 (female) for incident EPD, and 0.861 (male) and 0.884 (female) for incident insulin dependence. Established chronic pancreatitis risk factors (eg, obesity, tobacco dependence) are recovered with large effect sizes and high significance, supporting epidemiologic plausibility. Negative associations were consistent with etiologic subtypes, competing risks, and health care utilization patterns.
Conclusions
Using only existing coded longitudinal history from a US administrative claims environment, ZeBRA enables simultaneous, test-free prediction of early deterioration and delayed pancreatic sequelae after AP, providing a scalable and interpretable basis for early risk stratification and targeted follow-up across the AP-chronic pancreatitis continuum. Broader clinical adoption will require external validation, especially in non-US health systems and data sets with different coding and care-delivery structures.
DOI
10.1097/mpa.0000000000002697
PMID
42383454
Research metadata (titles, abstracts, authors, publication dates) is sourced from Europe PMC, which aggregates content from PubMed/MEDLINE, PubMed Central, and other providers under the U.S. National Library of Medicine (NLM/NCBI). NLM does not claim copyright on abstracts; the original journal publishers or authors may hold rights to the full articles. AI-generated summaries are derived only from the abstract, never the full text, and are always labeled as AI-generated.
Want more like this?
Browse more from the feed. A free account unlocks the full feed, saving and communities.
Back to the research feed