Related Experiment Video
Updated: Jul 2, 2026
Characterization of Metabolic Status in Nonhuman Primates with the Intravenous Glucose Tolerance Test
Published on: November 13, 2016
AI-driven Test-Free Prediction of ICU Admission, Insulin Dependence, and Exocrine Dysfunction after Acute
Ishanu Chattopadhyay1, Dmytro Onishchenko1, Philip Kern2,3
1Division of Biomedical Informatics, Department of Internal Medicine, University of Kentucky, Lexington, KY USA.
Insights
An AI platform accurately predicts acute pancreatitis (AP) complications, including ICU admission and long-term issues like exocrine pancreatic dysfunction and diabetes, using electronic health records. This tool aids early risk stratification for better patient management.
Area of Science:
- Artificial Intelligence in Medicine
- Health Informatics
- Clinical Prediction Models
Background:
- Acute pancreatitis (AP) presents diverse clinical trajectories, ranging from rapid deterioration requiring intensive care to delayed complications such as exocrine pancreatic dysfunction (EPD) and pancreatogenic diabetes.
- Current scoring systems for AP are often impractical for initial assessment and lack predictive power for long-term outcomes.
Purpose of the Study:
- To develop and validate an AI-driven platform for predicting both early and delayed complications following a first-time diagnosis of acute pancreatitis.
- To provide a scalable and interpretable tool for early risk stratification and targeted patient follow-up across the AP-chronic pancreatitis spectrum.
Main Methods:
- An AI platform was developed using routinely collected electronic health record (EHR) data from a large U.S. administrative claims database.
- The platform was trained and validated to predict three key outcomes: ICU admission (same-day, within 1 week, within 2 weeks), incident EPD, and incident insulin dependence.
- Model performance was assessed using the Area Under the Receiver-operating curve (AUC).
Main Results:
- The AI platform achieved high predictive accuracy for ICU admission, with AUCs of 0.986 (same-day), 0.933 (1-week), and 0.927 (2-week).
- Predictive accuracy for longer-term outcomes was also strong: AUCs for incident EPD were 0.913 (male) and 0.901 (female), and for incident insulin dependence were 0.861 (male) and 0.884 (female).
- The model identified established chronic pancreatitis risk factors (e.g., obesity, tobacco dependence) with significant effect sizes, confirming epidemiologic plausibility.
Conclusions:
- The AI platform, named ZeBRA, enables simultaneous, test-free prediction of early deterioration and delayed sequelae after AP using existing EHR data.
- ZeBRA offers a scalable and interpretable foundation for early risk stratification and tailored follow-up for patients with AP.
- External validation in diverse health systems is recommended for broader clinical adoption.
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 predict three outcomes after a first recorded AP diagnosis: (i) ICU admission (same-day, within 1 wk, within 2 wk), (ii) incident EPD, and (iii) incident insulin dependence among patients without prior diabetes diagnosis or anti-hyperglycemic prescriptions. Models were trained and validated using a U.S. administrative claims database comprising 164 million individuals.
Results:
We demonstrate Area Under the Receiver-operating curve (AUC) 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 (e.g., 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 healthcare utilization patterns.
Conclusions:
Using only existing coded longitudinal history from a U.S. 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-U.S. health systems and datasets with different coding and care-delivery structures.
Related Concept Videos
01:30Acute Pancreatitis II: Clinical Manifestations and Management
01:25Acute Pancreatitis I: Introduction
01:27Acute Pancreatitis I: Introduction
Acute pancreatitis is characterized by rapid inflammation of the pancreas, often caused by factors like gallstone blockage or excessive alcohol consumption. Chronic pancreatitis, on the other hand, is a slow, progressive inflammation that may result from long-term alcohol abuse, obstructions in the pancreatic duct, or genetic factors.
The causes of acute pancreatitis include:
01:12Type I Diabetes I: Introduction
01:29Chronic Pancreatitis II: Collaborative Care
Assessment:
01:21Acute Pancreatitis II: Pathophysiology