Learning the electronic health record at the minute-scale
- 1Stanford University
Abstract
Electronic health record foundation models are traditionally trained on the scale of years, days, or hours. These timescales, however, lack the minute-scale resolution required to directly guide clinical decisions at the bedside. We introduced minute-scale learning, a new paradigm for training and evaluating models, and developed MINT, a minute-scale foundation model for pediatric emergencies. MINT was pretrained and validated on 766,733 pediatric emergency department visits at five health systems, comprising 16 years of data from 10 hospitals. On minute-scale forecasting tasks, MINT outperformed and generalized to external health systems better than task-specific models (superior in 21 of 25 comparisons). MINT outperformed physicians in forecasting escalations of respiratory support. MINT demonstrated uniquely minute-scale capabilities including department-scale monitoring, dynamic risk explanations, individualized physiologic response forecasts, and hypothesis generation. Learning at the minute-scale improves performance, strengthens generalizability, and provides insights, actionability, and scientific capabilities that are not accessible at other timescales.
Related
- Machine Learning Models Using Electronic Health Record Data to Predict Obstructive Sleep Apnea
- Electronic Health Record-Based Machine-Learning for Cardiac Arrhythmia Identification in Adults with High-Risk Chronic Disease: Model Development and Validation Study
- Machine Learning-Supported Efficient VTE Risk Assessment using Routinely Collected Electronic Health Record Data
- Machine Learning–Based Prediction of Major Postoperative Complications After Noncardiac Surgery Under General Anesthesia Using Routinely Available Electronic Health Record Data: Retrospective Multicenter Development and Validation Study
- Comparative health outcomes of buprenorphine versus methadone treatment for opioid use disorder: a retrospective propensity score-matched cohort study in the global federated electronic health record network
BibTeX
@misc{narang2026learning,
title = {{Learning the electronic health record at the minute-scale}},
author = {K. Narang and N. Addo and J. Bains and M. Firnberg and S. Tat and A. Hooft and D. C. Arrubla and B. Abed and C. Singh and R. Kamaleswaran and J. Feng and A. E. Kornblith},
howpublished = {medRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.04.26364562},
url = {https://www.medrxiv.org/content/10.64898/2026.10.04.26364562v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Learning the electronic health record at the minute-scale
AU - K. Narang
AU - N. Addo
AU - J. Bains
AU - M. Firnberg
AU - S. Tat
AU - A. Hooft
AU - D. C. Arrubla
AU - B. Abed
AU - C. Singh
AU - R. Kamaleswaran
AU - J. Feng
AU - A. E. Kornblith
PY - 2026
DA - 2026/10/04
PB - medRxiv
DO - 10.64898/2026.10.04.26364562
UR - https://www.medrxiv.org/content/10.64898/2026.10.04.26364562v1
AB - Electronic health record foundation models are traditionally trained on the scale of years, days, or hours. These timescales, however, lack the minute-scale resolution required to directly guide clinical decisions at the bedside. We introduced minute-scale learning, a new paradigm for training and evaluating models, and developed MINT, a minute-scale foundation model for pediatric emergencies. MINT was pretrained and validated on 766,733 pediatric emergency department visits at five health systems, comprising 16 years of data from 10 hospitals. On minute-scale forecasting tasks, MINT outperformed and generalized to external health systems better than task-specific models (superior in 21 of 25 comparisons). MINT outperformed physicians in forecasting escalations of respiratory support. MINT demonstrated uniquely minute-scale capabilities including department-scale monitoring, dynamic risk explanations, individualized physiologic response forecasts, and hypothesis generation. Learning at the minute-scale improves performance, strengthens generalizability, and provides insights, actionability, and scientific capabilities that are not accessible at other timescales.
ER -