Maintaining EHR-based Predictive Clinical Decision Support in Mental Health against Temporal Shifts
- 1Peking University
Abstract
Electronic health record (EHR)-based risk prediction offers opportunities for early intervention in mental health, but models may remain in use as patient populations, care practices, and documentation change. Such temporal shifts can reduce predictive performance, while updating models from recent data alone may be unstable when adverse outcomes are uncommon. We developed Adversarial Drift-Aware Predictive Transfer (ADAPT), a distributionally robust approach to improve the temporal durability of mental health risk prediction. ADAPT uses variation in models estimated from previous calendar periods to characterize plausible changes in predictor-outcome relationships. Data from the current period constrain which historical patterns remain relevant, allowing the method to seek robustness against subsequent shifts while preserving contemporary discrimination. We retrospectively evaluated ADAPT using EHRs from 889,445 patients with mental health diagnoses at Mass General Brigham and Duke University Health System, covering one-year risk prediction for 28 outcomes. In the worst 20% of year-pair comparisons, ADAPT improved AUC over all comparators by 0.016-0.036 at Mass General Brigham, 0.011-0.018 at Duke, and 0.008-0.037 after transport. For suicide-attempt prediction, contemporary AUC at Mass General Brigham was 0.82 compared with 0.80 for the best comparator. Eight-year relative AUC degradation was 7.7%, compared with 12.1% for pooled-data training. These results suggest that information from prior temporal shifts can improve robustness to future changes without sacrificing predictive performance for current patients.
Related
- Decision-support modelling for liver paired exchange integrating clinical thresholds and operational feasibility
- Real-world evaluation of a multilingual agentic retrieval-augmented clinical AI framework for evidence-grounded medical decision support
- Extracting smoking history from clinical notes for lung cancer screening decision support: comparing a structured-judgment model with general-purpose large language models
- Evaluating a Clinical Decision Support System for Chronic Pain Management: A Per-Protocol Analysis
- An automated, explainable, NCCT-based clinical decision-support system for spontaneous intracerebral hemorrhage.
BibTeX
@misc{xiong2026maintaining,
title = {{Maintaining EHR-based Predictive Clinical Decision Support in Mental Health against Temporal Shifts}},
author = {X. Xiong and E. Shi and Z. Qian and H. Zhu and C. Hong and Z. Guo and J. W. Smoller and M. Liu and T. Cai},
howpublished = {medRxiv},
year = {2026},
month = oct,
doi = {10.64898/2026.10.28.26366683},
url = {https://www.medrxiv.org/content/10.64898/2026.10.28.26366683v1},
note = {CC BY},
}
RIS
TY - UNPB
TI - Maintaining EHR-based Predictive Clinical Decision Support in Mental Health against Temporal Shifts
AU - X. Xiong
AU - E. Shi
AU - Z. Qian
AU - H. Zhu
AU - C. Hong
AU - Z. Guo
AU - J. W. Smoller
AU - M. Liu
AU - T. Cai
PY - 2026
DA - 2026/10/28
PB - medRxiv
DO - 10.64898/2026.10.28.26366683
UR - https://www.medrxiv.org/content/10.64898/2026.10.28.26366683v1
AB - Electronic health record (EHR)-based risk prediction offers opportunities for early intervention in mental health, but models may remain in use as patient populations, care practices, and documentation change. Such temporal shifts can reduce predictive performance, while updating models from recent data alone may be unstable when adverse outcomes are uncommon. We developed Adversarial Drift-Aware Predictive Transfer (ADAPT), a distributionally robust approach to improve the temporal durability of mental health risk prediction. ADAPT uses variation in models estimated from previous calendar periods to characterize plausible changes in predictor-outcome relationships. Data from the current period constrain which historical patterns remain relevant, allowing the method to seek robustness against subsequent shifts while preserving contemporary discrimination. We retrospectively evaluated ADAPT using EHRs from 889,445 patients with mental health diagnoses at Mass General Brigham and Duke University Health System, covering one-year risk prediction for 28 outcomes. In the worst 20% of year-pair comparisons, ADAPT improved AUC over all comparators by 0.016-0.036 at Mass General Brigham, 0.011-0.018 at Duke, and 0.008-0.037 after transport. For suicide-attempt prediction, contemporary AUC at Mass General Brigham was 0.82 compared with 0.80 for the best comparator. Eight-year relative AUC degradation was 7.7%, compared with 12.1% for pooled-data training. These results suggest that information from prior temporal shifts can improve robustness to future changes without sacrificing predictive performance for current patients.
ER -