Related Experiment Video
Updated: Jan 9, 2026
07:31
Implementation of a Real-Time Psychosis Risk Detection and Alerting System Based on Electronic Health Records using CogStack
Published on: May 15, 2020
7.5K
ClinPreAI: An Agentic AI System for Early Postpartum Depression Risk Prediction from Multimodal EHR Data
Daniel Palacios1,2,3, Sukru Aras4,2,3, Yi Zhong4,2,3
1Quantitative Computational Biosciences, Baylor College of Medicine, Houston, Texas 77030, USA.
Medrxiv : the Preprint Server for Health Sciences
|December 3, 2025
Summary
A new AI system, ClinPreAI, can predict postpartum depression (PPD) risk using electronic health records. This autonomous agent improves early identification, making advanced predictive modeling more accessible for maternal mental health.
Area of Science:
- Artificial Intelligence
- Clinical Informatics
- Mental Health
Background:
- Postpartum depression (PPD) affects 10-15% of mothers annually, with early identification being a significant clinical challenge.
- Existing methods for PPD risk prediction often lack efficiency and accessibility in clinical settings.
Purpose of the Study:
- To introduce ClinPreAI, an agentic AI system designed for autonomous development and evaluation of PPD risk prediction models.
- To leverage multimodal electronic health record (EHR) data for enhanced PPD prediction accuracy.
Main Methods:
- Analysis of EHR data from 4,161 pregnant individuals, including 27 structured clinical variables and social worker notes.
- Development and implementation of ClinPreAI, an agentic AI system with five modules for iterative model refinement via autonomous experimentation.
- Evaluation of predictive performance using the Edinburgh Postnatal Depression Scale (EPDS) score ≥10 as the primary outcome.
Main Results:
- ClinPreAI achieved an F1 score of 0.68 ± 0.03 on structured data, surpassing traditional AutoML and commercial solutions.
- On multimodal data, ClinPreAI achieved an F1 score of 0.65 ± 0.04, matching custom LLM-XGBoost and outperforming zero-shot models.
- Demonstrated the capability of agentic AI in democratizing sophisticated predictive modeling for clinical applications.
Conclusions:
- Agentic AI systems like ClinPreAI can automate the design, implementation, and evaluation of clinical prediction tools.
- This approach lowers barriers for developing robust predictive models in healthcare, especially where ML expertise is limited.
- ClinPreAI represents a significant advancement in applying autonomous AI for perinatal mental health prediction and clinical decision support.
