Hierarchical machine learning model integrating clinical history and nursing observations for predicting violent behavior in hospitalized schizophrenia patients

Aug 5, 2026

Developed and validated a hierarchical machine learning model that fuses static clinical history with dynamic nursing observation indicators to predict violent behavior among hospitalized schizophrenia patients. A retrospective cohort of 346 patients (July 2021–July 2024) was labeled by documented aggressive incidents into violent and non-violent groups. Nineteen? no—18 static variables from electronic medical records and 39 weekly dynamic behavioral indicators were used. Regularized logistic regression was selected and combined via weighted fusion, achieving an AUC of 0.8741, outperforming single static (0.7953) and dynamic (0.8003) models.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages16

Reading Time40 min

AccessFree to download

Objective

Methods

Study design and participants

Features and predictive modeling

Results

Conclusion

TYPE Original Research PUBLISHED 22 September 2025 DOI 10.3389/fpsyt.2025.1644341 OPEN ACCESS EDITED BY Animesh Kumar Paul, University of Alberta, Canada REVIEWED BY Georgi Neichev Onchev, Medical University Soﬁa, Bulgaria Ivo Dönnhoff, Heidelberg University Hospital, Germany Bahareh Behroozi Behroozi Asl, University of Alberta, Canada *CORRESPONDENCE Ying Duan [Duanying1333@163.com](Duanying1333@163.com)[ ](Duanying1333@163.com)Wei Yang [epicard@163.com](epicard@163.com) †These authors have contributed equally to this work RECEIVED 10 June 2025 ACCEPTED 02 September 2025 PUBLISHED 22 September 2025 CITATION Meng X, Wang L, Duan Y, Zhu G, Wang J, Sun Y, Wang M, Liu M, Sun C, Pang L, Hu K, Yang W, Shao W, Ren J, Shao X and Zhang Y (2025) Hierarchical machine learning model integrating clinical history and nursing observations for predicting violent behavior in hospitalized schizophrenia patients. Front. Psychiatry 16:1644341 . doi: 10.3389/fpsyt.2025.1644341 COPYRIGHT © 2025 Meng, Wang, Duan, Zhu, Wang, Sun, Wang, Liu, Sun, Pang, Hu, Yang, Shao, Ren, Shao and Zhang. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms. Hierarchical machine learning model integrating clinical history and nursing observations for predicting violent behavior in hospitalized schizophrenia patients Xianfeng Meng 1†, Liang Wang 1†, Ying Duan 2*, Gang Zhu 3, Jinhuan Wang 1, Ying Sun 1, Mingtao Wang 1, Miao Liu 1, Chenhui Sun 1, Longlong Pang 4,5,6, Kunyuan Hu 4,5,6, Wei Yang4,5,6*, Wei Shao 7, Jintao Ren 1, Xiaojun Shao 3 and Yang Zhang 1 1 Liaoning Provincial Mental Health Center, Tieling, Liaoning, China, 2 Liaoning Maternal and Child Health Hospital, Shenyang, Liaoning, China, 3 Department of Psychiatry, The First Afﬁliated Hospital of China Medical University, Shenyang, Liaoning, China, 4 Key Laboratory of Networked Control Systems, Chinese Academy of Sciences, Shenyang, China, 5Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang, China, 6 University of Chinese Academy of Sciences, Beijing, China, 7Shengjing Hospital of China Medical University, Shenyang, Liaoning, China Objective: To develop and validate a hierarchical machine learning model integrating static clinical features and dynamic behavioral assessments for accurately predicting violent behaviors among hospitalized schizophrenia patients. Methods: This retrospective study included 346 schizophrenia patients hospitalized from July 2021 to July 2024 in Liaoning Province. Patients were categorized into violent (n = 123) and non-violent (n = 223) groups based on documented aggressive incidents. Eighteen static clinical variables (e. g., age, gender, history of violence, manic symptoms) were extracted from electronic medical records, and 39 dynamic behavioral indicators (e. g., anger expression, insomnia, auditory hallucinations) were assessed weekly using the Psychiatric Patient Nursing Observation Scale. Predictive models were separately developed using six machine learning algorithms: Regularized Logistic Regression (LR), Support Vector Machine (SVM), Extreme Gradient Boosting (XGBoost), Random Forest (RF), Multi-layer Perceptron (MLP), and K-Nearest Neighbor (KNN) . Regularized logistic regression was selected as the ﬁnal algorithm due to its superior predictive performance, indicated by the highest Area Under the Curve (AUC), in both static baseline and dynamic behavioral models. A hierarchical predictive model was then established using regularized logistic regression separately for static baseline risk and dynamic risk ﬂuc

FAQs

What is the purpose of the hierarchical model?

To accurately predict violent behavior in hospitalized schizophrenia patients by integrating static clinical history with dynamic nursing observation data.

How were violent and non-violent groups defined?

Patients were categorized using documented aggressive incidents recorded during hospitalization into violent (n=123) and non-violent (n=223) groups.

Which algorithm and evaluation metric produced the best performance?

Regularized logistic regression was chosen, and the integrated hierarchical model achieved the highest predictive performance with an AUC of 0.8741.

Aug 5, 2026

Upload by Clementine