A clinical prediction model based on interpretable machine learning algorithms for prolonged hospital stay in acute ischemic stroke patients: a real-world study

Aug 5, 2026

A clinical prediction model predicts prolonged length of stay (LOS) in acute ischemic stroke (AIS) using an interpretable machine learning approach. Patients were categorized into prolonged LOS versus no prolonged LOS, with prolonged LOS defined as hospitalization longer than 7 days. LASSO regression selected predictors and eight algorithms were compared for discrimination, calibration, and clinical utility. The Gaussian naive Bayes model achieved strong performance with high AUC in training and validation sets. Key predictors included pneumonia, dysphagia, thrombectomy, and stroke severity.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages11

Reading Time28 min

AccessFree to download

Objective

Methods

Study population and outcome definition

Feature selection and model comparison

Model interpretation and evaluation

Results

Conclusions

TYPE Original Research PUBLISHED 22 November 2023 DOI 10.3389/fendo.2023.1165178 OPEN ACCESS EDITED BY Prem P. Kushwaha, Case Western Reserve University, United States REVIEWED BY Jianbo Chang, Peking Union Medical College and Chinese Academy of Medical Sciences, China Hai-Yang Wang, Jining First People’s Hospital Afﬁliated to Shandong First Medical University, China Liang Pan, People’s Hospital of Deyang City, China Qiang He, Sichuan University, China *CORRESPONDENCE Wenle Li [drlee0910@163.com](drlee0910@163.com)[ ](drlee0910@163.com)Liangqun Rong [rongliangqun@163.com](rongliangqun@163.com)[ ](rongliangqun@163.com)Haosheng Wang [Dr_haosheng@163.com](Dr_haosheng@163.com) †These authors have contributed equally to this work RECEIVED 13 February 2023 ACCEPTED 21 April 2023 PUBLISHED 22 November 2023 CITATION Wang K, Jiang Q, Gao M, Wei X, Xu C, Yin C, Liu H, Gu R, Wang H, Li W and Rong L (2023) A clinical prediction model based on interpretable machine learning algorithms for prolonged hospital stay in acute ischemic stroke patients: a realworld study. Front. Endocrinol. 14:1165178 . doi: 10.3389/fendo.2023.1165178 COPYRIGHT © 2023 Wang, Jiang, Gao, Wei, Xu, Yin, Liu, Gu, Wang, Li and Rong. This is an openaccess 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. A clinical prediction model based on interpretable machine learning algorithms for prolonged hospital stay in acute ischemic stroke patients: a realworld study Kai Wang1,2†, Qianmei Jiang 3†, Murong Gao 4†, Xiu’e Wei 1,2, Chan Xu 5, Chengliang Yin 6, Haiyan Liu 1,2, Renjun Gu 7, Haosheng Wang 7,8*, Wenle Li 2,9* and Liangqun Rong 1,2* 1 Department of Neurology, The Second Afﬁliated Hospital of Xuzhou Medical University, Xuzhou, Jiangsu, China, 2 Key Laboratory of Neurological Diseases, The Second Afﬁliated Hospital of Xuzhou Medical University, Xuzhou, Jiangsu, China, 3 Department of General Practice, Xindu District People’s Hospital of Chengdu, Chengdu, Sichuan, China, 4 Department of Rehabilitation, Beijing Rehabilitation Hospital Afﬁliated to Capital Medical University, Beijing, China, 5 Department of Dermatology, Xianyang Central Hospital, Xianyang, China, 6 Faculty of Medicine, Macau University of Science and Technology, Macau, Macao SAR, China, 7School of Chinese Medicine and School of Integrated Chinese and Western Medicine, Nanjing University of Chinese Medicine, Nanjing, China, 8State Key Laboratory of Pharmaceutical Biotechnology, Division of Sports Medicine and Adult Reconstructive Surgery, Department of Orthopedic Surgery, Nanjing Drum Tower Hospital, The Afﬁliated Hospital of Nanjing University Medical School, Nanjing, Jiangsu, China, 9The State Key Laboratory of Molecular Vaccinology and Molecular Diagnostics and Center for Molecular Imaging and Translational Medicine, School of Public Health, Xiamen University, Xiamen, China Objective: Acute ischemic stroke (AIS) brings an increasingly heavier economic burden nowadays. Prolonged length of stay (LOS) is a vital factor in healthcare expenditures. The aim of this study was to predict prolonged LOS in AIS patients based on an interpretable machine learning algorithm. Methods: We enrolled AIS patients in our hospital from August 2017 to July 2019, and divided them into the “prolonged LOS” group and the “no prolonged LOS”group. Prolonged LOS was deﬁned as hospitalization for more than 7 days. The least absolute shrinkage and selection operator (LASSO) regression was applied to reduce the dimensionality of the data. We compared the predictive capacity of extended LOS in eight differ

FAQs

What outcome did the study aim to predict in acute ischemic stroke patients?

The study predicted prolonged length of stay (LOS) in acute ischemic stroke patients.

How was prolonged LOS defined?

Prolonged LOS was defined as hospitalization for more than 7 days.

Which model performed best and which factors were strongest predictors?

The Gaussian naive Bayes model showed the best prediction performance. Strong predictors included pneumonia, dysphagia, thrombectomy, and stroke severity.

Aug 5, 2026

Upload by Lucas Martin