Retrospective Study
Copyright ©The Author(s) 2025. Published by Baishideng Publishing Group Inc. All rights reserved.
World J Methodol. Sep 20, 2025; 15(3): 101837
Published online Sep 20, 2025. doi: 10.5662/wjm.v15.i3.101837
Machine learning-based models for prediction of in-hospital mortality in patients with dengue shock syndrome
Luan Thanh Vo, Thien Vu, Thach Ngoc Pham, Tung Huu Trinh, Thanh Tat Nguyen
Luan Thanh Vo, Thach Ngoc Pham, Tung Huu Trinh, Thanh Tat Nguyen, Department of Infectious Diseases, Children's Hospital No. 2, Ho Chi Minh City 700000, Viet Nam
Thien Vu, AI Nutrition Project, National Institutes of Biomedical Innovation, Health and Nutrition (NIBIOHN), Osaka 5670001, Japan
Thien Vu, NCD Epidemiology Research Center, Shiga University of Medical Science, Otsu, Shiga 5200003, Japan
Thanh Tat Nguyen, Department of Tuberculosis and Epidemiology, Woolcock Institute of Medical Research, Ho Chi Minh City 700000, Viet Nam
Co-corresponding authors: Thien Vu and Thanh Tat Nguyen.
Author contributions: Nguyen TT, Vo LT, Pham TN, Trinh TH contributed to conceptualization, funding acquisition; Nguyen TT, Vo LT contributed to data curation, investigation; Nguyen TT, Vu T contributed to formal analysis; Nguyen TT, Vo LT, Pham NT contributed to methodology; Vo LT, Vu T, Nguyen TT contributed to writing-original draft; Vo LT, Vu T, Pham NT, Trinh TH, Nguyen TT contributed to revision of the final manuscript. All authors have contributed to and approved the final manuscript.
Institutional review board statement: This sub-study stemmed from the primary published study, “Prognostic values of serum lactate-to-bicarbonate ratio and lactate for predicting 28-day in-hospital mortality in children with dengue shock syndrome”. The primary study was approved by the Scientific Committee and Institutional Review Board (IRB) of the Children’s Hospital No. 2, Ho Chi Minh City, Viet Nam (IRB. No. 893/QD-BVND2, signed on 06-June-2022).
Informed consent statement: We used a secondary dataset from primary research, which was considered to cause less than minimal risk to the participants. Therefore, the need for ethical approval was waived. This study was performed in accordance with the principles of Good Clinical Practice and ethical guidelines of the Declaration of Helsinki.
Conflict-of-interest statement: All authors declare that there is no conflict of interest.
Data sharing statement: The original contributions of this study are included in the article and Supplementary material. Further requests can be directed to the corresponding authors.
Open-Access: This article is an open-access article that was selected by an in-house editor and fully peer-reviewed by external reviewers. It is distributed in accordance with the Creative Commons Attribution NonCommercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited and the use is non-commercial. See: https://creativecommons.org/Licenses/by-nc/4.0/
Corresponding author: Thanh Tat Nguyen, MD, PhD, Senior Researcher, Department of Tuberculosis and Epidemiology, Woolcock Institute of Medical Research, Pham Ngoc Thach Street, Ho Chi Minh City 700000, Viet Nam. thanhhonor@gmail.com
Received: September 29, 2024
Revised: November 3, 2024
Accepted: November 19, 2024
Published online: September 20, 2025
Processing time: 159 Days and 7.6 Hours
Core Tip

Core Tip: The in-hospital mortality rate of children with dengue shock syndrome (DSS) at a large tertiary pediatric hospital in Viet Nam was 3%. The supervised models showed good predictive value. In particular, the Random Forest and Extreme Gradient Boost models demonstrated the highest model performance. The supervised machine learning model showed that the nine most important predictive variables included younger age, presence of underlying diseases, severe transaminitis, critical bleeding, platelet transfusion requirement, elevated international normalized ratio and blood lactate levels, and high vasoactive inotropic score (> 30). Identification of mortality predictors in patients with DSS will help optimize management protocols to enhance survival outcomes.