Identification of associative factors for recurrent respiratory tract infection using machine learning models in preschoolers

Authors: Li, Che; Wang, Qiong; Wang, Xiaoqian; Zhang, Wenqian; Zheng, Fangjieyi; Zhang, Zhixin; Niu, Wenquan

Source journal: Pediatric research (2026-11-16)

DOI: 10.1038/s41390-026-05565-5

Abstract

OBJECTIVES: Childhood recurrent respiratory tract infection (RRTI) poses a significant global public health burden. This study aimed to identify factors associated with RRTI in preschool children using machine learning (ML) and to develop a web-based tool for estimating individualized RRTI probabilities. METHODS: Two cross-sectional surveys using stratified cluster sampling were conducted in 2020 and 2025. RRTI was diagnosed according to the national criteria issued by the Chinese Medical Association and China Association of Chinese Medicine. Questionnaire data covering demographic characteristics, fetal/neonatal history, and lifestyle factors were collected. RESULTS: Total 15,075 eligible records from two surveys were included, reporting an RRTI prevalence of 14.55%. Ten ML models were trained on 29 candidate variables. The eXtreme Gradient Boosting (XGBoost) model outperformed the others, yielding an area under the curve of 0.733. SHapley Additive exPlanations analysis identified sleep duration, child's age, and screen duration as the top three associative variables for childhood RRTI. Forward feature selection was applied to derive a parsimonious nine-variable predictive set, which was subsequently deployed as an online application ( https://cipchina.streamlit.app/ ). CONCLUSIONS: In a large, population-based cohort of preschool children, three factors (sleep duration, child's age, and screen duration) associated with RRTI were consistently identified across multiple high-performing ML models. IMPACT: This study identified key lifestyle-related factors associated with recurrent respiratory tract infection in preschool children. Ten machine learning models were applied to analyze 29 variables from 15,075 Chinese preschoolers, with further importance interpreted via the SHapley Additive exPlanations. Sleep duration emerged as the primary lifestyle-related factor associated with childhood recurrent respiratory tract infection, followed by child's age and screen duration.

Read the full text

Related topics

Related open access articles

Abstract and bibliographic data from PubMed (U.S. National Library of Medicine). Copyright of the abstract belongs to the original publisher or authors.