Your new experience awaits. Try the new design now and help us make it even better

ORIGINAL RESEARCH article

Front. Nutr.

Sec. Nutritional Epidemiology

Volume 12 - 2025 | doi: 10.3389/fnut.2025.1650644

This article is part of the Research TopicRevolutionizing Nutritional Epidemiology: Harnessing Digital Health, AI, and Big Data for Population-Level Disease Prevention and ManagementView all articles

Dietary Intakes and Asthma-Coronary Heart Disease Comorbidity: Predictive Analytics via Machine Learning Model and Feature Explanation based on SHAP

Provisionally accepted
  • 1The First Clinical Medical College, Inner Mongolia Medical University, Hohhot, China
  • 2College of Chinese Medicine, Inner Mongolia Medical University, Hohhot, China
  • 3Emergency Medical Center, Inner Mongolia Autonomous Region People's Hospital, Hohhot, China

The final, formatted version of the article will be published soon.

Objective: The association between dietary nutrients and asthma-coronary heart disease (CHD) comorbidity remains unclear. We applied weighted quantile sum (WQS) regression model to evaluate dietary mixtures' joint associations. Machine learning models were developed to elucidate this relationship and assess dietary intake's predictive role.Methods: We analyzed 4,334 NHANES 2005-2018 participants, examining 45 dietary components and covariates. WQS assessed joint dietary associations. Six ML models were benchmarked using performance metrics, with SHAP values interpreting feature importance.Results: Among 4,334 participants, 442 had comorbidity. WQS revealed a joint negative association (estimate = -0.2968, OR = 0.74, 95%CI = 0.64-0.86, P < 0.001), with alcohol showing the highest relative association among the dietary components analyzed. Random Forest demonstrated optimal performance: AUC = 0.967, Pr-AUC = 0.985, accuracy = 0.923, F1 = 0.948, sensitivity = 0.985, specificity = 0.770. SHAP analysis highlighted key predictors: hypertension, diabetes, age, caffeine, copper, and added vitamin B12.Conclusion: WQS demonstrated a joint negative association of dietary nutrients with comorbidity. Random Forest showed superior predictive performance. SHAP analysis identified caffeine and copper as dietary factors with the strongest associations in the model.

Keywords: machine learning, Shap, Dietary intakes, Asthma, coronary heart disease

Received: 20 Jun 2025; Accepted: 04 Aug 2025.

Copyright: © 2025 Tang, Li and Lingchun. 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) or licensor 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.

* Correspondence: Xu Lingchun, Emergency Medical Center, Inner Mongolia Autonomous Region People's Hospital, Hohhot, China

Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.