Identifying key determinants of health among China's migrant population using machine learning methods: Evidence from the china migrants dynamic survey.
Where this comes from
- Record sourced from PubMed, PMID 41212830.
- Also identified by DOI 10.1371/journal.pone.0335168 and PMC identifier 12599968.
- No licence information is recorded for this record.
- Because redistribution is not established, this page shows the abstract only. Follow the links below for the full text.
Abstract
Continuously improving health security for the migrant population is a key component of China's healthcare system reform. Existing research indicates that migrant health is influenced by multiple factors, yet the relative importance of these factors remains inadequately measured. This study aims to analyze the current health status of China's migrant population and rank the primary factors influencing their health based on importance. Data were sourced from the 2018 China Migrants Dynamic Survey, including 108,669 cases after data cleaning. The health status of the migrant population was initially analyzed using frequency and percentage distributions. Logistic regression was then applied to examine the relationship between various factors and migrant health. Subsequently, six machine learning methods (Neural Network, Random Forest, Support Vector Machine, Gradient Boosting Machine, Extra Trees, and Decision Tree) were applied to rank the importance of these factors. A multidimensional performance metric system (accuracy, precision, recall, F1 score, and AUC value) was employed to comprehensively evaluate the classification performance of the models. SHAP (Shapley Additive Prediction) values were used to illustrate the contribution of different factors to the health status of the migrant population. The health status of China's migrant population is generally positive, though it is influenced by multiple factors, with varying degrees of significance. Among six distinct machine learning models, the Random Forest model demonstrated the best predictive performance. Its results indicate that the key factors affecting migrant health are age, employment, income, and education level. SHAP value analysis reveals that stable employment, higher education levels, and higher income are positively correlated with better health outcomes, while age was predominantly negatively correlated, indicating a detrimental effect on health status. The overall health status of China's migrant population is relatively optimistic. However, their disadvantaged positions in areas such as education and income expose them to higher health risks. To address these key determinants, further improvements in health safeguards should focus on: developing stratified intervention strategies based on age structure differences; optimizing work environments and employment security; enhancing health literacy; and strengthening public health emergency management and social support systems.
Medical subject headings
- Transients and Migrants
- Machine Learning
- Health Status