Influencing Factors of ADL Disability in Middle-aged and Elderly Based on Bayesian Network Model: A Study Based on the China Health and Retirement Longitudinal Study

Aim While Bayesian networks model offer a promising approach for discussing factors associated with many diseases, little attention has been poured into (ADL) disability using Bayesian network models. This study aimed to explore the complex network relationships between ADL disability and its related factors, and to achieve the Bayesian reasoning for ADL disability, thereby, offering a scientific reference for the prevention and treatment of ADL disability. Methods The data was downloaded from the Online Open Database of China Health and Retirement Longitudinal Study (CHARLS) 2020, a population-based longitudinal survey. Univariate analyses and Logistic regression analyses were used for preliminary screening of variables, “bnlearn” Bayesian network software package was used for model construction, and Netica software was used for model inference. Results Among12,656 individuals (9,313 men and 10,439 women), there are 3,060 participants with ADL disability (24.2%). The variables were screened by Logistic regression analysis model and included in the Bayesian network model. The Bayesian network were constructed with 12 nodes and 22 directed edges. The results showed that age, somatic pain and depressive symptoms are directly associated with ADL disability. Besides, cognitive function indirectly associated with ADL disability through depressive symptoms. Conclusions The Bayesian network model reveals the direct and indirect factors and correlation strength of ADL disability in middle-aged and elderly people, clarifies the complex network relationship between factors, and provides a scientific basis for early prevention of ADL disability in middle-aged and elderly people.