Choices of medical institutions and associated factors in older patients with multimorbidity in stabilization period in China: A study based on logistic regression and decision tree model

Abstract

Background

As China's population ages, its disease spectrum is changing, and the coexistence of multiple chronic diseases has become the norm with respect to the health status of its elderly population. However, the health institution choices of older patients with multimorbidity in stabilization period remains underresearched. This study investigate the factors influencing the choices of older patients with multimorbidity to provide references for the rational allocation of healthcare resources.

Methods

A multistage, stratified, whole-group random-sampling method was used to select eligible older patients from September to December of 2022 who attended the Community Health Service Center of Guangdong Province. We adopted a self-designed questionnaire to collect patients' general, disease-related, social-support information, their intention to choose a healthcare provider. A binary logistic regression and decision tree model based on the Chi-squared automatic interaction detector algorithm were implemented to analyze the associated factors involved.

Results

A total of 998 patients in stabilization period were included in the study, of which 593 (59.42%) chose hospital and 405 (40.58%) chose primary care. Our binary logistic regression results revealed that age, sex, individual average annual income, educational level, self-reported health status, activities of daily living, alcohol consumption, family doctor contracting, and family supervision of medication or exercise were the principal factors influencing the choice of medical institutions for older patients with multimorbidity (p < 0.05). The decision-tree model reflected three levels and 11 nodes, and we screened a total of four influencing factors: activities of daily living, age, a family doctor contract, and patient sex. The data showed that the logistic regression model possessed an accuracy of 72.9% and that the decision tree model exhibited an accuracy of 68.7%. Prediction using the binary logistic regression was thus statistically superior to the categorical decision-tree model based on the Chi-squared automatic interaction detector algorithm (Z = 3.238, p = 0.001).

Conclusion

More than half of older patients with multimorbidity in stabilization period chose hospitals for healthcare. Efforts should be made to improve the quality of healthcare services and increase the medical contracting rate and recognition of family doctors so as to attract older patients with multimorbidity to primary medical institutions.