AUTHOR=Zhang Min , Rong Jian , Liu Song , Zhang Beibei , Zhao Yaodong , Wang Haibo , Ding Hong 

TITLE=Factors related to self-rated health of older adults in rural China: A study based on decision tree and logistic regression model

JOURNAL=Frontiers in Public Health

VOLUME=Volume 10 - 2022

YEAR=2022

URL=https://www.frontiersin.org/journals/public-health/articles/10.3389/fpubh.2022.952714

DOI=10.3389/fpubh.2022.952714

ISSN=2296-2565

ABSTRACT=Objective: The aim of this study was to explore the related factors of self-rated health by using decision tree and logistic regression model among the older adults in rural China.
Methods: Convenience sampling was employed to enrolled 1223 respondents who met the inclusion criteria from 10 randomly selected villages in M County in China. The content of the questionnaire covered the demographic characteristics, physical and mental health, sleep status, and risk of falling. The Pittsburgh Sleep Quality Index (PSQI) and the Morse Falls Risk Scale (MFS) were used to evaluate sleep status and risk of falling respectively. Decision tree and logistic regression model were employed to analyze the related factors of self-rated health.
Results: 817 (68.7%) of the subjects had a good self-rated health. Logistic regression model showed that living standard, alcohol consumption, sleep quality, labor, hospitalization, discomfort, a number of chronic diseases and mental health were associated with self-rated health (P-value <0.05); decision tree model showed that a number of chronic diseases, sleep quality, mental health, hospitalization, gender and drinking were associated with self-rated health. The sensitivity and specificity of the logistic regression model was 67.7% and 75.5% respectively, and the area under the ROC curve was 0.789 (0.763, 0.816); the sensitivity and specificity of the decision tree model was 71.5%, and 61.4% respectively, and the area under the ROC curve was 0.733 (0.703, 0.763).
Conclusions: Decision tree and logistic regression model complement each other, and can describe the factors related to self-rated health of the rural elderly from different aspects. Our findings indicated that mental health, hospitalization, drinking, sleep quality. etc were the important associated factors.