Optimal 24-Hour Sleep Duration and Depression Risk in Chinese Middle-Aged and Elderly: A Machine Learning and RCS Analysis of Nonlinear Associations.

OBJECTIVES: To investigate the relationship between sleep (nocturnal sleep and total 24-hour sleep) and depressive symptoms in middle-aged and elderly people using data from the China Health and Retirement Longitudinal Study (CHARLS). STUDY DESIGN: Cross-sectional study. METHODS: A total of 13,718 middle-aged and elderly people aged 45 and above from the China Health and Retirement Longitudinal Study (CHARLS 2020) were included. A multivariate logistic regression model was used to analyze the correlation between sleep and depressive symptoms, and a random forest model was used to rank the importance of variables. Furthermore, the nonlinear relationship between sleep duration and depressive symptoms risk was fitted using a restricted cubic spline (RCS) model. RESULTS: Compared with the group with 24-hour sleep duration of > 7 to ≤ 9 h, the adjusted odds ratio (OR) for depressive symptoms in the group with 24-hour sleep duration of ≤ 7 h (OR = 1.649) and > 9 h (OR = 1.166) was higher. Random forest results showed that total 24-hour sleep ranked above average among variables affecting depressive symptoms. The RCS curve showed that a total 24-hour sleep time of > 7 to ≤ 9 h was associated with the lowest adjusted odds for depressive symptoms. CONCLUSIONS: There is a significant correlation between sleep duration and depressive symptoms in middle-aged and elderly people. Those who slept > 7 to ≤ 9 h per night had a lower incidence of depressive symptoms, and total 24-hour sleep of > 7 to ≤ 9 h showed a stronger OR value. Life satisfaction also emerged as an important factor influencing depressive symptoms.