Estimating the longitudinal causal effect of insufficient sleep on depressive symptoms: Evidence from the CHARLS study

Abstract

Background: The relationship between sleep and depression has been extensively explored; however, research on the long-term causal effects of chronic sleep insufficiency on depressive symptoms remains limited. This study employs advanced causal inference techniques to assess the longitudinal impact of sleep insufficiency on depressive symptoms, accounting for both time-invariant and time-varying confounders.

Methods: Data from the China Health and Retirement Longitudinal Study (CHARLS) were analyzed. Sleep insufficiency was defined as <6 h of sleep per day. Longitudinal targeted maximum likelihood estimation (LTMLE) was used to examine the impact of persistent sleep insufficiency over 2 to 9 years on depressive symptoms, measured by the CESD-10 scale. Subgroup analyses by gender and age, as well as the effects of napping duration, were also conducted.

Results: The study included 4362 participants, with a mean age of 55.90 years (SD = 7.71). At baseline, 924 participants reported insufficient sleep. The average treatment effect (ATE) for depressive symptoms increased initially, peaking at 3.75 points at 7 years, then slightly declining to 3.66 points at 9 years. The ATE was higher in females (4.57) than in males (2.80). Napping for over 30 min was associated with reduced depressive symptoms. Sensitivity analyses confirmed the robustness of these findings.

Conclusions: This study provides evidence of the longitudinal causal effect of insufficient sleep on depressive symptoms. Over 9 years, ATE initially increased, plateauing after 7 years. Napping for over 30 min was linked to lower depressive symptoms, especially in those with insufficient nighttime sleep. Subgroup analyses showed stronger effects in females.

Keywords: Depressive symptoms; Insufficient sleep; Longitudinal study; Targeted maximum likelihood estimation.