THE EFFECT OF SLEEP DURATION ON FALL RISK IN MIDDLE-AGED AND OLDER CHINESE ADULTS: DEPRESSIVE SYMPTOMS AS A MEDIATOR

Abstract

Objective: Standard mediation analysis approaches are not applicable in the assessment of direct and indirect effects with timevarying mediators. We aimed to estimate longitudinal effects of sleep duration on fall risks among middle-aged and older Chinese population, when considering depression as a time-varying mediator. Methods: Data were drawn from the nationally representative survey of the China Health and Retirement Longitudinal Study. We followed the participants who were 45 years and older from the baseline in 2011 to the follow-up in 2015. Sleep duration and fall were measured by self-reported interviews. Depressive symptoms were measured by the Centre for Epidemiological Studies Depression scale. We used the parametric mediation G-formula to estimate the interventional analogues of natural direct and indirect effects of sleep duration on fall risks. Results: At the end of follow-up, 18.4% of the 12,209 participants reported falling. Compared with sleep for 7 hours, the total effect in the intervention scenario of 5 hours sleeping per night was estimated to increase fall risk (Relative risk (RR) = 2.27, 95% CI: 2.12, 2.46). The direct effect and the indirect effect were estimated to increase fall risk by 97% (RR = 1.97, 95% CI: 1.82, 2.16) and 15% (RR = 1.15, 95% CI: 1.14, 1.16), respectively. Conclusion: Findings provide novel evidence for a mediating role of depression in pathways linking sleep duration to falls. Our study presents the first application of parametric mediation gformula in an epidemiologic study on fall risk. Multiple mediators require further study in order to fully explain the effect of sleep duration on fall risk among middle-aged and older adults.