Identifying potential action points for improving sleep and mental health among employees: A network analysis

Mental health issues are severe public health problems, inevitably affected by, also affecting, sleep. We used network analysis to estimate the relationship among various aspects of sleep and mental health simultaneously, and identify potential action points for improving sleep and mental health among employees. We used data from the baseline survey of the Chinese Cohort of Working Adults that recruited 31,105 employees between October 1st and December 31st, 2021. The mental health included anxiety (measured by the Generalized Anxiety Disorder-7), depression (Patient Health Questionnaire-9]), loneliness (Short Loneliness Scale), well-being (Short Scales of Flourishing and Positive and Negative Feelings), and implicit health attitude (Lay Theory of Health Measures). Seven dimensions of sleep were assessed by the Pittsburgh Sleep Quality Index. An undirected network model and two directed network approaches, including Bayesian Directed Acyclic Graphs (DAGs) and Evidence Synthesis for Constructing-DAGs (ESC-DAGs), were applied to investigate associations between variables and identify key variables. Depression, daytime dysfunction, and well-being were the "bridges" connecting the domains of sleep and mental health in the undirected network, and were in the main pathway connecting most variables in the Bayesian DAG. Anxiety constituted a gateway that activated other sleep and mental health variables, with sleep duration and implicit health attitude forming end points of the pathway. Similar directed pathways were confirmed in the ESC-DAG. Our network study suggests anxiety, depression, well-being, and daytime dysfunction may be potential action points in preventing the development of poor sleep and mental health outcomes for employees.