Abstract
<title>Abstract</title> <p> Background Adult psychological burden is often linked to social support, sleep problems, and addictive behaviors, but these factors are usually examined separately. It remains unclear whether sleep problems account for symptom burden beyond addictive behaviors, and whether repeated sleep-related hours are associated with later perceived stress. Methods Study 1 included 1,223 adults from a cross-sectional online survey. Parallel bootstrap mediation models with 5,000 resamples examined whether subjective sleep problems, alcohol-related problems, nicotine dependence, and internet gaming symptoms accounted for the association between perceived social support and somatic-psychological symptom burden after removing the sleep item. Study 2 used longitudinal China Health and Nutrition Survey data from up to 8,557 adults. HC3-robust regression models tested whether latest, mean, and slope-based sleep-related hours from 2004 to 2011 were associated with perceived stress in 2015. Additional analyses examined nonlinearity, sleep-hour categories, measurement heterogeneity, and age-stratified patterns. Results In Study 1, greater perceived social support was associated with lower symptom burden (total effect = -0.1504, 95% CI: -0.1972 to -0.1061). The largest indirect association was through subjective sleep problems (indirect effect = -0.0761, 95% CI: -0.1021 to -0.0532), while internet gaming symptoms and alcohol-related problems showed smaller indirect associations. Nicotine dependence did not show a stable pathway. In Study 2, longer latest pre-2015 sleep-related hours were associated with higher perceived stress in the fully adjusted model ( <italic>B</italic> = 0.193, 95% CI: 0.080 to 0.306). A more positive sleep-related hours slope showed a similar association ( <italic>B</italic> = 0.853, 95% CI: 0.045 to 1.662). Quadratic models did not indicate a U-shaped pattern; higher stress was mainly observed among participants reporting ≥ 9 h. Conclusions Sleep-related indicators were associated with adult psychological burden across both data sources. Subjective sleep problems were the main pathway linking lower perceived social support to greater symptom burden, while longer and increasing sleep-related hours were associated with later perceived stress. These findings support sleep-related profiling, but require cautious interpretation because sleep was self-reported, wording varied across CHNS waves, and the observational design does not support causal inference. Trial registration Not applicable. </p>