Sleep disturbance has an enormous impact on college students. Poor sleep is associated with low academic achievement, psychological distress and high health risk behaviors. College students with various mental health problems (e.g., anxiety and depression) are particularly at risk for sleep problems. The aim of the present study was to examine the impact of a wide range of internalizing/externalizing psychological problems on sleep disturbance. A total of 2134 Chinese college students (60.2% men) with mental health problems were selected as participants after completing the self-reported Chinese college student mental health screening scale. A web-based survey was used to assess a wide variety of internalizing/externalizing psychological problems and sleep disturbance. The results showed that hostility, somatic symptoms and obsessive-compulsive symptoms (OCS) were significantly associated with sleep disturbance. Somatic symptoms played a mediating role in the relationship between hostility and sleep disturbance. Moreover, the mediating effect was moderated by OCS, and a significant difference in the mediating effects was observed between low OCS and high OCS groups. Overall, our research findings indicate that a high level of OCS exacerbates the adverse effects of somatic symptoms on sleep disturbance, and suggests that assessment and improvement of hostility, somatic symptoms and OCS should be considered in facilitating better sleep among college students with mental health problems.

Sleep disturbance is common among college students, but the statistical relationships linking obsessive-compulsive symptoms (OCS), affective symptoms, and problematic Internet use remain unclear in nonclinical freshmen. In this cross-sectional study, 9,418 college freshmen aged 18–23 years (M = 19.60, SD = 0.65; 3,660 females) completed relevant subscales of the Chinese College Students’ Mental Health Screening Scale. Covariate-adjusted conditional process analyses (analytic N = 9,416) estimated separate and serial indirect associations through anxiety and depression and first-stage moderation by Internet addiction (IA). OCS was positively associated with sleep disturbance (B = 0.284, t = 20.58, p < .001, 95% CI [0.257, 0.311]). Indirect associations were observed through anxiety (estimate = 0.263, 95% bootstrap CI [0.232, 0.293]), depression (estimate = 0.098, 95% bootstrap CI [0.084, 0.114]), and anxiety followed by depression (estimate = 0.113, 95% bootstrap CI [0.098, 0.130]). The OCS × IA interaction in the anxiety equation was statistically significant but small (B = 0.0057, SE = 0.0015, t = 3.71, p < .001). Concurrent measurement prevents conclusions about temporal ordering or causality.