A cohort study on self-reported role stressors at work and poor sleep: does sense of coherence moderate or mediate the associations?

The aim of the present study was to examine the bidirectional associations between subjective role ambiguity and role conflicts at work, respectively, and self-reported sleep 2 years later. In addition, we also examine whether sense of coherence (SOC) moderate or mediate the association between role stressors and poor sleep and between poor sleep and role stressors.

We used questionnaire data collected in 2006 and 2008 from the Workplace Bullying and Harassment cohort. In 2006, 3363 responded to the questionnaire and in 2008 1671 responded. In total, 1569 participants responded in both 2006 and 2008 to the questions on role stressors (in terms of role ambiguity and role conflicts at work) and sleep problems in both 2006 and 2008. Sleep problems were assessed with the awakening index (AWI) and the disturbed sleep index (DSI). Moderation and mediation analyses of the association were estimated using structural equation modelling.

We found a prospective association between role stressors and sleep problems [beta values were 0.07 (95% CI 0.03-0.11) and 0.05 (CI 0.01-0.10) for DSI and AWI, respectively] when adjusting for sleep problems at baseline, age, sex, and life style factors (i.e. alcohol, smoking, and leisure time physical activity). SOC moderated the association showing that participants with lower SOC scores who reported higher role ambiguity reported sleep problems to a higher extent than participants with high SOC scores. SOC also mediated the association between role stressors and sleep problems. We also found support for sleep problems at baseline and role stressors 2 years later [DSI 0.04 (CI 0.00-0.08) and 0.15 (CI 0.09-0.21)] for role ambiguity and role conflicts, respectively. Similar results were observed for AWI.

Subjective role stressors were prospectively associated with sleep problems. Yet, sleep problems could also prospectively predict subjective role stressors (i.e. reverse causation). The analyses also showed that SOC may be regarded as both a mediating and a moderating factor of the association between subjective role conflicts and poor sleep. We found that SOC moderated the prospective association so participants with low SOC report more sleep problems with subjective role conflicts compared to participants with high SOC. Finally, we also found SOC mediated the prospective association between subjective role stressors and sleep problems and the reverse association.