Association between insomnia and job stress: a meta-analysis

Purpose Insomnia has become one of the foremost health concerns among workers. Despite a significant number of epidemiological studies have reported on the correlation between insomnia and job stress, comprehensive evidence remains insufficient. Therefore, this research seeks to provide evidence with greater reliability, through summarizing relevant contemporary literature via a meta-analysis. Methods Literature from across Europe and Asia that was of both a prospective and cross-sectional design was included, if well-controlled odds ratios were available. The meta-analysis was undertaken in accordance with the guidelines devised by PRISMA, including tests for publication bias and heterogeneity. Results High job stress was associated with a greater risk of suffering from insomnia (random OR = 1.73, 95% CI 1.46-2.05), and the correlation between effort-reward imbalance and insomnia was statistically significant (random OR = 2.63, 95% CI 1.22-5.69). Higher demand was correlated to a relatively greater risk of insomnia (random OR = 1.35, 95% CI 1.20-1.51), while the pooled effect of low control was not found to be statistically significant. The summary random odds ratio of heavy workload was 2.76, and a pooled odds ratio of 1.67 (fixed, 95% CI 1.11-2.52) was calculated in low social support. With regard to the overall population, work-family conflict was correlated with insomnia (random OR = 2.32, 95% CI 1.53-3.51). The subgroup analysis provided comparable outcomes, for both males (fixed OR = 1.97, 95% CI 1.50-2.57) and females (random OR = 2.80, 95% CI 1.30-6.05). Egger's regression indicated that publication bias may be apparent in the syntheses of effort-reward imbalance, low social support, and work-family conflict (p < 0.05). Heterogeneity was caused by design, measuring the exposure or outcome, in addition to the region where the research was conducted. Conclusions The correlation between insomnia and higher levels of job stress, effort-reward imbalance, high demand, heavy workload, and low social support was determined. Publication bias and heterogeneity were partially observed. Furthermore, future studies with improved methodologies and a focus on mechanisms are anticipated.