PDF
Citation
Print
INTRODUCTION
Insomnia is a disorder that affects the quality, quantity, and timing of sleep.1 Sleep is crucial for maintaining both physical health and mental well-being,2 and for maintaining quality of life.3 However, many individuals do not obtain sufficient quantity or quality of sleep. In the United States,4 Europe,5 and Asia,67 adults sleep less than the seven hours recommended for maintaining optimal level of physical and mental health; and more than one-third of the population experiences insomnia.8 In 2019, approximately 10% of the general Korean population suffered insomnia.9 Insomnia can lead to hypertension,10 cardiovascular diseases,11 diabetes,12 and loss of cognitive function and memory consolidation.1314
Insomnia among workers specifically deserves much attention because approximately 18% of workers in Europe15 and 23% in the United States16 report experiencing insomnia. In Korea, the prevalence of insomnia was reported to be 2.5% among male workers and 2.3% among female workers in 2011.17 Additionally, approximately 10%, 32%, and 8–26% of employees who work in day, night, and rotating shifts, respectively, experience insomnia.18 Insomnia is consistently associated with occupational injuries18 and may lead to increased accident risk, elevate absenteeism in the workplace, reduce work productivity, and inhibit career progression.19 Among workers, it is important in terms of their individual health problems, economic hardship, and poor company productivity. To prevent these problems, examining various risk factors for insomnia among workers is pivotal.
Risk factors for insomnia generally include psychosocial factors, concurrent medical conditions, and other variables (such as excessive use of alcohol and caffeine use and prescribed drugs).20 Psychosocial factors are work-related risk variables highlighted in various studies on insomnia.2122 For example, psychosocial factors related to work include job stress,2324 poor interpersonal relationships,20 overcommitment,25 low job control,202627 conflict and ambiguity in the role,28 high job demands,2027 poor social support at work,202529 and exposure to threats and violence at work.30 Recently, workplace-related factors, such as workplace bullying31 and violence at work,30 have emerged as workplace issues that are strongly associated with insomnia.22
‘Low social support,’ which is a determinant of mental health problems,32 critically leads to negative workplace outcomes including increased stress, burnout, and higher turnover rates globally.3334 The lack of social support in Korea is a powerful predictor of depression, more so than job demands or job control,35 due to the unique workplace culture. The impact of low social support in the workplace is especially pronounced in Korea’s collectivist culture, which emphasizes community and hierarchical relationships.20 These findings highlight the need for supportive work environments.
According to high association between low social support at work and insomnia, previous studies have examined the relationship between social support at work and insomnia.192022 For instance, Park et al.22 found that poor psychosocial working conditions, including experiences of sexual harassment, discrimination, workplace violence, poor work–life balance, low job satisfaction, and high job demands, were associated with an increased risk of work-related insomnia among Korean workers. However, the study used data from the first Korean Working Conditions Survey (KWCS), conducted in 2006, which is outdated.
Despite some existing studies on sex-based differences in the relationship between social support at work and physical health,93637 research on the association between insomnia and social support at work using sex-stratified analyses is lacking. Additionally, although social support at work has been indicated as a crucial factor in predicting job satisfaction38 and the relationship between job satisfaction and insomnia has been explored,3940 only a few studies have analyzed the interaction between social support, job satisfaction, and their combined impact on insomnia.
Hence, our study focused specifically on the relationship between social support at work and insomnia in the current sociocultural and occupational environment, using a nationally representative sample of Korean workers. We aimed to identify high-risk groups and examined the interaction effects of social support and job satisfaction on sleep disturbance. Additionally, we explored the mediating role of job satisfaction in this relationship, considering sex differences to provide a comprehensive understanding.
METHODS
Data collection and study design
This cross-sectional study utilized data from the 6th KWCS, a nationally representative survey of South Korean workers conducted by the Occupational Safety and Health Research Institute (OSHRI). The KWCS gathers data on sociodemographic characteristics, working conditions, and health of workers aged 15 years and older, using a multistage systemic sampling technique. The sixth KWCS, conducted between October 2020 and April 2021, included approximately 50,538 survey respondents who participated in one-on-one, in-person interviews conducted by professional interviewers from OSHRI. All respondents provided permission for the use of de-identified data in scientific studies. The KWCS has been shown to be valid and reliable, and its quality is guaranteed through the use of an organized questionnaire and a thoughtfully designed random sampling procedure.
A visual representation of the study sample selection process is shown in Fig. 1. First, 31,144 participants were excluded based on the following criteria: those aged < 19 years (n = 153) or ≥ 65 years (n = 8,404), those who were unemployed (n = 12,810), and those with missing values (n = 9,777). Finally, data from 19,394 participants (9,046 males and 10,348 females) were analyzed.
Outcomes and independent variables
The definitions of social support at work and insomnia used in our study are shown in Supplementary Table 1. Social support at work was assessed as an independent variable based on eight questions derived from the KWCS.414243 Each question was answered by participants using a five-point Likert scale, with scores on the scale reversed. The social support measurement in this study draws upon House’s Social Support Theory and Karasek & Theorell’s Job Demand-Control-Support (JDCS) Model.44 House’s theory conceptualizes social support as encompassing emotional, instrumental, and informational dimensions, all of which reduce workplace stress and improve psychological well-being. Items 1 and 2 evaluate direct emotional and instrumental support from colleagues and supervisors. In contrast, items 3 to 8 reflect organizational and supervisory behaviors, such as recognition, respect, and feedback. These organizational and supervisory behaviors coincide with the JDCS model’s emphasis on the stress-relieving effect of social support in high-demand, low-control environments. Regarding these items’ reliability, Cronbach’s α was 0.793 in a previous study,42 indicating a high level of internal consistency.
Based on previous study defined the mean score of 8 items as a social support variable,41 we calculated the sum score of 8 items, with higher scores indicating low social support at work. A cut-off score of 18 was used for the definition of social support at work, based on the average score of participants. The score of ≥ 18 in this study refers to strong social support at work, while the score of < 18 refers to poor social support at work.
Insomnia, the outcome of this study, was measured with the Minimal Insomnia Symptom Scale. The presence of insomnia was ascertained using three questions. The existence of insomnia was defined by a minimum insomnia severity score of ≥ 6. A cut-off score of 6 for the insomnia questions was sufficient to distinguish between those with and without insomnia, with a sensitivity of 0.82 and specificity of 0.86 when compared with the International Classification of Disease research criteria for insomnia in the general population aged 20–64 years.45
Other covariates
Potential confounders and covariates included age and sex as sociodemographic characteristics; education level and average monthly income as socioeconomic characteristics; and job satisfaction, working hours per week or month, quick return, and shift work as occupational characteristics. Age was categorized into 10-year increments: 20–29, 30–39, 40–49, 50–59, and 60–64 years. Income was classified into four quartiles (< 200, < 300, < 400, and ≥ 400) in millions of South Korean Won, based on average monthly income.18 Educational level was classified into three groups: middle school graduation or lower, high school graduation, and college graduation or higher. Job satisfaction was categorized as never satisfied, low satisfied, satisfied, and very satisfied (using the question “Are you generally satisfied with your work environment?”).46 Working hours were categorized into < 40 hours, 40–52 hours, and > 52 hours per week, considering the 40-hour legislative workweek and additional 12-hour overtime according to the Korean Labor Standards Act.47 Using the question, “Was there at least once last month that the interval from the end of work to the next day was less than 11 hours?” individuals were divided into quick and non-quick return groups. The question “Do you work shifts?” was used to separate individuals into shift and non-shift work groups.
Statistical analysis
All analyses were conducted by sex stratification. The baseline characteristics of the study population were compared using χ2 tests for categorical variables and t-tests for continuous variables. The 95% confidence intervals (CIs) and crude and adjusted odds ratios (ORs) for insomnia were estimated using univariate and multivariate logistic regression models. Furthermore, the status of social support at work was divided into quartiles based on the overall scores and named as lowest support (< 19), low-middle support (< 22), high-middle support (< 24), and highest support (≥ 24). The dose–response relationship between social support at work and insomnia was further examined. Moreover, interaction analyses of social support at work and job satisfaction with insomnia were conducted. The relative excess risk due to interaction (RERI),48 a common measure of additive interactions, was calculated using the delta method to analyze this interaction. The proportion of the mediating effect of job satisfaction on the relationship between social support at work and insomnia was estimated by causal mediation analysis using the “mediation” package in R (R Foundation for Statistical Computing, Vienna, Austria). Standard error was estimated using the robust sandwich method. When we analyzed the mediating effect of job satisfaction on this relationship, respondents who indicated “very satisfied” or “satisfied” were categorized as the satisfied group, and those who indicated “never satisfied” or “poorly satisfied” were categorized as the non-satisfied group.
For all analyses in all two-tailed statistical tests, a value of P < 0.050 was considered statistically significant. R software (version 4.2.1; R Foundation for Statistical Computing, Vienna, Austria) was used for all statistical analyses.
Ethics statement
The study protocol was approved by the Institutional Review Board (IRB) of Severance Hospital (IRB: 4-2021-1046) and followed the ethical guidelines of the 2013 Declaration of Helsinki. This study was retrospective in nature, and the requirement for obtaining informed consent was waived.
RESULTS
Table 1 presents the basic occupational characteristics of male and female workers based on their social support at work. Among the 19,394 workers (9,046 males and 10,348 females), 1,490 (16.5%) male and 1,678 (16.2%) female workers had low social support at work. Low social support at work was higher among male workers who were older, had insomnia, lower education levels, income levels, job satisfaction, and more quick returns, and shift work. Female workers who had low social support at work demonstrated similar features as their male counterparts.
Table 1
Baseline characteristics of participants by social support at work
Table 2 summarizes the basic occupational characteristics of male and female workers who experienced insomnia (524 [5.79%] male and 867 [8.38%] female). The prevalence of insomnia was higher among male workers who were older, were in the low social support group, had lower education levels, worked longer hours per week, had lower job satisfaction, had more quick returns, and performed shift work. Female workers who had insomnia showed similar features.
Table 2
Baseline characteristics of participants by insomnia
Table 3 presents the estimated ORs and 95% CIs in the final model, revealing the relationship between low social support at work and insomnia by using a multivariable logistic regression analysis. The fully adjusted OR (95% CI) for insomnia with social support at work was 1.47 (1.29–1.69; P < 0.001). Among male and female workers, the fully adjusted OR (95% CI) of insomnia was 1.71 (1.39–2.12; P < 0.001) and 1.34 (1.12–1.60; P < 0.001), respectively.
Table 3
Adjusted odds ratio of insomnia according to social support at work in the logistic model
Values are presented as odds ratio (95% confidence interval).
Model 1 was adjusted for sociodemographic characteristics, such as age, sex (only both workers odds ratio analyses). Based on model 1, model 2 was further adjusted for socioeconomic characteristics, such as education level, income. Based on model 2, model 3 was further adjusted for occupational characteristics, such as job satisfaction, working hours per week or month, quick return, shift work.
Bold values indicate statistical significance (P < 0.05).
According to the dose–response analysis, low social support at work exhibits a J-shaped relationship with insomnia in male workers while demonstrating a linear association in female workers (Fig. 2). Consistent with the main analysis results, the adjusted OR (95% CI) of insomnia for the lowest social support group (> 19), compared with the highest social support group (≥ 24), was 1.64 (1.30–2.07; P < 0.001) among male workers, and 1.61 (1.33–1.94; P < 0.001) among female workers. Moreover, a notable increase in insomnia among male workers was observed in the highest social support quartile, whereas a gradual and consistent increase in the probability of insomnia was observed with higher levels of social support among female workers.
Fig. 2
Dose–response relationship of social support at work and insomnia by sex: (A) male, (B) female.
OR = odds ratio, CI = confidence interval.
Table 4 shows the significant joint effects of simultaneous social support at work and job satisfaction on insomnia, regardless of sex. We found an additive interaction between social support at work and job satisfaction on insomnia in both male and female workers (RERI [95% CI], 1.91 [0.47–3.35]; P = 0.005 and 1.59 [0.69–2.49], P < 0.001; respectively). Moreover, the mediation analysis showed that the proportion of mediation of job satisfaction was 30.3% among both male and female workers in the relationship between social support at work and insomnia.
Table 4
Interaction between job satisfaction and social support at work on insomnia and mediation of job satisfaction on the relationship between social support at work and insomnia among male and female workers
DISCUSSION
Our comprehensive analysis of the nationally representative data on Korean workers highlighted that having low social support at work is significantly associated with the presence of insomnia in both sexes. Our findings remained consistent even after adjusting for covariates such as age, education level, income, working hours, shift work, quick returns, and job satisfaction. We also noted dose–response relationships between social support at work and insomnia, which varied by sex. Insomnia appeared to have joint effects on social support at work and job satisfaction among both male and female workers. Moreover, job satisfaction mediated the effect of social support at work on insomnia by approximately 30%. These findings are meaningful in determining the multifaceted characteristics and high-risk groups for insomnia related to the workplace environment.
Our findings align with previous studies using Korean data. Kim et al.20 demonstrated that the relationship between inadequate social support and insomnia was significant (OR, 1.30; 95% CI, 1.14–1.47), reflecting our result. These studies highlight the vital importance of social support in Korea workplaces, where hierarchical and collectivist structures exacerbate the impact of low social support on insomnia. Lifetime employment in Korea heightens interpersonal pressure, making conflicts harder to resolve, thus making social support a more crucial factor for insomnia than in western cultures.20
Additionally, this study’s results were consistent with previous studies from other countries. Gosling et al.49 investigated the influence of health status and social support on the relationship between job stress and insomnia and found that job stress had limited independent effects on insomnia, while the level of social support played a more prominent role in predicting the presence or absence of insomnia. Åkerstedt et al.50 suggested that individuals with high levels of social support at work had a lower risk of insomnia. Consequently, the relationship between social support at work and insomnia was more significant than the associations of other work-related factors with insomnia.
Åkerstedt et al.50 highlighted the close association between positive psychosocial resources and the restoration of sleep health, which is potentially explained by physiological mechanisms. Positive psychological indicators are linked to mitigating hormonal changes and positively affect sleep quality.51 Social support at work is associated with healthier hypothalamic–pituitary–adrenal axis functioning, which is known to greatly affect sleep patterns and emotions, and has been shown to be inversely related to long-term body mass index increases.52 These findings emphasize the global prevalence of insomnia and establish the significance of social support at work.
According to the dose–response relationship, an increase in the prevalence of insomnia was observed with decreasing social support at work, regardless of sex. However, sex-based differences in the steepness of the curve in the dose–response relationship were noted. Male workers demonstrated no significant impact of social support at work on insomnia except for the category with the lowest level of social support, where a notable increase in insomnia was observed. Contrastingly, for female workers, a linear dose–response relationship was observed between weak social support at work and insomnia. This might illustrate that social support at work is highly related to stress,53 as women are generally more affected by psychological health problems across different countries and situations.54 Our findings may be related to sleep reactivity, a trait, in which exposure to stress makes it difficult to fall or stay asleep.55 Notably, women exhibit higher levels of sleep reactivity than men, as reported by Kalmbach et al.55; this trait is associated with work-related factors. Male workers only showed a significant relationship between the lowest social support at work and insomnia; this might be explained by higher psychological resilience in male workers than in female workers.56 Consequently, different approaches need to be applied to manage social support related to sex-based insomnia.
This study explored the mediating and joint effects of job satisfaction on the relationship between social support and insomnia. We emphasized that job satisfaction is important in the interaction and causal mediation analyses of the relationship between social support at work and insomnia. According to previous studies, job satisfaction is highly associated with social support in the workplace,57 and also is related to insomnia.3940 The combination of low social support and low job satisfaction significantly increases the risk of insomnia, amplifying negative effects on mental and physical health. Interventions targeting both social support and job satisfaction in the workplace could synergistically improve sleep health and overall well-being for employees. Furthermore, the proportion of the relationship between social support at work and insomnia mediated by job satisfaction was considerably higher (both sexes, 30.3%). Previous studies have predominantly focused on analyzing associations independently rather than exploring their combined effects. A study conducted by Ye et al.58 indicated that job satisfaction was significantly and negatively associated with insomnia. Moreover, other studies have highlighted that social support at work is related to job satisfaction.5960 Overall, we suggest that job satisfaction and social support at work are crucial factors in managing insomnia. Therefore, developing guidelines focusing on both job satisfaction and social support at work to enhance workers’ overall sleep should be a priority in the future.
Our study had several limitations. Firstly, the cross-sectional design prevents establishing temporal associations among variables, necessitating longitudinal analyses to ascertain causality. Secondly, self-reported questionnaires may result in recall bias, warranting validation through objective sleep quality measures like polysomnography. Thirdly, generalization beyond Korean workers should be approached cautiously, considering potential cultural and contextual differences. Moreover, exploration of the relationship among self-employed individuals and contributing family workers remains necessary. Lastly, focusing on general insomnia using limited questions may overlook specific types of sleep issues, highlighting the need for studies examining diverse insomnia and their association with social support at work.
Our findings make a meaningful contribution to understanding the complex association between workplace support at work and insomnia among Korean workers, using data from the KWCS, a representative dataset. We employed advanced statistical methods to investigate whether job satisfaction mediates the relationship between social support at work and insomnia and explored potential additive interactions. Furthermore, conducting a sex-stratified analysis enhances comprehension of these relationships, considering variations between male and female participants. These findings will provide a basis for improving the sleep health of workers.
In conclusion, our study highlights the significant association between social support at work and insomnia, stratified by sex, among Korean workers. Policymakers should consider social support programs for the effective exploration of the specific characteristics of each sex and the role of job satisfaction as a mediator in a healthier work environment, ultimately improving overall sleep health among workers.
XML Download