Sleep Reactivity and Sleep Efforts in Shift Workers
Article information
Abstract
Objective
The current study aimed to investigate the differences in sleep reactivity and sleep effort differs among late night shift workers (LSWs) and non-late night shift workers (non-LSWs), and non-shift workers (non-SWs).
Methods
In total, 6,023 participants (1,613 non-SWs, 3,339 LSWs, and 1,071 non-LSWs) were recruited. Non-SWs was defined as those who works at fixed schedules during standard daylight. LSWs was defined as who work late night hours (10 PM–6 AM), while non-LSWs was SWs who did not work during late night. All completed the Ford Insomnia Response to Stress Test (FIRST), the Glasgow Sleep Effort Scale (GSES), the Pittsburgh Sleep Quality Index (PSQI), the Epworth Sleepiness Scale (ESS), the Insomnia Severity Index (ISI), and the short-term Center for Epidemiologic Studies-Depression scale (CES-D) through online survey.
Results
LSWs and non-LSWs reported higher FIRST, GSES scores than non-SWs. In addition, LSWs reported higher FIRST, GSES scores than non-LSWs. FIRST scores were correlated with CES-D, PSQI, ISI, and ESS for LSWs, non-LSWs, and non-SWs alike. GSES scores were also correlated with CES-D, PSQI, ISI, and ESS for LSWs, non-LSWs, and non-SWs alike.
Conclusion
SWs showed higher sleep reactivity and sleep effort than non-SWs. LSWs had higher sleep reactivity and sleep effort than non-LSWs, and these variables are associated with insomnia, daytime sleepiness, and depressive symptoms. Our findings suggests that late night schedule, may increase sleep reactivity and sleep effort, which are associated with sleep and mood disturbances.
INTRODUCTION
In industrialized countries, 24-hour operations are important for public safety, public health, and the economy [1]. Consequently, in many nations more than 16% of wage and salary workers are shift workers (SWs) who work outside of the standard work hours (9 AM to 5 PM) [1-3]. Individuals engaged in shift work commonly experience more sleep disturbance and excessive sleepiness than individuals who work during standard hours [4,5]. Approximately 26% of SWs develop shift work sleep disorder (SWSD) and approximately 45% of SWs suffer from insomnia. SWSD is characterized as insomnia when sleep is expected and/or having excessive sleepiness during the desired waking period accompanied by reduction of total sleep time associated with shift work [1].
Many studies have reported that SWs are exposed to more stressful events such as safety perceptions, conflicts at work, and low decision latitude than day workers [6-9]. Previous studies have found that stressful events or challenges to the sleep system, including environmental factors and/or circadian phase shifts, impair normal sleep function [10-12]. Sleep reactivity is tendency to exhibit pronounced sleep disturbance in response to sleep challenges such as stressors.
Because sleep reactivity may cause difficulty in falling and staying asleep after a stressor, higher premorbid sleep reactivity is known as one of the risk factors of developing future insomnia [13,14].
When considering the fact that SWs report higher fatigue, and more stresses related administrative/professional pressure, and physical/psychological danger than day workers, sleep reactivity may correlate with more sleep problems in SWs. However, few studies have explored the relationship between sleep reactivity and shift work.
According to the cognitive model of insomnia, insomnia is developed and maintained when dysfunctional beliefs and attitudes regarding sleep trigger excessively negative presleep cognitive activity [15]. Sleep effort is a contributor factor of these dysfunctional beliefs and attitudes regarding sleep. Sleep effort is a unique psychological factor that can exacerbate the negative presleep cognitive processes [16]. It might escalate insomnia and make it severe by distorting perceptions of sleep [17]. Also SWs with SWSD report more presleep cognitive arousal, dysfunctional beliefs, and selective attention toward worries and helplessness than SWs without SWSD [18].
Sleep disturbances in SWs can vary depending on their working shift or patterns because shift work encompasses many different times and schedules [19]. A few studies have reported that shift work schedules reflect the severity of sleep disturbance [20,21]. In particular, night shift work which is related to the circadian misalignment might induce severe sleep disturbances.
However, whether the work schedule is associated with sleep reactivity or sleep effort in SWs remains unclear. The present study explored differences in sleep reactivity and sleep effort between SWs and non-SWs and examined whether levels of sleep reactivity and/or sleep effort are associated with sleep and/or mood in SWs. We hypothesized that SWs would have higher levels of sleep reactivity and sleep effort than non-SWs and late-night SWs (LSWs) would have higher levels of sleep reactivity and sleep effort than other SWs.
METHODS
Study population
Initially, 1,254 workers (aged 32.58±7.93 years; 448 males and 806 females; 961 SWs and 293 non-SWs) were recruited through online advertisement. An additional 5,400 participants (aged 38.33±9.89 years; 2,693 males and 2,707 females; 3,600 SWs and 1,800 non-SWs) were recruited through an online survey company (Macromill Embrain Co. Ltd., Seoul, Korea) to ensure the study population included more males, middle-aged workers, and non-SWs. The inclusion criteria were at least 18 years of age and having full- or part-time employment. The exclusion criterion was the inability to complete the online surveys.
In total, 6,665 participants (aged 37.00±9.82 years; 3,149 males and 3,516 females) were recruited. Subjects with work schedules that were difficult to classify were excluded (n=11), as were those with inconsistent schedules: identified themselves as daytime non-SWs but their work schedule during 10 PM to 6 AM (n=480) and identified themselves as late-night workers but not working during 10 PM to 6 AM (n=151). Finally, 6,023 participants (aged 37.18±9.83 years; 2,772 males and 3,251 females) were included in statistical analyses. No significant differences were observed in age or sex between the 642 excluded and the 6,023 included subjects. All participants provided written informed consent. All procedures were conducted in accordance with the ethical standards of our national and institutional committees on human experimentation and the Helsinki Declaration of 1964 (as revised in 2013). The Institutional Review Board of Samsung Medical Center approved the study protocol (approval no. 2019-04-095).
Work schedules
Work patterns were operationally defined based on previous categorization [22-24]. Because working time is a main determinant for sleep disturbances in SWs [19], shift work patterns were divided into two types depending on whether the working time included late night hours (approximately 10 PM to 6 AM):
1) Non-SWs: subjects who work a fixed schedule during standard daylight time (7 AM to 6 PM; 1,613 participants, aged 37.78±9.81 years; 685 males and 928 females).
2) LSWs: SWs who work late night hours (10 PM to 6 AM; 3,339 participants, aged 36.30±9.45 years; 1,688 males and 1,651 females).
3) Non-LSWs: SWs who work from 7 AM to 10 PM (1,071 participants; aged 39.04±10.67 years; 399 males and 672 females).
Measures
The Ford Insomnia Response to Stress Test (FIRST) is a self-reported questionnaire assessing sleep reactivity [24]. It consists of 9 items examining the likelihood that subjects would experience sleep disturbance in response to 9 hypothetical stressful situations. The Glasgow Sleep Effort Scale (GSES) measures the core components of an overall model on persistent preoccupation with sleep [25].
It consists of 7 items reflecting the individual’s level of sleep effort to sleep. The GSES is based on the Pittsburgh Sleep Quality Index (PSQI), which measures sleep quality and disturbance [26]. The PSQI consists of 19 items examining 7 factors: subjective sleep quality, latency, duration, habitual sleep efficiency, sleep disturbance, use of sleeping medication, and daytime dysfunction. The Epworth Sleepiness Scale (ESS) is a well-validated self-administered questionnaire assessing the subjective level of daytime sleepiness in 8 common situations [27]. The Insomnia Severity Index (ISI) is a commonly used 7-item selfreporting instrument assessing the severity of both daytime and nighttime components of insomnia [28,29]. The Center for Epidemiologic Studies-Depression scale (CES-D) is a 20-item questionnaire used to evaluate depressive symptoms experienced in the past week [30,31]; the valid and reliable short form of the CES-D was used in the present study [32,33]. State-Trait Anxiety Inventory (STAI) was used to evaluate anxiety levels [34]. This is a self-report questionnaire consisting of 40 questions. The first 20 questions, STAI-1 used in the paper, referred to how the patient felt overall right now (anxiety as a state). The scale is a 4-point Likert scale. The Korean version of the somatization subscale of the Symptom Checklist 90-item version (SCL-SOM) was used to evaluate somatization symptoms [35]. The SCL-SOM consists of 12 questions related to subjective distress of bodily dysfunction [35]. A higher SCL-SOM score indicates more severe somatization [36]. The Korean version of the SCL-SOM has good reliability and validity [37].
The Korean version of International Restless Legs Scale (K-IRLS) is a self-report scale with 10 items rating the severity of restless legs syndrome (RLS) symptoms as well as the impact of these symptoms on daily life, sleep, and mood [38]. The K-IRLS score ranges from 0 to 40, and a higher score indicates more severe RLS symptoms. The Berlin Questionnaire (BQ) was a questionnaire assessing the risk of sleep apnea [39]. The BQ includes 10 questions divided into three categories; snoring (category 1), daytime somnolence (category 2), and hypertension/obesity (category 3). The risk of sleep apnea is determined if two or more categories are positive.
Online survey also included demographic data such as age, sex, marital status (0 [married], 1 [single], and 2 [divorced/widowed]), monthly income (from 1 [no income] to 6 [>4,500,000 KRW]), weekly working hours, drinking (from 1 [never] to 4 [more than once a week]), physical illness (0 [absence] or 1 [presence]), and job type. Job type was written in a subjective form, and we classified it into 3 categories (A [White collar: managers, professionals, and clerks], B [Service: service workers and sales workers], and C [Blue collar: skilled agricultural, forestry and fishery workers, craft and related trades workers, and elementary workers]) based on the Korean Standard Classification of Occupations [40].
Statistical analysis
Among-group differences in continuous variables were compared using the one-way analysis of variance. Categorical data (e.g., sex) were compared using the chi-square test. Due to the possible confounding effects of covariates, analysis of covariance was also used for among-group comparisons after controlling for age, sex, marital status, working hour, job type, income, physical illness, drinking, CES-D, SCL-SOM, STAI-X, K-IRLS, and BQ. The Kolmogorov–Smirnov test was used to evaluate normality; the Fisher’s least significant difference was applied in post-hoc analysis. All statistical analysis were performed using IBM SPSS software (version 29; IBM Corp., Armonk, NY, USA) [27].
RESULTS
Significant differences were observed in age (F=36.06, p<0.001), sex (χ2=68.95, p<0.001), working hour (F=76.54, p<0.001), job type (F=126.52, p<0.001), marital status (F=8.75, p<0.001), income (F=3.96, p<0.001), physical illness (F=5.23, p<0.001), dirking (F=15.79, p<0.001), CES-D (F=57.54, p< 0.001), STAI (F=17.32, p<0.001), SCL-SOM (F=42.99, p<0.001), and K-IRLS score (F=195.92, p<0.001) among non-SWs, LSWs, and non-LSWs. However, no significant differences were observed in BQ (F=5.12, p<0.006) in three groups. LSWs had significantly higher STAI, SCL-SOM, and K-IRLS score than non-SWs and non-LSWs (all p<0.001). Non-LSWs had significantly higher in these scores than non-SWs (p<0.001). Working hour of LSW was longer than non-SWs and non-LSWs (all p<0.001). Non-LSWs had significantly longer working hour than non-SWs (p<0.001). The monthly income of non-SWs was higher than that of LSWs and non-LSWs (all p<0.001). LSWs drank alcoholics more frequently than non-SWs and non-LSWs (all p<0.001). White color job was more common in non-SWs than in LSWs and non-LSWs (all p<0.001). Physical illness was more common in non-LSWs and LSWs than in non-SWs (all p<0.001). There were more married people in non-LSWs than in LSWs (p<0.001).
Non-LSWs were significantly older than LSWs (p<0.001) and non-SWs (p<0.001). Non-SWs were significantly older than LSWs (p<0.01). Non-SWs had significantly lower CES-D scores than LSWs and non-LSWs (all p<0.001). However, CES-D scores did not differ between LSWs and non-LSWs.
The differences among non-SWs, LSWs, and non-LSWs were significant based on FIRST (F=14.67, p<0.001) and GSES (F=49.32, p<0.001) questionnaires after controlling for age, sex, marital status, working hour, job type, income, physical illness, drinking, CES-D score, SCL-SOM, STAI-X-2, K-IRLS, and BQ (Table 1, Figures 1 and 2). In post-hoc analysis, LSWs and non-LSWs had significantly higher FIRST scores than non-SWs (p<0.001 for LSWs, p<0.05 for non-LSWs). LSWs had significantly higher FIRST scores than non-LSWs (p<0.05). LSWs had significantly higher GSES scores than nonLSWs (p<0.001 for LSWs, p<0.001 for non-LSWs).
Significant group differences were also observed in PSQI (F=38.28, p<0.001), ISI (F=51.89, p<0.001), and ESS (F=3.95, p<0.019) after controlling for age, sex, marital status, working hour, job type, income, physical illness, drinking, CES-D score, SCL-SOM, STAI-X-2, K-IRLS, and BQ (Table 1). Compared with non-SWs, SWs had significantly higher PSQI scores (p<0.001 for LSWs, p<0.001 for non-LSWs) and ISI scores (p<0.001 for LSWs, p<0.001 for non-LSWs). LSWs had significantly higher PSQI scores (p<0.01) and ISI scores (p<0.001) than non-LSWs. LSWs also had significantly higher ESS scores than non-SWs (p<0.01). However, no significant differences were observed in ESS scores between LSWs and non-LSWs or between non-LSWs and non-SWs.
For all participants, FIRST scores significantly correlated with CES-D scores (r=0.54, p<0.001), PSQI (r=0.45, p<0.001), ISI (r=0.56, p<0.001), and ESS (r=0.29, p<0.001) (Table 2). GSES scores also significantly correlated with CES-D scores (r=0.51, p<0.001), PSQI (r=0.54, p<0.001), ISI (r=0.69, p<0.001), and ESS (r=0.27, p<0.001) (Table 2). Similar results were obtained when correlation analysis was conducted for non-SWs, LSWs, and non-LSWs.
DISCUSSION
This study investigated levels of sleep reactivity and sleep effort in SWs (LSWs and non-LSWs) and non-SWs. SWs (LSWs and non-LSWs) had higher levels of sleep reactivity and sleep effort than non-SWs. In particular, LSWs reported higher levels of sleep reactivity and sleep effort than non-LSWs. Higher levels of sleep reactivity and sleep effort were associated with depressive symptoms and sleep disturbance such as poor subjective sleep quality, insomnia, and excessive daytime sleepiness.
Both LSWs and non-LSWs reported poor sleep quality and insomnia compared with non-SWs. LSWs exhibited excessive daytime sleepiness compared with non-SWs. These findings are consistent with prior studies reporting vulnerability to insomnia or daytime sleepiness among SWs [19,23,24], and were expected because insomnia and daytime sleepiness are major characteristics of SWSD [41].
Together, the findings confirm that shift work is associated with high levels of sleep reactivity. In other words, sleep of SWs (especially that of LSWs) were likely to be more negatively affected under stressful situations. Additionally, LSWs had higher levels of sleep reactivity than non-SWs. This is the first study to report the effects of different shift work on sleep reactivity.
Notably, even after controlling for depressive symptoms, sleep reactivity remained significantly different among groups. A previous study found that individuals with high levels of sleep reactivity were more likely to experience depression after starting shift work [14]. However, the results of the present study indicate that the relationship between shift work and sleep reactivity is independent of depression, although shift work and sleep reactivity are both closely correlated with depression [14]. Although the circadian inconsistency of shift work negatively affects mood regulation [42], sleep reactivity may be increased through other pathways independent of mood regulation. Increased sleep reactivity among SWs may be due to greater daily perceived stress [43] or shift work-related circadian misalignment causing stress vulnerability [44].
These sleep disturbances may cause SWs to be more attentive to and concerned about their sleep and sleep problems, consequently increasing their level of sleep effort. Sleep effort is a crucial behavioral and cognitive factor inducing chronic insomnia, so the increased level of sleep effort among SWs paradoxically worsens their insomnia and induces negative cognitive arousal. Thus, sleep problems among SWs may increase their level of sleep effort, which in turn may aggravate their sleep disturbance, thus creating a vicious cycle.
SWs also had higher levels of sleep effort than non-SWs. SWs are more likely to be poor sleepers due to circadian misalignment and irregular sleep patterns caused by their work schedules. According to previous study [45] sleep effort presents in two ways; 1) direct effort to get to sleep or to get back to sleep when awake in bed and 2) indirect effort to sleep by manipulating the setting conditions for normal sleep. The “effort” might be crucial behavioral and cognitive factor inducing chronic insomnia. Thus, sleep problems among SWs may increase their level of sleep effort, and increased sleep effort paradoxically worsens their insomnia which could create a vicious cycle of sleep problems and sleep efforts.
In the current study, LSWs reported higher sleep efforts than non-LSWs. The sleep problems of SWs depend on their work shift [45,46]. LSWs had more insomnia symptoms than day or evening SWs [19]. Therefore, more prominent sleep problems of LSWs may lead to a stronger cognitive arousal such as preoccupation or worry about sleep, which eventually lead to greater sleep effort of LSWs compared to non-LSWs. A previous study also reported that circadian misalignment of LSWs was associated with impaired cognitive function such as attention and information processing [47].
In the present study, non-LSWs reported higher levels of sleep reactivity and sleep effort than non-SWs, although both non-LSWs and non-SWs worked from approximately 6 AM to 10 PM. This result can be explained by the irregular work schedules experienced by non-SWs and non-LSWs. An irregular sleep-activity cycle increases vulnerability to stress [48], which may explain the higher levels of sleep reactivity among non-LSWs than among non-SWs in the present study. The higher levels of sleep effort among non-LSWs compared with non-SWs also suggest that an irregular work schedule may induce dysfunctional sleep effort even without late night work.
Sleep reactivity and sleep effort were associated with increased depression, insomnia, and drowsiness regardless of work schedule. As expected, sleep reactivity and sleep effort affected depressive symptoms, insomnia, and sleepiness in both SWs and non-SWs. These results indicate that higher sleep reactivity and sleep effort of SWs may be a risk factor for SWSD. A previous study suggested that sleep reactivity is a key factor for vulnerability to developing SWSD [14]. Additionally, given the bidirectional relationship between sleep problems and depressive symptoms [49,50], SWSD may worsen the depressive symptoms, while depressive symptoms may raise the risk of SWSD. Higher risk of SWSD raised by sleep reactivity or sleep efforts might be inter-related with depressive symptoms which were also associated with sleep reactivity or sleep efforts.
The present study had several limitations. First, its design was cross-sectional, so temporal or causal associations could not be investigated between shift work exposure and sleep reactivity or sleep effort. Second, although the study population included workers in various types of employment, all subjects were South Korean; this may limit the applicability of the results in other countries due to differences in internal and external environments or culture. Third, participants were permitted to self-report answers to online questionnaires; this facilitated recruitment of study participants with diverse work schedules, but in future research more objective methods will be needed, such as actigraphy and polysomnography. Fourth, although a number of potential covariates were statistically controlled for, some factor which can affect sleep reactivity or sleep efforts, such as anhedonia or psychiatric past history were not controlled.
In conclusion, the results of the present study confirmed higher levels of sleep reactivity and sleep effort in SWs, especially those who worked late nights. Late night work and irregular schedules can increase vulnerability to stress associated with sleep and dysfunctional effort to sleep. Because sleep reactivity and sleep effort are correlated with sleep and mood, increased sleep reactivity and sleep effort may also aggravate sleep and mood problems in SWs.
Notes
Availability of Data and Material
The datasets generated or analyzed during the study are available from the corresponding author on reasonable request.
Conflicts of Interest
The authors have no potential conflicts of interest to disclose.
Author Contributions
Conceptualization: Seog Ju Kim. Data curation: Yunjee Hwang, Hyeyeon Jang, Seog Ju Kim. Formal analysis: Yunjee Hwang, Seog Ju Kim. Funding acquisition: Seog Ju Kim. Investigation: Jooyoung Lee, Sehyun Jeon, Jichul Kim, Somi Lee. Methodology: Yunjee Hwang, Hyeyeon Jang. Project administration: Seog Ju Kim. Resources: Jooyoung Lee, Sehyun Jeon, Jichul Kim, Somi Lee. Software: Jooyoung Lee, Sehyun Jeon, Jichul Kim, Somi Lee. Supervision: Seog Ju Kim. Validation: Hyeyeon Jang, Seog Ju Kim. Visualization: Yunjee Hwang, Hyeyeon Jang. Writing—original draft: Yunjee Hwang, Hyeyeon Jang. Writing—review & editing: Hyeyeon Jang, Seog Ju Kim.
Funding Statement
This research was supported by the National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT) (No. 2022R1A2C2008417), the Bio & Medical Technology Development Program of the National Research Foundation (NRF) funded by the Korean government (MSIT) (No. 2020M3E5D9080561), and a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (No. HR21C0885).