Prospective changes in sleep problems in response to the daily rest period among Japanese daytime workers: A longitudinal web survey

Summary The daily rest period (DRP) is the daily inter‐work interval and can include a sleep opportunity, leisure time, and other non‐work time. A longer DRP may allow workers to increase time in bed (TIB) and adjust sleep timing, and that may reduce sleep problems such as short sleep duration, sleep debt, social jetlag, and poor sleep quality. The present study examined the longitudinal association between the DRP and these sleep problems among Japanese daytime workers. The DRP, TIB on workdays, sleep quality (Pittsburgh Sleep Quality Index [PSQI]), sleep debt and social jetlag were measured in November 2016 (n = 10,000) and February 2019 (n = 3,098). Of these, 955 permanent daytime workers were divided into five groups based on the change in the DRP duration: shortened ≥2 hr, shortened ≥1 hr, no change (<1 hr), extended ≥1 hr and extended ≥2 hr. Linear mixed‐model analysis revealed significant interaction (group × time) effects on the TIB, PSQI score and sleep debt (all p < 0.001), but not on social jetlag (p = 0.476). Post hoc comparisons revealed that the TIB was decreased, and the sleep debt was increased in the shortened ≥2 hr group, whereas the TIB was increased and PSQI score was improved in the extended ≥2 hr group (all p < 0.01). These findings suggest that an extension of the DRP improves sleep quantity and quality but not sleep debt and social jetlag. Aside from extending the DRP, ensuring a sufficient sleep duration and adjusting sleep timing during the DRP may also be needed to prevent sleep problems.

The European Union (EU) working time directive states that EU workers have the right to take 'a minimum daily rest period (DRP) of 11 consecutive hours every 24 hr' (European Parliament Council, 2003).
In Japan, a 'work interval system', which requires employers to ensure workers get a certain interval of hours from the end of one workday to the start of the next (i.e. DRP; Ministry of Health, Labour, and Welfare, Japan, 2017), has been an obligation to make an effort to attain by the amendment of the 'Act on Special Measures for Improvement of Working Hours Arrangements', which came into effect on April 2019 (Ministry of Health, Labour, and Welfare, Japan, 2021). The DRP is expected to include activities and/or behaviours normally performed outside work hours, i.e. sleep, leisure time, commute time, and other non-work time. Therefore, a longer DRP may result in a longer sleep opportunity. As a longer sleep opportunity could increase the time in bed (TIB) and allow a suitable adjustment of the sleep timing on workdays (social time) to non-workdays (biological time), a longer DRP may improve sleep duration, sleep debt, and social jetlag. In addition, given that workers with many quick returns (short DRP of <11 hr) have frequent insomnia (e.g. Eldevik, Flo, Moen, Pallesen, & Bjorvatn, 2013), extending DRPs could also improve sleep quality. Our previous study examined the cross-sectional association between DRP and these sleep problems and revealed that workers with shorter DRP had significantly shorter sleep duration, poorer sleep quality, and greater amounts of sleep debt and social jetlag (Ikeda et al., , 2021.
However, those previous studies were based on a cross-sectional design, which does not address the causal links between DRP and sleep problems. A full understanding of the link between prospectively extending the DRP and the potential reduction in these sleep problems needs to be established.
The present study sought to address the research gap by examining the longitudinal association between DRP and sleep problems among Japanese daytime workers. We hypothesised that a longer DRP is longitudinally associated with fewer sleep problems.

| Survey and sampling
This study was approved by the Research Ethics Committee of the National Institute of Occupational Safety and Health, Japan (H2742 and H3027). The baseline survey was conducted in November 2016 and the follow-up survey in February 2019. The surveys were conducted via a website by a third-party research company (IDEA PROGGET Co., Ltd.) with a voluntary registrant database comprising almost 3.5 million individuals. In the baseline survey, the company randomly emailed invitations to participate in the survey to 60,000 registered workers. The individuals interested in the survey accessed the website Uniform Resource Locator (URL) enclosed in the email. There, they provided web-based informed consent before they completed the survey. Subsequently, the participants received rewards points from the company. The first 10,000 workers (aged 20-64 years) who met the sample population criteria based on a composition ratio of sex, age group, and industry type as reported in the Labour Force Survey (Statistics Bureau, Ministry of Internal Affairs, and Communications, Japan, 2016) were recruited. In the follow-up survey, participation requests were sent by email to the same 10,000 workers, and 3,098 of them completed the follow-up survey.
As this study focussed only on permanent daytime workers, nightshift workers and workers with non-permanent employment (e.g. part-time or temporarily employed; n = 1,623) were excluded from subsequent analyses. In addition, permanent daytime workers with outlier data (±3 standard deviations [SD] from the mean of permanent daytime workers on the start of work, end of work, bedtime and/or wake-up time; n = 520) were also excluded. The final sample size (i.e. permanent daytime workers without outliers) was 955 ( Figure 1).

| Data collection
Demographic data were collected, namely, sex, age (years), smoking habit (current smoker or ex-and non-smoker), alcohol status (non-consumption or more than once per week), employment type F I G U R E 1 Participant enrolment. SD, standard deviation (permanent worker, part-time worker, dispatched worker, contract employee, entrusted employee and other), presence or absence of night shift work (from 10:00 p.m. to 5:00 a.m.), and industry type (16 types).
The Workers' Living Activity-Time Questionnaire (Matsuo, Sasai, So, & Ohkawara, 2016), a self-administered questionnaire, was used to ask about the averages in the previous month for the following: bedtime of the previous workday and non-workday; wake-up time on workdays and non-workdays; presence or absence of the need to commute; if commuting, the start and end times of the commute; and the work beginning and end times. The DRP was calculated as the interval from 'the end of working hours' to 'the beginning of working hours'. The TIB on workdays was calculated from 'bedtime of the previous workday' to 'waking time on workday'. The TIB on non-workdays was calculated from 'bedtime of the previous nonworkday' to 'waking time on non-workday'. Sleep debt was then derived by determining the difference in TIBs, which was calculated by subtracting the TIB on workdays from that on non-workdays.
To assess the longitudinal association between change in DRP and sleep problems, linear mixed-model analysis with an unstructured covariance matrix was conducted for the TIB (workday and non-workday), PSQI score, sleep debt, social jetlag, bedtime (workday and non-workday), wake-up time (workday and non-workday), mid-sleep (workday and non-workday), leisure time, and round-trip commute time. The groups (five groups) and survey times (baseline and follow-up) were set as fixed effects and individual subjects as random effects. In addition, sex, age, smoking habit, and alcohol status were included in the model as fixed effects as confounding variables. The significance level was set at 0.05. Post hoc comparison for interaction effects was conducted by linear mixed-model analysis with an unstructured covariance matrix in each group and adjusted by Bonferroni corrections (the significant level for the post hoc tests was set at 0.05/5 = 0.01). The random effect was individuals, and the confounding variables were sex, age, smoking habit, and alcohol status.
All statistical analyses were conducted using SPSS version 23.0 for Microsoft Windows (SPSS Software Inc.).

| RE SULTS
The demographic data are shown in Table 1

| DISCUSS ION
The present study examined the longitudinal association between DRP and sleep problems among Japanese daytime workers, and our findings thus partially support our hypothesis that a longer DRP is longitudinally associated with fewer sleep problems.
The DRP includes sleep opportunity, and a longer sleep opportunity may allow for a longer TIB. The previous cross-sectional studies found that a longer DRP was associated with a longer TIB  and sleep duration  on workdays. Therefore, a change in DRP may influence the actual TIB.
The present longitudinal study showed that a change in DRP influenced the TIB on workdays, that is, a longer DRP increased the TIB on workdays. However, the TIB on workdays was significantly changed in the DRP extended/shortened ≥2 hr groups, but not in the DRP extended/shortened <2 hr groups. In contrast, leisure time was significantly changed in the DRP extended/shortened ≥2 and ≥1 hr groups. A previous cross-sectional study found that the DRP correlated with the TIB and leisure time on workdays; however, the correlation between the DRP and the TIB was significantly weaker compared with leisure time . These results suggest that when the DRP change is small, the change may be mainly   (Roenneberg et al., 2007). Further studies are needed to solve the issue.
Previous studies found that DRP and quick returns are associated with sleep quality (Eldevik et al., 2013;Ikeda et al., 2018).
In the present study, the analysis of the longitudinal association between the DRP and sleep quality showed that the workers who extended their DRP showed improved sleep quality. However, the cause could not be determined in the present survey. One possible reason is the reduction of stress by extended leisure time and/or shortened working hours on workdays rather than the extension of the DRP per se. Further studies would be needed to clarify the causes.  were unable to work as hard as they used to after an illness or accident. In these cases, sleep problems are unlikely to be improved even if the DRP were extended. Further interventional studies are needed to clarify the issue.
Despite these limitations, the present study established a longitudinal association between the DRP, which contains sleep opportunity, and sleep problems in Japanese daytime workers. Although the work interval system has been promoted in Japan, only 4.2% of 4,191 Japanese companies have adopted the system (Ministry of Health, Labour, and Welfare, Japan, 2020). Introduction of the work interval system and extension of the DRP may help improve sleep quantity and quality. In addition, ensuring the recommended sleep duration and adjusting sleep timing during the DRP may reduce sleep debt and social jetlag. Overall, these measures may reduce workers' sleep problems and related health issues.

This study was supported by the Industrial Disease Clinical
Research Grants from the Ministry of Health, Labour and Welfare, Government of Japan (150903-01).

CO N FLI C T O F I NTE R E S T
The authors have no conflicts of interest to declare.

AUTH O R CO NTR I B UTI O N
HI designed the study. TK, TS, XL and TM conducted the survey. HI and YN carried out analyses. HI wrote the main text. MT supervised the study and managed the research grant. All authors contributed to the final version of the manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
Research data are not shared because we did not inform the participants of the data transparency nor declare the possibility on the institutional review board.