A cross-sectional study of the relationship between exercise, physical activity, and health-related quality of life among Japanese workers during the COVID-19 pandemic

Background: Exercise and physical activity positively affect physical and mental health, and healthy workers contribute to increased work productivity. This study aimed to investigate time spent on exercise during leisure time and physical activity, including time at work, in relation to health-related quality of life (HRQOL) among Japanese workers. Methods: An Internet-based national health survey --Collaborative Online Research on Novel-coronavirus and Work study (CoRoNaWork study)-- was conducted among 33,087 Japanese workers in December 2020. After excluding invalid responses, 27,036 participants were categorized into four and five groups according to exercise time and physical activity time, respectively. Each group's scores were compared on each of the four questions on the Japanese version of the Centers for Disease Control and Prevention Health-Related Quality of Life (CDC HRQOL-4) between each group using a linear mixed model. Age-sex adjusted and multivariate models were used to compare each index of the CDC HRQOL-4.

Interventions to encourage daily exercise even for a short period of time may be an effective strategy to improve workers' health and productivity.

Background
People who frequently engage in physical activity and have good exercise habits are known to have lower morbidity and mortality rates for all-cause mortality, ischemic heart disease, hypertension, diabetes, obesity, osteoporosis, and colon cancer, as well as improved depression and anxiety symptoms and reduced prevalence of panic attacks, social phobia, and plaque phobia [1][2][3][4]. However, about one in four men and one in three women worldwide do not engage in the physical activity and exercise, necessary to maintain good health, as recommended by the World Health Organization (WHO) [5]. Surveys conducted in Japan reported that 32% of men and 26% of women exercised for 30 minutes or more at least twice a week [6], and that 33% of workers participated in regular exercise [7].
Some studies show that absenteeism due to health problems decreased with increased physical activity [8], that regular aerobic exercise and strength training prevented work productivity loss due to presenteeism [9], and that increased physical activity reduced work restrictions [10]. These findings suggest that exercise and physical activity have a positive impact on work productivity. Thus, workers need to have good exercise habits, not only for their health but also for their work performance.
The global epidemic of the coronavirus disease 2019 (COVID-19) will continue to significantly impact . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 18, 2021. ; https://doi.org/10.1101/2021.06.15.21259002 doi: medRxiv preprint daily life, work, and healthcare worldwide in 2021. Studies have shown that public health restrictions related to COVID-19 have reduced the amount of time spent outdoors [11], and that 70% of people surveyed showed a moderate or greater change in physical activity during the period when voluntary restraint was recommended [12]. In Japan, the government declared a state of emergency in April 2020, and people were asked to refrain from going out. In addition, to prevent the spread of infection, facilities where many people gather, such as department stores, movie theaters, gyms, and sports grounds, were requested to close [13]. These measures affected people's exercise habits and reduced the time spent in physical activities, such as walking.
In the present study, we focused on the relationship between physical activity and health-related quality of life (HRQOL). This multidimensional and broad concept includes health, work, culture, and religion, and is limited to physical and mental health factors. The Centers for Disease Control and Prevention (CDC) defined HRQOL as an individual's or group's perceived physical and mental health over time [14].
Four core components have been used to assess HRQOL: self-rated health, the number of days in the past 30 days that a person was physically or mentally unhealthy, and the number of days when unhealthy conditions limited a person's activities. The CDC HRQOL-4, developed by the CDC as a tool for public health surveillance, measures these four factors [15]. International studies have provided support for the reliability and validity of the CDC HRQOL-4. In 2020, the Japanese version of the CDC HRQOL-4 was developed, and evaluated for its reliability and validity for use with workers [16]. Furthermore, because the CDC HRQOL-4 was found to be associated with indicators of work functioning impairment [17], the . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 18, 2021. ; https://doi.org/10.1101/2021.06.15.21259002 doi: medRxiv preprint measurement of HRQOL is helpful for both health and labor-management assessment of workers.
Understanding the relationship between time spent exercising during leisure time, time spent engaged in physical activity including work, and HRQOL will provide important information for developing interventions to improve workers' health and work productivity. Furthermore, we thought that it would be useful to investigate these relationships during the COVID-19 pandemic, mainly because physical activity tends to be restricted. As such, this study investigated the relationship between exercise, physical activity, and HRQOL using a large-scale Internet survey of workers conducted during the COVID-19 pandemic.

Study design and setting
The present study was cross-sectional and used a part of the baseline data of a prospective cohort study that was conducted by a research group at the University of Occupational and Environmental Health in Japan. The study is called the Collaborative Online Research on Novel-coronavirus and Work study (CORoNaWork study). A Japanese Internet survey company-Cross Marketing Inc.
Tokyo-implemented the study, which involved a self-administered questionnaire survey. Baseline survey data was collected from December 22-25, 2020. Incidentally, the baseline survey occurred during the third wave of the pandemic when the number of COVID-19 infections and deaths was overwhelmingly higher than in the first and second waves; therefore, Japan was on high alert. The details of the study protocol are provided by Fujino et al. [18].
. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted June 18, 2021. ; https://doi.org/10.1101/2021.06.15.21259002 doi: medRxiv preprint

Participants
The participants were between 20 and 65 years and working at the time of the baseline survey. A total of 33,087 adults participated in the CORoNaWork study, which employed stratified cluster sampling by gender, age, region, and occupation. A database of 27,036 individuals was created by excluding 6,051 who were determined to have provided invalid responses.

Questionnaire
The questionnaire items are described in detail by Fujino et al. [18]. We used questionnaire data on sex, age, presence of illness under medical treatment, average exercise time during leisure time per day, average physical activity time including work/day, educational background, area of participants' residence, job type, working hours per day, telecommuting frequency, and HRQOL. The CDC HRQOL-4 was used to assess overall HRQOL, including the following four items: (a) self-rated health, (b) the number of physically unhealthy days in the past 30 days, (c) the number of mentally unhealthy days in the past 30 days, and (d) the number of days with activity limitation in the past 30 days.
Self-rated health was recorded as excellent, very good, good, fair, or poor [19]. We used the Japanese version of the CDC HRQOL-4 developed by Japanese epidemiologists and occupational physicians [17].

Outcome variables
The outcome variables were CDC HRQOL-4 scores for self-rated health (5-point Likert scale: 1 = excellent, 2 = very good, 3 = good, 4 = fair, and 5 = poor) and the number of days in the past 30 days . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

Predictor variables
We classified the participants into four groups according to their exercise time during leisure time: ≥ 60 minutes/day, 30-59 minutes/day, 1-29 minutes/day, and almost never. Participants were also classified into five groups according to their physical activity time, including work: ≥ 120 minutes/day, 60-119 minutes/day, 30-59 minutes/day, 1-29 minutes/day, and almost never. These variables were used as the predictor variables.

Potential confounding variables
The following items were used as confounding factors. Sex, age (20-29, 30-39, 40-49, 50-59, and ≥ 60 years of age), educational background (junior or senior high school, junior college or vocational school, university or graduate school), and presence of illness under medical treatment were used as personal characteristics. Occupation (regular employee, manager, executive, public service worker, temporary worker, freelancer or professional, other), telecommuting frequency (4 days/week or more, 2-3 days/week, 1 day/week or less, almost never), working hours per day (less than 8 hours/day, 8-9 hours/day, 9-11 hours/day, 11 hours/day or more) were used as work-related factors. The prefecture of the participants' residence was also used.

Statistical method
We used a linear mixed model (LMM) to analyze the relationships between the four items of the CDC HRQOL-4 and the four groups of exercise time, and the five groups of physical activity time. Two . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 18, 2021. ; models were analyzed for each predictor variable. In the age-sex adjusted model, we treated the four classifications of exercise time, the five classifications of physical activity time, age, and sex as fixed effects, and treated the prefecture of residence as random effects. In the multivariate model, we added personal characteristics and work-related variables to the fixed effects of the age-sex adjusted model.
. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

Comparison of the scores on the CDC HRQOL-4 among the exercise time groups
In the age-sex adjusted and multivariate models, we used LMM to compare the CDC-HRQOL-4 scores for each of the four groups of exercise time (Tables 2 and 3). The two models showed a main effect for self-rated health. Post hoc tests showed that shorter exercise time tended to be associated with . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 18, 2021. significantly lower self-rated health (P = 0.016 for the pair of 30-59 minutes/day group and 1-29 minutes/day group in the age-sex adjusted model, and P < 0.001 for the others; P = 0.017 for the pair of 30-59 minutes/day group and 1-29 minutes/day group in the multivariate model, and P < 0.001 for the others). The two models also showed the main effect for the three items on physically unhealthy days, mentally unhealthy days, and days with activity limitation. Post hoc tests showed that the almost never group had significantly more physically unhealthy days than the ≥ 60 minutes/day group and the 30-59 minutes/day group (P = 0.002 and P < 0.001 in the age-sex adjusted model, respectively, and P = 0.008 and P < 0.001 in the multivariate model, respectively), and significantly more mentally unhealthy days than the other three groups (P < 0.001 for both the age-sex adjusted model and multivariate model). The almost never group had significantly more days with activity limitation than the 30-59 minutes/day group and the 1-29 minutes/day group (P < 0.001 and P = 0.028 in the age-sex adjusted model, and P = 0.001 and P = 0.047 in the multivariate model, respectively).

Comparison of the scores on the CDC HRQOL-4 among the physical activity time groups
The age-sex adjusted model and multivariate model were also used to compare the CDC-HRQOL-4 scores for each of the five groups of physical activity time (Tables 2 and 3). The two models showed a main effect for self-rated health and mentally unhealthy days. In the age-sex adjusted model, the ≥ 120 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 18, 2021. ; https://doi.org/10.1101/2021.06.15.21259002 doi: medRxiv preprint minutes/day group had significantly lower self-rated health than the 60-119 minutes/day and 30-59 minutes/day groups (P = 0.001 and P = 0.012, respectively). Self-rated health was significantly lower in the almost never group than in the 60-119 minutes/day, 30-59 minutes/day, and 1-29 minutes/day groups (P < 0.001, P < 0.001, and P = 0.003, respectively). The ≥ 120 minutes/day group had significantly more mentally unhealthy days than the 30-59 minutes/day and 1-29 minutes/day groups (P = 0.010 and P < 0.001, respectively). The mentally unhealthy days were significantly more in the almost never group than in the 1-29 minutes/day group (P = 0.024). In the multivariate model, the almost never group had significantly lower self-rated health than the ≥ 120 minutes/day, 60-119 minutes/day, 30-59 minutes/day, and 1-29 minutes/day groups (P = 0.006, P < 0.001, P < 0.001, and P = 0.009, respectively), and the ≥ 120 minutes/day group had significantly more mentally unhealthy days than the 1-29 minutes/day group (P = 0.017). For physically unhealthy days, the age-sex adjusted model showed a main effect, and post hoc tests showed that the ≥ 120 minutes/day group had significantly more unhealthy days than the 1-29 minutes/day group (P = 0.015). No main effect was observed for days with activity limitation.

Interaction of the exercise time groups and physical activity time groups
There was an interaction between exercise time and physical activity time in the two models for days with activity limitation. In the age-sex adjusted model, among the almost never group for time spent on exercise, those who almost never engaged in physical activity had significantly more days with activity limitation than the 30-59 minutes/day group (P = 0.005). In the multivariate model, among the almost . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 18, 2021. ; https://doi.org/10.1101/2021.06.15.21259002 doi: medRxiv preprint never group for time spent on exercise, those who almost never engaged in physical activity had significantly more days with activity limitation than the 30-59 minutes/day group and 1-29 minutes/day group (P = 0.007, P = 0.046, respectively).

Discussion
This study clarified the relationship between HRQOL (self-rated health and unhealthy days), exercise time during leisure time, and physical activity time including work. The relationship between exercise time and HRQOL showed that higher self-rated health was associated with increases in time spent exercising, even after adjusting for personal characteristics and work-related factors. In addition, the results suggest that an exercise habit of 30 minutes/day or more may help reduce the number of physically unhealthy days. Exercise, even for a short time every day, may effectively reduce mentally unhealthy days. The results of this study were similar to those of previous studies conducted on adults in various regions [20][21][22][23]. Another study has shown that moderate-to-vigorous physical activity of fewer than 30 minutes/day is associated with a reduced mortality rate [24], thereby implying that even short exercise sessions may contribute to better health and fewer mentally unhealthy days. These results provide valuable information while considering interventions to address exercise habits among workers.
As for the relationship between physical activity, including work, and HRQOL, no time spent or excessive time spent in physical activity was associated with lower self-rated health. In addition, the number of unhealthy days increased with the amount of time spent engaged in physical activity. The no time spent in physical activity result in part may be due to the work-related tasks of the individuals in . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 18, 2021. ; the almost never group. In other words, it may be that more people in this group perform most of their work tasks in a sitting position. A study found a negative relationship between sitting time and self-rated health, although the sample was older adults [25]. It is also possible that the limited amount of time spent engaged in physical activity is due to a low sense of subjective health.
With regard to physical activity time and mental health, a previous study found that more exercise time during leisure time was negatively associated with depression but not significantly associated with more work-related physical activity time [26]. Furthermore, some studies have shown that the lower the socioeconomic status, the higher the work-related physical activity [27], and that people of lower socioeconomic status were at a higher risk of mental disorders and mood disorders [28]. Therefore, longer physical activity time, including work, may be associated with higher stress at work, leading to an increase in mentally unhealthy days.
As for the days with activity limitation, the results suggest that less than 60 minutes/day of exercise can reduce the number of activity limitation days. However, this relationship was different from that found for the other indicators of HRQOL (self-rated health, physically unhealthy days, mentally unhealthy days), where daily exercise, even for a long time, may have a positive impact. This difference may be because daily exercise for a long time increases the risk of injury that may interfere with daily life. Studies have shown that the average recreational runner suffers an injury rate of 2.5 to 12.1/1000 hours during running, which is one of the most popular forms of exercise [29]. In addition, workers who had no exercise habits and no physical activity time had more days with activity limitation, so it is . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 18, 2021. ; especially important for workers who spend little time engaged in some form of physical activity to exercise during their leisure time.
These results indicate that it is crucial to have an exercise habit, regardless of the amount of time spent engaged in physical activity (including work), to have a good health-related quality of life. In addition, the results suggested that between 30 and 60 minutes of exercise daily was the most effective amount of exercise time for having a good health-related quality of life. Conversely, daily exercise, even for a short time, is thought to improve self-rated health and decrease mentally unhealthy days. Therefore, encouraging workers with no exercise habit to exercise daily even for a short time may benefit their health and work productivity.

Limitations
This study has four limitations. First, the generalizability of the results is unclear because the CORoNaWork study is an Internet-based survey. However, to reduce sampling bias, sampling was conducted by generation, residence, and occupation. Second, because this was a cross-sectional study, the causal relationship between exercise time, physical activity time, and HRQOL could not be determined. Further research is needed to investigate the causal relationships using experimental designs with workers or employees as the sample population. For example, Emerson et al. [30] demonstrated that the effects of an exercise and nutrition workplace wellness program reduced stress and improved participants' quality of life. Third, the questionnaire only asked about the duration of exercise in leisure time and physical activity, including work, not for exact exercise intensity and physical activity intensity.
. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 18, 2021. ; Fourth, this study was conducted during the COVID-19 pandemic. As such, the impact of exercise and physical activity on HRQOL needs to be considered within the environmental context of the pandemic, which might differ under normal circumstances. Considering that the duration of physical activity was affected by the COVID-19 pandemic and that the participants were anxious about COVID-19 infection, a continuous evaluation is necessary.

Conclusions
This study investigated the relationship between exercise time in leisure time, physical activity time, and HRQOL. These results suggest that daily exercise has a positive impact on HRQOL. Interventions to encourage workers with no exercise habit to exercise daily, even for a short time, may be an effective strategy to improve both workers' health and work productivity during the COVID-19 pandemic. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 18, 2021. ;

Ethics approval and consent to participate
This study was approved by the Ethics Committee of the University of Occupational and Environmental Health, Japan (No. R2-079). Informed consent was obtained from the participants' website.

Consent for publication
Not applicable.

Availability of data and materials
Due to the nature of this research, the participants of this study did not agree to their data being publicly shared, and hence, supporting data was not available.

Competing interests
The authors declare that they have no competing interests. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 18, 2021. ; 8. López-Bueno R, Smith L, Andersen LL, López-Sánchez GF, Casajús JA. Association between physical activity and sickness absenteeism in university workers. Occupational Medicine.
. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 18, 2021. ; is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted June 18, 2021. . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint