Changes in Workers’ Physical Activity and Sedentary Behavior during the COVID-19 Pandemic

Background: COVID-19 has changed the world and strongly affected the health of the people and the quality of their life. These changes might impact employees’ physical activity (PA) and sedentary behavior (SB). This study aimed to summarize the literature focusing on the COVID-19-caused changes in physical activity and sedentary behavior among the adult working population. Methods: Literature searches were conducted in June 2021 using multiple electronic databases. The following keywords and synonyms were used during the searching process: physical activity, sedentary behavior, COVID-19, employee. After the data cleaning process (duplicates, inclusion criteria), the title and the abstract of all manuscripts from the searches were screened independently by two reviewers. Results: Thirty-nine manuscripts were selected as a result of the searching process. Of these, 5 were SB related, 15 were PA related and 19 addressed both PA and SB. There were longitudinal (10), cross-sectional (28) and a case study (1) in the selected manuscripts. The majority of studies were conducted in the USA (6) and Europe (18), and four studies examined the PA and/or SB in multiple countries. The majority (34 studies) of the studies used subjective, self-reported, but mostly before-validated questionnaires. Objective measures were less common and used only 12.8% of the examined studies. Moreover, 76.4% of the studies described an overall decrease in the amount of PA during the COVID-19 pandemic. In three cases, researchers observed an increase in PA among the workers. Five studies reported no significant changes in the amount of PA during the pandemic. As far as SB is concerned, 18 out of 24 of the studies reported an overall increase in the amount of SB between the two periods. Four manuscripts reported no significant change in the amount of SB, and there was only one manuscript in the examined studies that reported an overall decrease in the SB time before and during the COVID-19 pandemic. Discussion: There is no doubt that the COVID-19 pandemic, including lockdown and work from home (WFH) policies, impaired the PA and SB level of the populations. This period has delivered an important message for the adult working population as well. They should be as active as possible and avoid high levels of SB and uninterrupted sitting time. Therefore, organizations, policies and public health bodies should motivate workers, especially office workers, to be more active and interventions must be developed to mitigate the negative effects of the COVID-19 on PA and SB.


Introduction
People's lifestyles have changed significantly in recent years, with an increasing number of people living a sedentary lifestyle, mainly in developed countries. According to

Materials and Methods
In June 2021, a literature search was performed, and the scientific journal articles were retrieved from ScienceDirect, PubMed and Scopus. We used these sources since all of them provide immediate access to the relevant studies. Manuscripts were selected according to the following inclusion and exclusion criteria: Inclusion criteria: • Study populations involving healthy adult employees (in any country); • Study outcomes investigating any form of PA and/or SB changes before and during the COVID-19 lockdown; • Manuscripts written in English language; • No restriction used in connection with study design; • Publication type involving research papers.
Exclusion criteria: • Study population involving children, adolescents, older or retired people or with a specific disease condition; • Study outcomes not investigating the change of the PA and/or SB before and during the COVID-19 lockdown; • Study not examining the employees' subgroup; • Publication type involving letters to editor, comments, editorials, reviews and recommendations.
At the end of the journal database search and selection process, an additional academic search was conducted, which was not restricted to only journal articles. Dissertations, master theses and project works were examined in connection with the topic. The same inclusion and exclusion criteria were applied.
To cover the relevant studies, literature searches will only include papers published since November 2019. All articles whose topics were related to the present study were taken into consideration. During the search, Boolean operators (such as AND or OR) and the following keywords were used (Table 1).
Duplicates were removed with the use of Zotero reference management software. The title and the abstract of all manuscripts from the searches were screened independently by two reviewers (G.R., É.B.B.), who determined the acceptability of the studies according to the inclusion and exclusion criteria. If disagreements aroused between the reviewers during the selection process, a third reviewer (K.R.-Ó.) was consulted; this was necessary on five occasions. Selected full-text manuscripts were then examined by G.R. and reviewed by É.B.B., K.R.-Ó., K.K., Z.B., I.F. and P.L. In order to identify other sources, the reference lists of the selected full-text manuscripts were further manually searched.
Reviewers picked out the following data from the selected studies: author (year of publication), the title of the manuscript, country, sample size, PA or SB assessment, study design, working population of the sample, change direction in SB, key findings related to SB, change direction in PA, key findings related to PA, main outcomes and types of outcomes. Scopus (TITLE-ABS-KEY (COVID-19 OR "Novel Coronavirus" AND TITLE-ABS-KEY ("Physical activity" OR "Physical inactivity" OR "Sedentary behavior" OR Sitting) AND TITLE-ABS-KEY (Worker OR "Working from home" OR Employment OR Employee OR Occupation)) ScienceDirect COVID-19 OR Coronavirus AND physical activity OR sedentary OR physical inactivity OR sitting AND worker OR working from home OR employment OR employee OR occupation.

Results
Our search identified a total of 982 records. When duplicates were removed, studies were limited to English and human samples, and inappropriate manuscript types were excluded: a total of 508 records were retained. After title and abstract screening, we located 80 full-text articles. The citations of these papers were then screened. An additional nine eligible papers were identified and one of these was retained after full-text review. At the end of the selection process, additional sources were screened to find other relevant studies, and finally, four studies were added to the selection. Due to the actuality of the research topic, we have found several early-stage studies, and three of them were selected, for which the publications are currently in the peer-review process phase [42][43][44]. In two cases, we selected a high-quality master's thesis or project work [45,46].
The flowchart of the search and selection process is shown in Figure 1. Table 2 contains a summary of the characteristics of the studies reviewed. Thirty-nine manuscripts were selected as a result of the searching process. Five manuscripts were SB related, 15 manuscripts were PA related and 19 manuscripts addressed both PA and SB. There were longitudinal studies (10), cross-sectional studies (28) and a case study (1) in the selected manuscripts. The majority of the studies were conducted in Europe (18) and the USA (6), and four studies examined the PA and/or SB in multiple countries.

Main Characteristics of the Studies
Of the studies, 20 out of 39 had focused on only the employees, and 16 manuscripts had studied the general population but directly investigated the employment status. An average 56.35% of the examined population was employed and investigated directly of these studies. In three cases [47][48][49], there was no information about the ratio of employees in the examined population, but the daily occupational PA was determined, and we considered this information. The total of 39 included manuscripts yielded a total of 118,022 participants with a range of job types. Twelve studies especially examined office workers [23,46,[50][51][52][53][54][55][56][57][58][59], and 11 studies [22,23,42,[44][45][46]52,[60][61][62][63] examined WFH or remote working.

Measurement Types and Outcomes of the Studies
Objective measures were less common and were used in only 12.8% of the examined studies. Different smart technologies were used to identify PA patterns; two times activity trackers [50,51] and accelerometers [52,53] were used, one time a mobile application was used [64] and one time a mixed methodology was used (next to the objective measurement self-report measure International Physical Activity Questionnaire-short form also employed to collect additional information) [65]. The majority (35 studies) of the studies used subjective, self-reported but generally before-validated questionnaires (e.g., International Physical Activity Questionnaire, Global Physical Activity Questionnaire, Occupational Sitting and Physical Activity Questionnaire).
A range of outcomes was used to categorize PA and SB changes in the form of time (e.g., METs/min/week, step counts/day or different types of PA levels in minutes or hours/day or week) and as a percentage of the population. A range of outcomes was used to categorize PA and SB changes in the form of time (e.g., METs/min/week, step counts/day or different types of PA levels in minutes or hours/day or week) and as a percentage of the population.      The total PA did not decline during the COVID-19 lockdown. There was no evidence of a decline in any of the three types of PA (walking, moderate-intensity activity and vigorous-intensity activity). Across the three categories "high," "moderate" and "low" PA, 17% of the participants became less active during the lockdown while 29% became more active.

MET minutes/w eek
Walking, moderate-intensity activity, vigorous-intensity activity.    The total PA did not decline during the COVID-19 lockdown. There was no evidence of a decline in any of the three types of PA (walking, moderate-intensity activity and vigorous-intensity activity). Across the three categories "high," "moderate" and "low" PA, 17% of the participants became less active during the lockdown while 29% became more active.

MET minutes/w eek
Walking, moderate-intensity activity, vigorous-intensity activity.    The total PA did not decline during the COVID-19 lockdown. There was no evidence of a decline in any of the three types of PA (walking, moderate-intensity activity and vigorous-intensity activity). Across the three categories "high," "moderate" and "low" PA, 17% of the participants became less active during the lockdown while 29% became more active.  The total PA did not decline during the COVID-19 lockdown. There was no evidence of a decline in any of the three types of PA (walking, moderate-intensity activity and vigorous-intensity activity). Across the three categories "high," "moderate" and "low" PA, 17% of the participants became less active during the lockdown while 29% became more active.

MET minutes/w eek
Walking, moderate-intensity activity, vigorous-intensity activity.     The total PA did not decline during the COVID-19 lockdown. There was no evidence of a decline in any of the three types of PA (walking, moderate-intensity activity and vigorous-intensity activity). Across the three categories "high," "moderate" and "low" PA, 17% of the participants became less active during the lockdown while 29% became more active.    The total PA did not decline during the COVID-19 lockdown. There was no evidence of a decline in any of the three types of PA (walking, moderate-intensity activity and vigorous-intensity activity). Across the three categories "high," "moderate" and "low" PA, 17% of the participants became less active during the lockdown while 29% became more active.

MET minutes/w eek
Walking, moderate-intensity activity, vigorous-intensity activity.   The total PA did not decline during the COVID-19 lockdown. There was no evidence of a decline in any of the three types of PA (walking, moderate-intensity activity and vigorous-intensity activity). Across the three categories "high," "moderate" and "low" PA, 17% of the participants became less active during the lockdown while 29% became more active.

PA in Adult Working Populations
In total, 76.4% of the studies described an overall decrease in the level of PA during the COVID-19 pandemic. In two cases, researchers observed an increase in PA among working adults. In the case of Franco et al. [54], respondents have been participating in health promotion programs during COVID-19 and maybe for this reason an increase has been detected in PA levels. In the other case [42], there were no specific promoting circumstances related to the PA level increase. Five studies [22,55,57,58,76] reported no significant changes in the level of PA between the periods. Studies addressed changes in PA in different forms: time spent on PA, step counts, percentage of the sample, comparison of results with the previous national data.

SB in Adult Working Populations
As far as SB is concerned, 18 out of 24 of the studies reported an overall increase in the amount of SB between the two periods. Four manuscripts [52,53,57,72] reported no significant change in the amount of SB, and there was only one manuscript in the examined studies that reported an overall decrease in the SB time before and during the COVID-19 pandemic among the working adult population. Studies addressed changes in SB in different forms: time spent on PA, step counts, percentage of the sample, comparison of results with the previous national data.
One paper [71] did not directly examine the changes of the amount of PA or SB pre-COVID-19 versus post-COVID-19 lockdown; however, the possible effect of lockdown can be observed. In Taeymans et al. study [71], the employees had a higher median daily sitting time and lower median summed MET minutes per week than other examined groups and the prevalence of long sitting time is higher compared to national data.

Discussion
The present review provides an overview of studies examining the changes in the employment's PA and SB during the COVID-19 pandemic. We identified 39 studies including different types of employments from several countries. Regardless of the applied methodology, a greater part of the investigated studies found that SB level increased and PA level decreased among the employees during the lockdown period.
Three-quarters of the studies reported negative effects on PA during the COVID-19 pandemic. In two cases, researchers observed an increase in PA among the employees, and six studies reported no significant changes in the amount of PA pre-COVID-19 versus post-COVID-19 lockdown among employees. Although [78] examined the impacts of the pandemic on a general population, our results are in line with the findings by this systematic review [78], which revealed that most of the selected studies found that the level of PA declined during the lockdown period. Several other studies examining different population groups found similar effects of the COVID-19 lockdown on participants' PA and sedentary patterns. For example, an online survey that examined the effects of the COVID-19 pandemic on PA among children in the USA reported that the parents perceived a decline in the PA level of the children as compared to before the pandemic [79]. In a mini-review [80], the authors reported a significant reduction of PA levels among university students compared to pre-lockdown values. Moreover, another review study investigating adults with physical disabilities and/or chronic diseases found a decrease in PA during the first wave of COVID-19 in almost every examined study [81].
A variety of methodologies and PA measures were used in the investigated articles ( Table 2). Most of the studies had a cross-sectional design and applied mostly validated self-reported questionnaires, and only 12.8% of the research used objective measures. To avoid the limitations of the self-reported questionnaires, using objective measurement (e.g., accelerometer, pedometer) might be a solution [64]. Future research should focus on objective, directly comparable, accelerometer-based data where it is possible. For example, worldwide data from different wearable activity trackers revealed an overall decrease in step counts during the COVID-19 lockdown [82][83][84]. Tison et al. [85] examined the changes in step count before and during the COVID-19 lockdown. The study found a decline in smartphone-derived step counts based on aggregated data. Differences can be observed between regions due to the regional variation in COVID-19 timing, enforcement, and behavior change [85].
Another study [64] also examined smartphone-based step counts from 815 Chinese adults before, during and after the lockdown. The study reported a sharp decline in daily step counts due to the abrupt nature of the lockdown, followed by a slow, steady increase, and almost four months after reopening, the total step counts were still slightly below the baseline level [64]. Objective data would be able to highlight the long-term consequences of the pandemic on peoples' PA level changes. The prediction of the long-term health effects of the lower PA level during and after the COVID-19 pandemic is difficult. However, a study from Japan reported a lasting significant decrease in PA level over three years following the 2011 earthquake and tsunami [24,86].
A potential reason for the decline in PA level among working adults might be that, on the one hand, during COVID-19, national governments prohibited several activities such as the majority of outdoor and social activities and most gyms, leisure and sporting facilities closed, therefore, people found it difficult to be active. On the other hand, additional responsibilities (e.g., school-aged children at home) for working adults and especially home-working employees also decreased opportunities to be active.
In contrast with the studies that showed a decline in PA levels among working adults, in two studies from Sweden and two studies from Switzerland, no changes in PA level were reported during the COVID-19 pandemic. These results can be explained by the less strict lockdown measures in the two countries: the stricter the confinement measures, the more likely that PA patterns are affected.
In most cases, people working remotely were less active during the COVID-19 lockdown [44,46,62], and those who continued to work at normal routine usually did not observe such a high level of changes in their PA level [52,58]. Dispersedly mode-specific and subgroup-specific increases can be found in different PA levels. However, the total PA levels decreased [70]. For example, in an Estonian [56] study, office workers reported a significant decrease in self-reported PA levels (sport, leisure), although an increase was reported in work-related PA level. However, the total level of PA decreased.
To slow down the spread of the coronavirus, several countries introduced enforcing social distancing, and the severity of these containment measures varies between countries. Lockdown restrictions, including self-isolation, stay at home, physical distancing, travel limitations, restrictions on public movement and banned sports activities and exercise outdoors are just some examples introduced during the pandemic [87]. This new virus has significantly altered people lives, and employment as well, businesses had to close permanently or constantly and many workers (especially office-based workers) were required to shift to a remote working environment (WFH) to stay safe [22]. For example, in Spain, remote work rose from almost 5% to 34% during quarantine [54]. Another study reported that nearly 50% of the participated organization had more than 80% of their employees working from home during the first wave of the COVID-19 pandemic [88]. The rise of remote working during the COVID-19 pandemic was confirmed in several other studies as well [89][90][91][92]. Switching to home offices may have a negative impact on employments' healthy habits. Employees and especially office workers are already at risk for being physically inactive (spending three-quarters of the workday in SB) [93], and it is especially true during the lockdown and the rise of WFH, which was associated with extended sedentary periods and increased screen time [22,52,54,78,94]. Employees who worked from home spent longer uninterrupted SB time during work time than those who never worked at home [95]. Even more alarming, if people spend more time sitting during the workday, they will spend more time sitting during their leisure time [4,96]. Furthermore, additional responsibilities (e.g., school-aged children) for home-working employees also decrease the opportunities to be active. According to a UK survey, 85% of UK employees with school-age children are balancing homeschooling with work [97].
Not surprisingly, 18 studies out of 24 reported an increase in SB time. In contrast with the studies that showed an increase in SB level among working adults, only four studies reported no changes in SB level during the COVID-19 pandemic (two studies from Sweden and two studies from Switzerland) [52,55,57,71]. These different results could be explained by the less strict lockdown measures in the two countries: the stricter the confinement measures, the more likely SB patterns are affected.
Although Stockwell et al. examined the impacts of the pandemic on a general population, our results are in line with the findings by this systematic review [78], which revealed that most of the selected studies found that the level of SB increased during the lockdown period. Moreover, in one study, despite employees have been participating in a health promotion program (10,000 steps/day), SB was negatively affected during the COVID-19 outbreak. The ratio of sedentary participants significantly increased during the pandemic [54].
Another review study investigating adults with physical disabilities and/or chronic diseases found an increase in SB during the first wave of COVID-19 in almost every examined case [81].
According to McDowell et al. switching to WFH due to the COVID-19 pandemic was associated with greater time spent sitting and was associated with more time spent in SB each day [22,45,46]. Fukushima et al. also support the fact that office workers are very much sedentary while working from home during the COVID-19 regulations, as they reported that WFH workers have 111 more minutes of sedentary behavior per day than non-WFH workers [95].
On the other hand, working from home, or in the office is even better than being unemployed during the pandemic as regards PA and SB. Recent research [22,60,63,69], in accordance with previous research [98], reported a lower level of PA and increased level of SB among unemployed people. In line with these results, previously inactive or low active people at baseline had become less physically active during the COVID-19 pandemic [29,57]. It is important to note that activity change during the COVID-19 lockdown depends on the level of lockdown and the habitual PA [64].
Lack of PA and accumulated SB is a known major health risk factor for poor overall health, premature mortality, and NCDs such as obesity, diabetes, and cardiovascular disease [8,9,11,99,100]. Moreover, it is even more important for older people and people with different chronic illnesses to be active in order to avoid the deterioration of their condition [70,101].
Office or desk workers probably have lower PA levels than other workers [102], therefore, it would be important, on the one hand, for organizations to be able to identify and help to preserve their employees' health regardless of whether they work from home or go to the office, and on the other hand, for governments and health professionals to encourage workers to be more active. Public health interventions are needed to break uninterrupted SB for all workers and motivate them to increase their PA, especially those who work from home. Innovative smart technologies may also contribute to prevent employees from a sedentary lifestyle and increase PA. For example, digital health coach technology can play in enhancing people's healthy behavior [103].
A strength of our study was that we conducted a systematic review in connection with the COVID-19 pandemic impacts on a specific population: the employees' PA and SB. Hence, our review may provide initial insight into the effects of the pandemic on the workers' activity behavior, and the results may contribute to the development of public health interventions among the adult working population.

Limitations
Although this is a review, the study has some limitations. First of all, the still ongoing pandemic and the novelty of the topic is one of the main limitations of the manuscript. Reviewers made great efforts in the searching process to find all relevant manuscripts; however, some articles may have been overlooked due to the examined databases and the selected searching terms. The majority of the studies applied self-reported subjective questionnaires, which were often based on participants' retrospective answers where accuracy and objectivity may be questionable. In several studies, the selected sample did not represent the population at a national level. The different occupations and workplaces included in the selected studies also can be a bias. Different methods of assessing SB and PA in the studies also affect the comparison of the results. Finally, our inclusion criteria could be another bias since the present study investigated only the adult working populations and excluded other parts of the population such as adults with physical disabilities and/or chronic diseases or adolescents and children.

Conclusions
Despite the variety of measurement types and study methodologies of the selected studies, the majority of them reported that PA levels have significantly decreased and, at the same time, that SB levels have significantly increased in the adult working population during the COVID-19 pandemic. The stricter the confinement measures, the more likely that PA and SB patterns are affected. People working from home were less active during the COVID-19 lockdown. In contrast, those who continued to work in a normal routine usually did not observe such a high level of changes in their PA level. On the other hand, being unemployed is even worse than WFH, or working from the office, because unemployed people have a lower level of PA and increased level of SB. The lack of PA and accumulated SB is a known major health risk factor for poor overall health, premature mortality and NCDs. The findings of the present review and the well-known health risk factors of the inactive lifestyle illustrate the urgent need to support people, especially sitting-based workers, to increase PA and decrease sedentary time during the pandemic. The workplace is an outstanding opportunity for promoting PA and reducing sedentary time in the working population. Considering that the COVID-19 is an ongoing pandemic, the adult working population should be continuously surveyed, and interventions and strategies should be developed by organizations, governments and health professionals to increase the level of PA and decrease sedentary time among workers, with a special focus on people working from home.