Working from home during the COVID-19 outbreak in Sweden: effects on 24-h time-use in office workers

Since the COVID-19 outbreak, working from home (WFH) has become the new normal for many workers with office-based jobs [1]. Recommending and implementing WFH is considered an important action to reduce exposure to virus and thereby decrease the risk of infection [2]. However, the increase in WFH during the COVID-19 pandemic may have both positive and negative impacts on working conditions as well as physical and mental health among employees [3].

Physical (in) activity and sleep are important determinants of health [4] and thus an essential issue in public health recommendations [5]. More physical activity is considered beneficial to health while too much sitting is detrimental. Sleep duration shows a U-shaped relationship with health, where both too much [6] and too little [7] sleep is associated with poor health outcomes, including cardiovascular diseases and mortality. Data collected before and during COVID-19 at the population level suggest that physical behaviors, i.e. different types of physical (in) activity, and sleep, have changed; physical activity has decreased, while sitting and total sleep time have increased [8‐10]. One explanation may be that WFH is associated with other physical behaviors than working at the workplace. Thus, although research is sparse, some studies suggest that WFH is associated with changes in the time spent in different physical behaviors [11], but little is known specifically regarding the extent to which physical behaviors of office workers differ between WFH days and days at the office. Also, previous studies may not apply to WFH during the COVID-19 outbreak. First, the extent of WFH has increased largely during the pandemic [12], now including workers with limited previous experience of WFH. Second, as WFH is strongly recommended during the COVID-19 outbreak, and in some cases even mandatory, it is likely not voluntary to the same extent as before the pandemic [13, 14]. Voluntary WFH is often adopted by employees as a strategy to adapt work to private life, motivated by family and household needs [14], and may be practiced mainly by employees having favorable conditions for WFH in terms of family life, household chores and workstation at home. Less voluntary WFH during the pandemic, likely performed even under unfavorable conditions, may be associated with changed physical behaviors because of these contextual differences [15, 16]. More research on the trade-off between positive and negative aspects of WFH is therefore needed as a basis for policy recommendations during and beyond the pandemic.

A day is comprised of 24 h during which time is allocated to different physical behaviors distributed among different domains (e.g. work and leisure). More time can be spent in one behavior only at the cost of reducing time in another behavior. For instance, increasing time in sleep will inevitably lead to less time awake; more time at work will lead to less leisure time; and more time spent being physically active will result in less inactive time. This inherent correlation in time-use data implies that standard statistical procedures cannot be used unless data are first processed using specifically adapted procedures. In Compositional data analysis (CoDA), data are expressed in terms of log-transformed ratios expressing relations between compositional parts (in the present case, different physical behaviors, and sleep), and those ratios can be analyzed using standard statistics [17, 18]. Thus, we used a CoDA approach as a basis for examining how WFH influences the 24-h time-use composition of sitting, standing, moving and sleeping; compared to days working at the office (WAO).

We used wearable technical measurement devices (accelerometers) to accurately determine physical behavior, complemented with a diary identifying work, leisure and sleep time. We used a within-subject repeated-measures design to compare WFH days and WAO days, as this design increases statistical power and controls for between-worker confounding by, e.g. income and skill, which would likely be higher among office workers with extensive WFH than among office workers more confined to the physical office space [19].

The present study focused on office workers working from home during the COVID-19 outbreak in Sweden. We aimed to determine the extent to which 24-h time use differs between days working at the office (WAO) and days working from home (WFH), in terms of wake time relative to sleep, time spent working relative to leisure, and time spent sitting relative to standing and moving.

We conducted a within-subject repeated-measures study to determine the extent to which working from home (WFH) during the COVID-19 outbreak influences the 24-h time-use composition in office workers, compared to days when workers go to the office (WAO). Using a CoDA approach for processing and analyzing behavior compositions, we found that days WFH were associated with more time spent sleeping relative to awake, while behaviors during work and leisure did not change markedly, compared with days WAO.

The observed 34 min increase in sleep time during days WFH occurred at the expense of reduced work (26 min) as well as leisure time (7 min). The CoDA showed that the relative proportions of work and leisure “within” the time awake were almost similar for days WAO and WFH (Table 3). This corroborates previous cross-sectional surveys conducted during the pandemic, suggesting that WFH during the pandemic is associated with longer sleep time [31] and shorter working hours [32] than before the restrictions. A longer sleep time during confinement was also reported in a survey of 14,000 workers from 11 countries [33]. A possible explanation for some waking hours (in our study 34 min) being reallocated to sleep during days WFH is that workers take the opportunity to sleep longer when they do not need to commute to the workplace in the morning. Moreover, the trend towards reduced working time from 512 min/day (8.5 h) on WAO days to 486 min (8.1 h) during days WFH suggests that people tend to work less overtime during WFH days, although this needs to be confirmed in larger studies. This trend is in contrast with previous pre-COVID-19 studies, reporting WFH to be associated with longer working hours [34].

We could not confirm that WFH had any marked influence on proportions of time awake spent in sedentary behavior, standing and moving in our sample of office workers. This is in contrast to a recent survey during the pandemic, reporting that respondents who were now working from home had more self-reported sedentariness than respondents whose working conditions were unchanged [35]. However, among our participants, behaviors at work might have changed during the pandemic. We did not have any pre-COVID-19 data in this population to further examine that issue. Population-based studies indicate that the pandemic may have resulted in reduced physical activity [9, 36], although representative national data confirming that are currently lacking.

Recent 24-h movement guidelines [5] recommend that in order to reduce health risks, adults should spend 7–9 h every night in sleep, at the most 8 h per day sedentary, and at least 150 min per week in moderate to vigorous physical activity, along with several hours of light activity and standing. On average, our sample was a bit closer to achieving a balanced 24-h time use according to these guidelines during days WFH than on days WAO; in particular, sleep increased relative to wake time, which may have beneficial implications for health and wellbeing [7].

However, we also found that our sample of office workers spent more than 10 h per day being sedentary – in conflict with the cited guidelines – and that 59% of this time was accumulated during work. A total sedentary time of this size, which is consistent with other studies of office workers [37, 38], has been associated with considerable health risks and early mortality, particularly among those not also engaged in physical activity of sufficient intensities [4, 39]. Thus, our results suggest that policies should encourage office workers to be more active, regardless of whether they work from home or go to the office.

Although the average worker did not report any major changes in working conditions, performance, wellbeing and boundaries between work and private life during the pandemic compared to before (see Fig. 2), we found a considerable heterogeneity between workers in these outcomes; some workers perceived changes to be beneficial while others reported negative changes. For instance, 28% reported that workload had changed for the worse. Although the factors underlying these results remain unknown, our findings suggest that some workers could be in need of additional support from the organization in order to cope with their work situation during the pandemic [40], and that it would be important for the organization to be able to identify and help these workers. When the COVID-19 pandemic is eventually over, it will be important to compare 24-h WFH behaviors with evidence from before the pandemic, so as to understand the extent to which COVID-19 experiences have led to a “new normal”. Also, more research is needed on the trade-off between negative and positive effects of WFH, together with other factors of relevance to work environment and health, to support recommendations for WFH after the pandemic.

The present study has several strengths, including the access to accurate exposure data based on accelerometry, and the use of a within-subject design which minimized risk of bias and confounding. The quite small sample size is a limitation, since it precluded us from detecting small, yet relevant, effect sizes with a sufficient certainty, and limited the opportunity to investigate possible effect modifiers, such as employment, age, gender, and occupational status. We only collected data in one organization, which limits the generalizability of the results. Also, we only collected data during an early phase of the pandemic (May–July 2020), and behaviors may have changed during later phases. Nevertheless, our results may extend to WFH practices after the pandemic better than results of studies addressing workers strictly forced to work from home, since our data were collected in a population with high occurrence of WFH, yet with an option of also working at the office.

We found that office workers during the COVID-19 outbreak in Sweden spent more time sleeping relative to awake during days when they worked from home, compared to days when they went to the office, while physical behaviors during work and leisure did not change markedly. The observed changes in 24-h time use during days working from home may be beneficial to health. Some workers in this sample reported that their workload, performance and wellbeing changed for the worse during the pandemic, suggesting that they would need interventions to prevent potential health risks.