1. Introduction
In response to coronavirus disease (COVID-19), a contagious condition first identified in December 2019, some world governments put in place stay-at-home orders and other movement and social restrictions to minimise the spread of the virus. Movement restrictions varied between New Zealand and Australian states and territories as both countries pursued a suppression or elimination policy, with restrictions including the temporary closure of public spaces and gyms, shops, businesses, schools, and tertiary institutions. In addition, many employees were encouraged or required to work from home, with only essential workers able to continue transiting to the workplace. This is likely to have resulted in changes to employment and health behaviours. Findings from studies across a range of countries indicate that community-wide PA participation decreased [
1,
2], sedentary behaviour increased [
1,
3], and poorer mental wellbeing was associated with lower activity levels during this time [
1,
2,
4,
5]. These behaviours have important health ramifications, as regularly engaging in sufficient amounts of physical activity (PA) and minimising sedentary behaviour is beneficial for health and important for preventing and managing chronic diseases [
6,
7]. The World Health Organization recommends that adults engage in 150–300 min of moderate-intensity or 75–150 min of vigorous-intensity PA (or a combination of both) each week, with muscle-strengthening exercise performed on at least 2 days [
8]. Even before the pandemic, however, less than one half of all adults globally met these guidelines [
9], and in Australia, only 15% of adults reported meeting both the aerobic and strength exercise guidelines on a weekly basis [
10,
11]. This poses a serious public health problem, as physical inactivity is a leading cause of chronic disease, morbidity, and premature mortality and adds to the global economic burden [
12].
Based on previous research, it could be theorised that the reduction in PA during the COVID-19 pandemic could negatively impact work ability and job performance. In an occupational context, physical inactivity is estimated to cost INT
$14 billion in lost productivity globally each year [
12] and is associated with lower physical and mental work ability [
13,
14]. However, systematic reviews conducted before COVID-19 reported that regular exercise may increase work ability [
15] and reduce absenteeism and presenteeism [
16]. These are important findings, as productivity losses due to chronic disease-related absenteeism (i.e., poor health resulting in sick leave) and presenteeism (i.e., when a person at work is unable to perform at full capacity due to illness, stress, or other issues) represent 10–15% of global economic output [
17]. Recent evidence suggests a favourable relationship between higher MVPA and lower absenteeism when accounting for various health factors. Observational (each weekly hour of reported PA was associated with the equivalent of 1.2 days per year lower sickness absence amongst Spanish university employees [
18]) and interventional data [
19] demonstrate positive outcomes for both absenteeism and presenteeism following weekly vigorous intensity aerobic and strength exercise. While findings are not ubiquitous [
15], there is evidence demonstrating improved physical and mental work ability following exercise interventions [
20,
21]. Exercise, therefore, has the potential to mitigate the chronic disease risk associated with sedentary occupations, reduce employee absenteeism and presenteeism, and improve work ability [
16]. However, whether relationships between PA, work ability, and job performance exist when various community-wide movement and social restrictions are in place is unknown.
The aim of this study was to investigate if Australian and New Zealand employees who were meeting aerobic and strength exercise guidelines reported higher work ability and lower work-related absenteeism and presenteeism than physically inactive employees during the COVID-19 pandemic, which was being managed with various movement and social restrictions.
2. Materials and Methods
2.1. Participants and Recruitment
A cross-sectional study design was used to collect self-reported data. Australian and New Zealand adults were invited to participate in an online survey using REDCap (version 10.0.29, Vanderbuilt University, Nashville, TN, USA) a secure, web-based software platform designed to support data capture for research studies [
22,
23]. The survey was advertised through La Trobe University, University of Auckland, and University of Notre Dame Australia research newsletters and social media networks, local news media, and online community noticeboards from 9 June–9 August 2020. This recruitment period was selected to maximise the response of a broad representation of the targeted population, and multiple avenues of participant recruitment were deliberately chosen to engage a diverse range of participants from broad demographic backgrounds. A total of 456 respondents completed the survey. During this period, different movement and social restrictions were in place across Australia’s states and territories, placing various limitations on group sizes for gathering indoors and outdoors.
In New South Wales in June, people could meet outside in groups of up to 20, up to 100 people were allowed inside gyms, and indoor gym classes allowed up to 10 people, providing there was only one person per four square meters. From 1 July, restrictions eased such that there was no upper limit on the number of people allowed at an indoor venue, with the four square meter rule still in place. In Victoria, outdoor exercise classes were limited to 10 people, with no sharing of equipment and 1.5 m distancing between people. From 30 June, residents across Melbourne and surrounding postcodes entered a period of lockdown and were only allowed two hours of outdoor exercise per day with one other person or with household members, while indoor exercise venues were closed. On 2 August, a State of Disaster was declared, with metropolitan Melbourne moving into Stage 4 restrictions, which included no travel further than 5 km from home without a work permit. In Western Australia, non-work indoor gatherings allowed up to 100 people at any one time per single undivided space, up to a total of 300 people per venue, providing there was only one person per two square meters. New Zealand was at Alert Level 1—Prepare, with no restriction on personal movement.
Australian and New Zealand residents aged 18 years or older who were employed at the time of the survey were eligible to participate and provided their informed consent before commencing. No other inclusion or exclusion criteria were applied. The survey collected information on participant demographics and characteristics, including gender, age, ethnicity, marital status, education and employment history, height, weight, medical history, and dependent relationships. The survey also collected information on current health-related lifestyle behaviours, including smoking, alcohol consumption, PA, sleep quantity, and information on work ability, absenteeism, and presenteeism. The study was approved by the La Trobe University Human Research Ethics Committee (HEC10200) and is reported according to the STROBE Statement [
24] (
Appendix A). Data were collected and managed using the REDCap electronic data capture tool hosted at La Trobe University.
2.2. Physical Activity and Exercise Participation
Participants responded to 11 questions about their PA and sedentary behaviour over the past week. This information was collected using the short-form International Physical Activity Questionnaire (IPAQ) [
25] and strength exercise questions that have been used in previous epidemiological research [
10]. Moderate-vigorous intensity aerobic physical activity (MVPA) participation was calculated as weekly energy expenditure (MET-min·wk
−1) using the validated IPAQ formula [
26], and strength exercise participation was reported as the number of days in the past week that the participant performed muscle strengthening exercises. Aerobic and strength exercise participation were compared to current MVPA (i.e., ≥500 MET·min·week
−1) [
8,
27] and strength exercise (≥2 days each week) [
8] guidelines, respectively. Sedentary behaviour was reported as time spent sitting on a typical weekday during the last seven days.
2.3. Other Lifestyle Behaviours
Participants were asked whether or not they were a current smoker or if they had quit in the last six months and how often (if ever) they consumed alcohol. Participants were also asked how many hours of actual sleep they got at night on average over the last month, after reporting their usual bedtime, wake time, and the duration taken to typically fall asleep.
2.4. Work Ability
Three questions from the Work Ability Index [
28] were included. The first question, validated as a single item [
29], asked participants to rate their current work ability compared to their lifetime best using an 11-point scale ranging from 0 (cannot work) to 10 (lifetime best). The next two questions asked participants to rate their current physical work ability with respect to the physical demands of their work and their current mental work ability with respect to the mental demands of their work, each using a 5-point scale ranging from 0 (very poor) to 4 (very good). To enable participants to respond accurately, Tengland’s (2011, p. 275) definition of work ability was provided: “Having the occupational competence, the health required for the competence, and the occupational virtues that are required for managing the work tasks, assuming that the tasks are reasonable and that the work environment is acceptable” [
30].
2.5. Absenteeism and Presenteeism
Absenteeism and presenteeism (constructs of employee productivity) over the past month were measured using the validated World Health Organisation’s Heath and Work Performance Questionnaire following the provided instructions [
31]. To measure absenteeism, participants reported their expected weekly hours of work and their actual number of hours of work over the past month after accounting for time off due to health reasons. This was converted to a measure of relative absenteeism, expressed as the difference between actual hours worked and expected hours of work, as a fraction of expected hours of work, and expressed as a percentage. Negative values indicated that participants worked more hours over the past month than what was expected of them. To measure presenteeism, participants were asked to rate their overall job performance on the days they worked over the past month using an 11-point scale ranging from 0 (total lack of job performance) to 10 (no lack of job performance), which was then expressed as a percentage (0–100%). Presenteeism (i.e., lack of job performance) was calculated as the difference between the reported score and 100%, with higher presenteeism values indicating poorer job performance.
2.6. Statistical Analyses
Univariate (unadjusted) and multivariate (adjusted for potential confounding variables including age, gender, ethnicity, marital status, level of education, employment level, employment group, BMI, number of medical conditions, medications, number of dependents, homeschool responsibilities, smoking status, alcohol consumption, sleep quantity and sedentary behaviour) regressions were used to investigate relationships between PA, mental and physical work ability, absenteeism, and presenteeism. Logistic regression was used for binary outcomes, and ordinal logistic regression for ordered categorical outcomes. The proportional odds assumption was checked and was met for the ordered logistic regressions. Quantile linear regression for the median was used for continuous outcomes. Quantile regression was chosen, as the assumptions behind ordinary linear regression were not met, namely normality and constancy of variance of the residuals. Non-linear fits were also checked for the continuous covariates in the quantile regressions, but the Akaike Information Criterion (AIC) indicated that linear functional forms were better fits. To arrive at a parsimonious model of best fit, a backward elimination process based on the AIC was used. Mean and 95% confidence intervals (CIs) are reported unless otherwise indicated.
The Chi-Square test of association was used to investigate whether participants who reported meeting the MVPA guideline were more or less likely to be meeting the strength exercise guideline (and vice-versa), and to further investigate whether those with medical conditions or those in part-time employment were more or less likely to be meeting the MVPA or strength exercise guidelines, with the phi coefficient (correlation coefficient) indicating the effect size (0.1 = small, 0.3 = moderate, 0.5 = large). The number of participant responses for each outcome is reported in each table. Data processing, chi-square, and
t-tests were conducted with Stata (version 16.1, StataCorp LLC, College Station, TX, USA). The regressions were carried out in freeware R [
32]. The ordered logistic regressions were conducted with the polr function from the MASS library [
33]. The quantile regressions were conducted with the quantreg library (R package version 5.67) [
34].
4. Discussion
Australian and New Zealand adult employees who reported meeting the MVPA guideline reported higher physical and mental work ability (i.e., occupational competence, health, and occupational virtues) and higher job performance compared to those who were not meeting the guideline during the early phase of the COVID-19 pandemic. Participants with one or more medical conditions were less likely to meet the MVPA guideline, while the difference between part-time (less likely) and full-time employees (more likely) approached statistical significance.
The association between meeting the MVPA guideline and higher work ability was expected, as previous research has reported that higher PA is associated with better cognitive and physical function and greater mental and physical wellbeing in healthy adults in the workplace [
13,
14,
35]. Our finding that participants who met the MVPA guideline reported higher job performance and work ability during the COVID-19 pandemic compared to those who did not meet the guideline is important, as other research has reported reduced job performance (i.e., greater presenteeism) in adults who transitioned from the workplace to working either fully or partly from home during this time [
36]. Critically, though, the potential effect for PA to influence job performance during work from home was not investigated [
36], as was the case in our study. Furthermore, software professionals working from home during the pandemic reported that distractions, boredom, and the need for competence were negatively associated with self-reported job performance, but in contrast to our findings, there was no association between job performance and PA [
37]. Our findings suggest that engagement in PA can support employee job performance in changing work environments, which is an important consideration for employers.
General work ability was 20% below lifetime best, and participants reported working 4.0% hours more over the past month than usually required (i.e., overtime) but reported a 30.0% lack of job performance, which may reflect job-related [
5] or other stress and may lead to poorer health outcomes over time [
38]. Factors negatively affecting work ability have been identified as a lack of leisure time, vigorous PA, poor musculoskeletal health, older age, high mental work demands, and poor work environment [
39]. During COVID-19, many employees were required to work from home, resulting in increased distractions and fewer peer-to-peer social interactions, potentially negatively affecting wellbeing during this time [
5,
37]. The pandemic also appears to have negatively impacted mental health, which could be a factor affecting work performance [
1,
4,
5]. In addition, COVID-19 may have compounded already burdened employees with additional personal responsibilities while working from home, such as homeschooling or caring for relatives. For example, a survey of US employees working from home during the COVID-19 pandemic found that a reduction in physical and mental wellbeing was associated with several factors, including less exercise, having at least one child at home, distractions while working, less co-worker communication, higher workload, having to adjust work hours, and lower satisfaction with indoor workspace set-up [
35]. Participants in our study may have been experiencing similar burdens, as they reported a 30% lack of job performance despite working more hours than was expected of them, noting that those who with higher MVPA participation reported lower presenteeism.
Part-time employees were less likely to meet the MVPA guideline than full-time employees and were less likely to report a rather good or very good mental work ability. Part-time employees were more likely to be female and reported having additional responsibilities; specifically, they were more likely to have caring and homeschooling responsibilities, which could have reduced their opportunities for PA participation. Given the evidence that increasing PA improves physical and mental work ability [
20,
21], part-time employees might require additional support to promote engagement in PA and exercise, particularly if they are not commuting to the workplace. Furthermore, those who had one or more medical conditions were less likely to meet the MVPA guidelines, while people with three or more medical conditions were less likely to report a rather good or very good physical work ability. Reports of severity of COVID-19 infection in individuals with existing medical conditions such as diabetes and obesity may have deterred individuals from completing MVPA for want of lowering their exposure risk, or those with medical conditions may have been less physically active as a result of their condition(s). Either way, employees with medical conditions might require additional exercise behaviour change support or education on how to exercise safely, particularly in these circumstances.
The amount of weekday sedentary behaviour that people reported (~7.9 h) was similar to that of the average healthy adult population (median 8.2 h) prior to the COVID-19 pandemic [
40]. Previous research in a similar population demonstrated that employees spend around 77% (~6.6 h) of their working hours and almost two-thirds (~4.3 h) of non-working hours each weekday in sedentary behaviour [
41], and sedentary time was potentially 50% greater in those working from home compared to those working at the workplace during the COVID-19 pandemic [
3]. In our study, the likelihood of reporting poorer mental work ability and reduced job performance increased with increases in sedentary time.
The proportion of participants in this study who reported meeting the MVPA and strength exercise guidelines was higher than that reported in adult Australian and international populations prior to COVID-19 [
9,
10,
11]. Other data have indicated that some people may have engaged in more PA because they had more time due to not having to commute to work and recognised the need to participate in regular PA during the pandemic [
2]. Of the 185 participants who met the strength exercise guideline, 81.1% also met the MVPA guideline. This finding suggests that promoting strength exercise participation might be an effective strategy to increase adherence to both the MVPA and strength exercise guidelines when movement and social restrictions are in place. With appropriate education, strength exercise in the form of using body weight as resistance can be easily completed in the home, allowing people with existing medical conditions to restrict potential exposure to COVID-19 while increasing their PA participation.
It must be acknowledged that the results are only generalisable to the sample included in the survey and not the wider working population. Furthermore, there are limitations with the use of the self-reporting tools to collect data on PA, and the IPAQ measures PA over the past 7 days, whereas presenteeism and absenteeism were measured over the past month. The specific restrictive measures implemented across Australia and New Zealand differed between jurisdictions throughout the survey period, and it is possible that movement and social restrictions changed between the earliest date requested to consider activities in the survey tools and when the respondent completed the survey. Therefore, it is not clear how the current level of restriction in relation to any previous restrictions influenced outcomes, and a decision was made not to include degree of restriction as a covariate.