Validity of the occupational sitting and physical activity questionnaire (OSPAQ) for home-based office workers during the COVID-19 global pandemic: A secondary analysis

∗ Corresponding author. Faculty of Health Sciences, The University of Western Ontario, London, Ontario, N6A 3K7, Canada.

Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.

Associated Data

Abstract

High levels of occupational sitting is an emerging health concern. As working from home has become a common practice as a result of COVID-19, it is imperative to validate an appropriate self-report measure to assess sitting in this setting. This secondary analysis study aimed to validate the occupational sitting and physical activity questionnaire (OSPAQ) against an activPAL4™ in full-time home-based ‘office’ workers (n = 148; mean age = 44.90). Participants completed a modified version of the OSPAQ and wore an activPAL4™ for a full work week. The findings suggest that the modified OSPAQ has fair levels of validity in terms of correlation for sitting and standing (ρ = 0.35–0.43, all p < 0.05) and agreement (bias = 2–12%) at the group level; however, estimates were poor at an individual level, as suggested by wide limits of agreement (±22–30%). Overall, the OSPAQ showed to be an easily administered and valid questionnaire to measure group level sitting and standing in this sample of adults.

1. Introduction

Sedentary behaviour is defined as any waking behaviour in a seated, lying or reclining posture while expending less than or equal to 1.5 metabolic equivalents (Tremblay et al., 2017). Increased time spent sedentary has been associated with a higher risk of type 2 diabetes, cardiovascular disease and all-cause mortality (Katzmarzyk et al., 2019), independent from physical activity levels (Owen et al., 2010). Office-working adults have been shown to spend up to 77% of their workday sitting (Thorp et al., 2012), and therefore represent an at-risk population for high levels of sedentary time. The health concerns associated with high amounts of sedentary behaviour are increasingly being recognized with, for example, the recent launch of the Canadian 24-Hour Movement Guidelines for Adults, which provide time-specific recommendations for limiting daily sedentary time (Ross et al., 2020). Accordingly, it is not surprising that numerous interventions have targeted sedentary behaviour reduction in the workplace (Blackburn et al., 2020).

In response to the SARS-CoV-2 (COVID-19) global pandemic, many desk-based workers have pivoted from working at the office to working from home. For instance, findings from a survey targeting American full- or part-time office-workers (n = 5858), found that only 20% reported working from home almost all the time or full-time pre-COVID; however, since the onset of COVID-19, these proportions have risen to 71% working from home most of the time or full-time (Pew Research Center, 2020). This rapid, unplanned and unequipped transition from office-to home-based settings for many office workers, in addition to social distancing and lockdown measures, has been linked to significant decreases in overall physical activity as well as significant increases in total daily sitting time (Ammar et al., 2020; Fitbit, 2020; Tison et al., 2020). Evidence suggests that these decreases in physical activity may also be having a negative impact on mental health outcomes, including increased depression, loneliness, stress and decreased positive overall mental health (Meyer et al., 2020).

This work-from-home trend seems likely to outlive the global COVID-19 pandemic, as this pivotal transition seems to have already changed the environment of future office work. For instance, many workers have reported both preference and employer granted options for in-office/work-from-home flexibility indefinitely (Anderson et al., 2021; Pew Research Center, 2020). Hence, it is important that interventions begin to target this new growing population's sedentary behaviour patterns. Currently, evidence is sparse regarding interventions aimed at reducing sedentary behaviours in office workers who work from home. In addition to effective interventions, accurate measures to capture sedentary behaviour in this new “office environment” need to be tested and validated. Ideally, a measure that is less expensive and can be easily and quickly distributed, such as a self-report questionnaire, is urgently needed to advance sedentary research in this new segment of the working population.

While the preferred method of sedentary behaviour measurement is with a device that can differentiate sitting from standing (i.e., activPAL™), this type of device-based measurement is usually expensive (e.g., costs associated with purchasing each device, delivery to participants, dressings needed) and relatively invasive to ask participants to wear. Due to this cost barrier and added participant burden, there are a number of self-report questionnaires that have been developed and used in the literature. Questionnaires that have been previously used to assess sedentary behaviours in office working adults include, but are not limited to, the International Physical Activity Questionnaire (IPAQ; Ekelund et al., 2006), the Workforce Sitting Questionnaire (Aittasalo et al., 2017), the Sedentary Behaviour Questionnaire (SBQ; Rosenberg et al., 2010) and the Occupational Sitting and Physical Activity Questionnaire (OSPAQ; Chau et al., 2012). Amongst these, the OSPAQ has been used in an array of populations such as university students (Dillon et al., 2021), university employees (Headley et al., 2018), sedentary obese individuals (Lohana and Yadav, 2020), health professionals (Zafiropoulos et al., 2019) and of relevance, office workers (Nelson-Wong et al., 2020; Rollo and Prapavessis, 2020; Urda et al., 2017), to measure time spent sitting, standing, walking and perfoming heavy labour tasks during work hours. The OSPAQ measures sitting and standing as separate behaviours, thus, making it an ideal self-report tool to properly classify sedentary behaviour separate from physical inactivity. It is also very easy to implement as it only consists of three questions, minimizing participant burden. Validation studies using the OSPAQ have previously been conducted in various populations and demonstrated mixed levels of agreement and reliability depending on the occupation (i.e., sedentary versus non-sedentary) and device-based measure used (i.e., Actigraph versus activPAL™) (Chau et al., 2012; Jancey et al., 2014; Maes et al., 2020; van Nassau et al., 2015). Whether these findings can be replicated among traditional office-workers now working from home warrants investigation. It is imperative to establish the ‘construct validity’ of this questionnaire in an at-home ‘office’ worker population to allow future research to confidently assess sedentary behaviour within this setting, without the need for costly device-based measures. It is also important to note that these previous validation studies carry several limitations and pose risk of bias due to inadequate sample size (van Nassau et al., 2015) or use a device that cannot accurately differentiate sedentary behaviour (i.e., sitting) from physical inactivity (i.e., standing) (Chau et al., 2012; Jancey et al., 2014).

Validity evidence is lacking towards a questionnaire that can be administered to home-based office workers. Hence, a secondary analysis of data from an unpublished randomized controlled trial (NCT04488796) was undertaken to examine the measurement of agreement between the OSPAQ and the activPAL4™ inclinometer for estimating percentage of time spent sitting, standing and moving (i.e., walking) during work hours in office-working adults who had transitioned to working from home due to the COVID-19 pandemic.

2. Material and methods

2.1. Study design & population

We performed a secondary analysis on data from an unpublished pre-registered randomized controlled trial (NCT04488796) that aimed to decrease and break up time spent sedentary among home-based office workers. Data were collected from September to December 2020. Participants were full-time, home-based office workers living in London, Ontario or the surrounding area. Individuals were eligible to participate if they were 18 years or older, self-declared working full-time (i.e., employed 30+ hours/week) 5 days per week (i.e., Monday to Friday), self-declared working at least 3 days per week from their home, were able to read and write in English and had access to a computer with Internet and email. Participants were ineligible if they were planning on leaving their current employer or taking a leave of absence/vacation for more than three consecutive workdays for the duration of the study. They were also ineligible if they self-declared having a medical condition or physical limitation that prevented them from being physically active.

Participants were recruited using a number of strategies. First, contact was made via email with relevant liaisons and/or senior executives of potential businesses of interest (i.e., offices/businesses that were known to be working from home due to COVID-19). If interested, they were asked to email all full-time employees within their respective office/business inviting them to participate. Second, recruitment emails were sent directly to home-based office-working employees whose contact information was publicly available on company or institution websites (e.g., employee directories). Third, home-based office workers were recruited via recruitment posters distributed on various social media platforms (i.e., Facebook, Instagram, Twitter, LinkedIn). The recruitment emails included relevant study details (i.e., objective, eligibility criteria, brief procedures) and instructed interested individuals to contact the researcher via email if they wished to participate or wanted to receive additional details prior to making a decision. The study was approved by the institutional research ethics board.

2.2. Procedure & measures

After receiving a study invitation email, interested participants were sent a link with a unique authorization code and asked to complete an online questionnaire through a survey website called SoSci (www.soscisurvey.de). The online questionnaire consisted of a Letter of Information, informed consent and a baseline questionnaire assessing relevant demographic characteristics and outcomes of interest (i.e., primary and secondary measures including the OSPAQ). Upon completion of the first questionnaire, participants were emailed a PDF version of the Letter of Information/Informed Consent and were asked to sign the form (digitally) and send it back to research personnel, along with their address for activity monitor delivery. Participants received the activPAL4™ device via courier and were instructed to apply the device on Sunday evening and to wear the device all day for a period of 5 working days (Monday through Friday). Upon receiving the activPAL4™, participants also received a link (via email) to a detailed video outlining the proper procedures on how to apply the device. If there was any confusion, they were asked to either email or call one of the researchers. Upon finishing, they were instructed to place the device into the return envelope that was provided, and it was picked up via courier the following Saturday. Participants then underwent a 4-week intervention period, filling out the OSPAQ at the end of each workweek (i.e., Friday). During the fourth week, they again wore the activPAL™ device and this was the period used for this secondary analysis validation study.

2.3. OSPAQ-revised

The percentage of time spent sitting, standing and moving (i.e., walking) during work hours was measured using a modified version of the OSPAQ (Chau et al., 2012). Due to the sedentary ‘office’ setting, “heavy labour or physically demanding tasks” was removed from the questionnaire. This decision was made based off previous work done in the field that have reported low or zero prevalence of this behaviour in the workplace (Chau et al., 2012; Jancey et al., 2014; van Nassau et al., 2015). Participants were asked to record both the total number of days and hours they worked in the last 7 days. Participants were then asked to record the percentage of time spent sitting, standing, and moving (i.e., walking) on a typical workday in the last 7 days (i.e., “How would you describe your typical workday for the last 7 days? This involves only time spent in work-related activities and does not include what you did in your leisure time."). The sum of all percentages were to equal a total of 100% (e.g., 80% occupational sitting, 10% occupational standing and 10% walking). Time spent in each behaviour (minutes) was calculated as follows: [Minutes worked in the last 7 days/Days at work in the last 7 days] × [Percentage of the behaviour reported (i.e., sitting/standing/moving)/100].

2.4. activPAL™

The activPAL™ is currently considered the most accurate field-based measure of sitting time and sit-to-stand transitions (Kozey-Keadle et al., 2011). The activPAL4™ was the model used in the present study and is a small device worn on the midline anterior aspect of the thigh (right or left) that can differentiate between sedentary, standing and free moving activity using proprietary algorithms (Intelligent Activity Classification, PAL Technologies). Participants were instructed to wear the device for a full work week (i.e., Monday-Friday) at baseline as well as during the last intervention week. The activPAL™ monitor has been shown to be highly accurate as direct observation has shown a perfect correlation for time spent sitting/lying, standing and walking in primary school aged children (Aminian and Hinckson, 2012) and has been used in many previous validation studies involving adults (Clark et al., 2013; Júdice et al., 2015). The activPAL™ default settings were used, the validation wear time protocol was set to the ‘24-hour protocol’ (allowing 4 hours of non-wear per day), and data were downloaded in custom duration epochs (15 seconds) via activPAL™ Professional Software (version 8.11.4.61) and transferred to Microsoft Excel (version 16.44). Participants were required to have at least three valid workdays from Monday-Friday to be used in data analysis, which is consistent with previous studies (Edwardson et al., 2017). In the baseline questionnaire, participants were asked to report the start and end time of their workday (i.e., What are the hours you work in-between?). The data analyzed for each participant's workday included the data between the self-declared start time (i.e., 9:30am) up to (and including) the last 15 seconds (5:29:45pm) before the official end time (i.e., 5:30pm). Average daily sedentary time (minutes per day) was calculated [total amount of time/average number of days] using two different equations. First, as the sum of ‘sedentary’, ‘primary lying’ and ‘secondary lying’ time. Second, all the behaviours included in the first approach plus time spent in ‘seated transportation’. Time spent standing was calculated from the ‘upright time’. Time spent moving was done as two separate calculations, the first consisting solely of ‘stepping time’ and the second combining ‘stepping time’ with ‘cycling time’. Each valid day of data was totaled and then averaged for the number of valid days to calculate average daily time (minutes) for the week. The percentage of time spent sitting, standing and walking from the activPAL4™ was calculated as follows [average minutes spent in the behaviour (i.e., sitting, standing or moving): per workday/total minutes of work time (i.e., 9:30 a.m. to 5:30 p.m. = 8 hours*60)] × [100].

2.5. Statistical analysis

Statistical procedures were conducted in SPSS Statistics, Version 27 (SPSS Inc., Chicago, Illinois), GraphPad Prism version 9.0.2 (GraphPad Software Inc., San Diego, CA) and Stata Statistical Software Release 11.0 (StataCorp LP, College Station, TX) software programs. The level of significance was set at p < 0.05. Descriptive statistics were calculated for all demographic variables collected at baseline and are shown as mean (standard deviation (SD)) or number (percentage) of the sample. Univariate ANOVAs (continuous variables) and chi-square tests (categorical variables) were conducted to ensure that there were no systematic differences between participants with valid and invalid data (all p-values > 0.05). Bland and Altman (1999) do not recommend excluding outliers; however, they do suggest assessing the influence of outliers on the results. Therefore, we ran the analysis both before and after removing extreme outliers with a winsorization technique (Guttman and Smith, 1969). A total of 11 data points were imputed this way. The removal of extreme outliers did not impact the results and were therefore left in the analysis.

Spearman correlation coefficients were calculated to assess the degree of association between the activPAL4™ and modified OSPAQ. The strength of the correlation was interpreted as poor (<0.30), fair (0.30–0.50), moderately strong (0.60–0.80), or very strong (>0.80) (Chan 2003). Limits of agreement between the activPAL4™ and the modified OSPAQ were determined according to the recommendations by Bland and Altman (Bland and Altman, 1986). The difference [OSPAQ − activPAL4™] of the two paired measurements (as a percentage) was plotted against the average [(OSPAQ + activPAL4™)/2] of the two measurements (as a percentage). Percentage was deemed the most appropriate way to express the data because the OSPAQ is asked and interpreted as a percentage. The Bland-Altman plots expressed in minutes can be found in supplementary data Fig. 1 , Fig. 2 , Fig. 3 , Fig. 4 , Fig. 5 . The mean difference, or bias, between the methods and the 95% limits of agreement intervals were calculated. Linear regression was used to determine linear bias. Significant linear bias indicates that the variability remained constant across average values while the mean difference increased significantly as average values increased. Therefore, where linear regression showed to be significant, the Bland–Altman plot presents the trend line for mean difference obtained from the regression and limits of agreement ( ± 1.96 SD).

Bland–Altman plot of agreement of total self-report (OSPAQ) sitting time (including transportation time) with device-derived (activPAL4™) sitting time (n = 95). The y axis is the difference between the two measures and the x axis is the average of the two, both expressed as a percentage. The bolded dashed line shows the mean difference between the two measures (−5.63), with the dashed lines representing the limits of agreement ( ± 29.87).

Bland–Altman plot of agreement of self-report (OSPAQ) sitting time (excluding transport time) with device-derived (activPAL4™) sitting time (n = 95). The y axis is the difference between the two measures and the x axis is the average of the two, both expressed as a percentage. The bolded dashed line shows the mean difference between the two measures (−1.58), with the dashed lines representing the limits of agreement ( ± 28.31).

Bland–Altman plot of agreement of self-report (OSPAQ) standing time with device-derived (activPAL4™) sitting time (n = 95). The y axis is the difference between the two measures and the x axis is the average of the two, both expressed as a percentage. The bolded dashed line shows the mean difference between the two measures (−11.71), with the dotted lines representing the limits of agreement ( ± 22.34).

Bland–Altman plot of agreement of self-report (OSPAQ) moving time (including cycling time) with device-derived (activPAL4™) moving time (n = 95). The y axis is the difference between the two measures and the x axis is the average of the two, both expressed as a percentage. The bolded dashed line shows the mean difference between the two measures (+6.44), with the dashed lines representing the limits of agreement ( ± 15.36).

Bland–Altman plot of agreement of self-report (OSPAQ) moving time (excluding cycling time) with device-derived (activPAL4™) moving time (n = 95). The y axis is the difference between the two measures and the x axis is the average of the two, both expressed as a percentage. The bolded dashed line shows the mean difference between the two measures (+6.60), with the dashed lines representing the limits of agreement ( ± 14.93).

2.6. Missing data

On any given variable at a single assessment point, the maximum percentage of missing data was 28% (n = 41). Of the 148 participants that filled out the baseline questionnaire, 108 of them had valid activPAL4™ data and of those, 95 had valid self-report data. Independent samples t-tests revealed that those who had valid activPAL4™ data were not different from those who did not have valid data on all demographic variables (p-values > 0.05). Taken together, all missing data were considered random. Hence, we decided to exclude missing data from the analysis.

3. Results

Descriptive statistics for the demographic variables, days worked from home and minutes worked per week are shown in Table 1 . Percentages and minutes of device based (activPAL4™) and self-reported (OSPAQ) behaviour characteristics during work hours are illustrated in Table 2 . The spearman rank correlation coefficient data between the activPAL4™ device and modified OSPAQ are displayed in Table 3 . All the spearman correlations were found to be significant (p < 0.05). The correlation of the activPAL4™ device with sitting and standing were fair (ρ = 0.35–0.43) and the correlation with moving was poor (ρ = 0.21–0.22).

Table 1

Participant characteristics presented as mean (SD) or count (%) of group.

a In the past 3 months, have you been active for a minimum of 30 min/day on at least 3 days of the week? (i.e., jogging, biking, swimming).

Table 2

Percentages and minutes of device-based (activPAL4™) and self-reported (OSPAQ) behaviours during work hours.