Analysis of 24-hour physical behaviours among office workers: approaches, impacts of the pandemic, and methodological considerations

Abstract

Introduction: Physical behaviours over a 24-hour day—including sitting, standing, moving, and sleeping—are critical to health. Office workers, particularly those who are overweight or obese, are at risk of developing health issues due to prolonged sedentary time. The COVID-19 pandemic further complicated physical behaviour patterns, especially with the shift to remote work. While interventions to change physical behaviours exist, little is known about their effectiveness in high-risk populations or how study design factors influence the accuracy of observed group differences in physical behaviours. This thesis aimed at addressing these gaps through three different studies: (1) evaluating the effectiveness of different intervention approaches to improve 24-hour time-use (e.g., decrease sitting), (2) investigating the impact of work location and weight status on temporal patterns of physical behaviours during the pandemic, and (3) examining the methodological factors influencing the accuracy of group differences in 24-hour time-use. Study 1 – Objective: To compare the effectiveness of the traditional “reduce sitting at work approach” and the “24-hour time-use approach” in modifying 24-hour time-use compositions of physical behaviours among overweight and obese office workers. Methods: A non-randomized pilot study was conducted with 45 office workers assigned to three groups: “Reduce sitting at work” (n=15) receiving an intervention focused on reducing sitting time at work; the “24-hour” (n=15) receiving an intervention aiming to reduce sitting at work as well as promoting behavioural changes around 24 hours (e.g., sedentary lifestyle, benefits of physical activity, and healthy sleep hygiene); or “control” (n=15) without any intervention. Daily time spent in physical behaviours (sitting, standing, active, and in bed) was monitored for 7 days using a thigh-worn accelerometer at baseline, and at the 3- and 6-month follow-ups. Intervention effects were analysed using linear mixed models, with a compositional data analysis approach. Results and Conclusion: No significant advantage of the “24-hour time-use approach” over the “reduce sitting at work approach” was observed. Both intervention approaches resulted in reduced sitting time and increased standing time compared to the control group at 3-month follow-up, but these effects were not sustained at 6-month follow-up. Domain-specific (i.e., work and leisure) analysis revealed that minimal changes occurred during leisure hours. Workplace-delivered interventions remain limited in promoting consistent behaviour changes throughout the entire day. Study 2 – Objective: To examine differences in temporal patterns of physical behaviours, assessed using accelerometry, on working and non-working days among normal-weight (body mass index [BMI] <25 kg/m2) and overweight (BMI ≥25 kg/m2) office workers who were either working exclusively at the office (WAO) or exclusively from home (WFH) during the COVID-19 pandemic. Methods: This cross-sectional study included 43 workers WAO (21 normal-weight and 22 overweight) and 73 workers WFH (33 normal-weight and 40 overweight). Physical behaviours were assessed for 7 days using a thigh-worn accelerometer, and compositions were analysed using compositional data analysis approach. Two-way analysis of covariance was used to test differences in behaviours during working days and non-working days, with workplace and weight status entered as between-subject factors, controlling for gender and age. Results and Conclusion: During working days, workers WFH spent more time sitting and in bed, with less temporal variation in behaviours, compared to workers WAO. Overweight workers spent more time sitting and exhibited less temporal variability in sitting behaviours than normal-weight workers. On non-working days, behaviours did not differ significantly by work location, while overweight workers spent more time sitting in total, and less time sitting in short and moderate uninterrupted bouts compared to normal-weight workers. These findings highlight the need for interventions to address prolonged sitting and low variability in behaviours, particularly among workers WFH and overweight workers. Study 3 – Objective: To determine how the number of participants in an overweight and a normal-weight group, the number of measurement days per participant, and balancedness of group sizes influence the accuracy of group differences in 24-hour physical behaviours. Methods: Simulated datasets based on real accelerometry data from 225 office workers (103 overweight and 122 normal-weight) were used to examine 96 (8x4x3) sampling strategies. Trueness (the size of the difference), precision (the 95% confidence interval width), and likelihood of significant group differences were evaluated under varying the total number of participants (8 alternatives: 18, 36, 54, 72, 90, 108, 126, 144), the number of working days per participant (4 alternatives: 1, 2, 3, 4), and the balance between populations (3 alternatives: ratios 1:1 [balanced], 1:2 [unbalanced favouring normal-weight], and 2:1 [unbalanced favouring overweight]). Results and Conclusion: Trueness (i.e., bias) was influenced by the number of measurement days but not by the number of participants. Precision improved with more participants and days, and balanced samples provided more precise estimates. Future studies should carefully balance participant recruitment and measurement days to optimize accuracy and cost-efficiency.