It is well recognized that the workplace is an important setting for physical activity promotion initiatives . This study highlights that it is also a key setting for addressing prolonged sedentary time – an independent risk factor for early death and poor health outcomes. In this sample of Australian employees, sedentary time (derived from accelerometers) comprised more than three-quarters of total work hours and work time also involved a considerable amount of sedentary time that was accrued in prolonged bouts (≥20 or ≥30 minutes), particularly within call-centre employees. Furthermore, the differences that were observed between work days and non-work days (where employees spent a significantly greater proportion of time sedentary and in prolonged sedentary bouts, and proportionately less in light-intensity activity) appeared to be attributable primarily to the differences between work and non-work hours.
Our findings are similar to those of a recent observational study from the USA of 21 desk-bound workers , which found that sitting, assessed using a monitor that incorporates both inclinometers and accelerometers, was higher on workdays than non-work days by 110 minutes per day. The reported difference, equating to approximately 9.9%, was similar to our observed 7.5% difference between work and non-work days.
The observation that employees spent proportionally more time in MVPA on work days compared with non-work days is consistent with several studies [30, 31] that have reported a tendency for sedentary office employees to engage in at least as much physical activity (as measured by step count) on work days as non-work days.
Our analysis of the accelerometer output by hour of the day suggests that differences observed between work days and non-work days are strongly influenced by sedentary and activity patterns during employees’ work hours. Specifically, the time period of 09:00–16:59, when many participants were working, was when significant differences in both sedentary time and time spent in light-intensity activity were observed between work days and non-work days. By contrast, the time periods of 06:00–06:59 and 17:00–18:59 which are typically outside of usual working hours was when MVPA was significantly higher on work days than non-work days. The occurrence of additional MVPA on work days, outside of work hours, may explain why work days, but not work hours involved proportionally more MVPA than non-working days. Additional data collection on how employees’ activity was accumulated, for example, in the form of structured exercising after work or via active transportation could yield important insights relevant to promoting activity with time- and context- specific interventions such as mobile text messaging.
In our pooled study sample, employees were sedentary for an average of 6.6 hours while at the workplace. This is considerably higher than figures based on self-report in Dutch full-time workers (2.7 hours inclusive of work-related travel)  and Australian workers in professional and white collar occupations (3.5 hours and 4.1 hours, respectively) . Our observation that approximately three quarters of work time was spent sedentary is comparable to a recent study of 140 Swedish call-centre operators that used inclinometers (mean 77% vs 75% respectively) . While most self-report studies examining workplace sitting time do not report on employees start and finish work times [11, 14–16], there is evidence to suggest that the high sedentary time seen in our study (when reported as a proportion of work hours) is not entirely due to our sample of employees reporting longer than normal work hours. Professional and white collar workers have been reported to spend an average of 81% of their working hours performing activities of light-intensity or lower ; in the present study, the combined averages for light-intensity (22%) and sedentary time (77%) were 97%. Caution is warranted in making direct comparisons between the studies, as employees were recruited from different occupational groups.
In addition to our sample of workers spending the majority of time at work sedentary, we observed that nearly half of sedentary time at work was accrued in prolonged sedentary bouts of at least 20 or 30 minutes (i.e., 33.5% and 21.5% respectively of total work hours). To date, only one other study has objectively quantified the sitting patterns of office-based workers across an assumed typical work day (9:00–17:00) . This study, which used the ActivPal activity monitor on 83 office workers, found an even higher proportion of sitting time at work (52% and 67%) was accrued in bouts of prolonged sitting that lasted longer than 20 or 30 minutes in duration, respectively.
Our study adds to a growing body of evidence suggesting that the workplace is a key setting for sedentary behaviours. Additional studies are needed to support these findings and to further explore the potential impact of sedentary behaviour at work on activity patterns outside of work. Such information is needed to inform public health guidelines aimed at reducing sedentary time, particularly those strategies that might target the workplace setting. These studies need high-quality assessment of behaviour that occurs during employees actual working hours, e.g. as reported in diaries, rather than assumed working periods of eight hours duration . This will help to avoid potentially substantial misclassification that might occur from MVPA occurring immediately after work, but within these assumed working time periods.
A novel element of our study was the examination of accelerometer-derived sedentary time and physical activity across different workplace settings. Here, variations across the three workplace settings selected were considerable. Not surprisingly, call centre workers were the most sedentary and least physically active during work hours. Call centre employees also accrued more of their sedentary time through prolonged bouts. In contrast, customer service employees had the lowest levels of prolonged sedentary time. These differences likely reflect variations in the opportunities of employees to interrupt sedentary periods through task-based activities. For employees working in settings that afford little or no task-based opportunities to interrupt sedentary time (such as call-centres), alternative options for reducing workplace sedentary time may be required. For example, introducing sit-to-stand workstations into call centres could provide employees with additional opportunities to interrupt prolonged sedentary time by enabling them to transition from sitting to standing and work in an upright posture intermittently throughout the work day.
A strength of our study was the combined use of an accelerometer and a self-report daily diary to provide individually tailored segmentation of sedentary and physical activity time occurring during work and non-work hours on work days. A key limitation of the accelerometer was that it was not able to provide a postural assessment of sitting, which is possible with other activity monitors such as the ActivPAL device [40, 41]; thus some time spent standing still may have been classed as sedentary time. In this study, a widely used cut point of <100 cpm was selected to estimate sedentary time. Whilst this is considered adequate for use in adult studies [26, 36, 42] there is no universally accepted cut point and other cut points, e.g. <150 cpm , have also been advocated. Similarly, light and MVPA were estimated using the well regarded Freedson cut points , although these are also not universally accepted. The selection of a different epoch length, different cut-points, or criteria for determining a bout of prolonged sedentary time would have led to different estimates of the mean amount of each activity and prolonged sedentary time. However, as this would be likely to occur equally for all days and times, our comparisons between work and non-work are unlikely to be affected. As is often the case in such studies, wear time did not appear to cover all waking hours and was estimated indirectly rather than observed, which may have affected results, since more non-wear time tended to be detected outside of work hours than during work hours, and on non-work than work days. Other limitations include our use of a convenience sample, and low numbers of call centre and customer service employees, limiting generalizability to the broader working population. Studies that use more sophisticated technologies cable of differentiating between standing and sitting time, along with probabilistic sampling covering a broader spectrum of employee and job attributes, are needed to better assess sedentary and physical activity time in the workforce and understand differences across workplace settings.