The effect of sit-stand schedules on office work productivity: A pilot study

1 Introduction

Prolonged sedentary time has been associated with various health risks [1]. Routine office or computer work are linked to some public health concerns due to their sedentary nature. To address such risk, office workers are suggested to add regular breaks or to use different working postures throughout a workday, according to expert opinions [2, 3]. Mechanical (manual) or motorized (electric) sit-stand workstations are suggested capable of counteracting the risks arising from prolonged seated work, because such workstations allow operators to perform regular office duties with alternative postures [4–9]. These workstations may be prescribed for employees based on their presumed medical or ergonomics needs. Recently, they are more common in workplaces with the goal of reducing sedentary work in a larger scale [6, 8–13]. However, one of the concerns about adopting alternative workstations is the effect on employee productivity [14].

Previous laboratory studies have demonstrated that cognitive function and motor efficiency are not impacted by the use of sit-stand or active workstations [15–17]. Several recent literature reviews concluded that sit-stand workstations do not result in decreased productivity [14, 18]. While promising, laboratory studies typically use surrogate measures that lack practical office performance applications and hence suffer from this inherited limitation [18]. Furthermore, most office studies about the posture effect on performance rely on self-reports of work postures (sit vs. stand). For example, Garrett et al. [19] did not instruct the users about the sit-stand schedule, and only collected self-reported standing times in the six months of the study period. Davis and Kotowski [4] used a computer program to prompt posture change every 30 minutes. The participants could ignore the messages; therefore, the actual standing time was not known. In this regard, studies that claim the effect of sit-stand workstations on productivity or performance are limited to inaccurate information about sitting and standing durations. The gradient effect of the ratio of sitting to standing on office work productivity or performance is unclear. Therefore, this naturalistic study was conducted to examine the effect of four different controlled sit-stand schedules on office performances. This pilot study was undertaken in an attempt to explore if various computer usage activities and organizational productivity measures would be affected by different sit-stand ratios.

2 Methods

2.2 Study design and protocol

Per organization policy, each participant was equipped with a standard electronic height-adjustable sit-stand desk alongside with one fixed-height desk for the total work surface need in each cubicle. The standalone sit-stand desk had a surface dimension of 147 cm wide and 74 cm deep. Each sit-stand desk was equipped with an electronic control box, a motor actuator, and two lifting columns (Linak, Nordborg, Denmark) via a control panel. The control panel could store three heights in the memory. Prior to the data collection, the sitting and standing heights were set for each individual by professional ergonomists based on the current office ergonomics guidelines (Fig. 1). For each employee, the computer, two monitors, and accessories such as keyboard and mouse, were positioned on this adjustable desk. The other desk was mainly used for non-computer tasks or as spare area for documents.

OPEN IN VIEWER

Fig. 1

Sitting on the left, and standing on the right.

This study employed a full factorial design, where all participants received identical treatments. The only independent variable was the sit-stand ratio per workday. Four different sit-stand ratios (sitting time: standing time = 1 : 1, 2 : 1, 3 : 1, and 7 : 1) were administered in a random order on four different 8-hour workdays. Each 8-hour workday consisted of two coffee breaks (about 15 min each) and one lunch break (about 30 to 60 minutes) to form four sit-stand work periods (cycles), roughly 1.5–2 hours each. Within each cycle, the assigned sit-stand ratio was applied, and sitting always preceded standing (Fig. 2). Experimenters were on-site to remind or enforce the participants to change postures.

OPEN IN VIEWER

Fig. 2

Sit-stand timeline on one exemplar day with a sit-stand ratio of 2:1. Lunch break duration was either 30 or 60 minutes depending on individual-human resource agreement, and not counted toward work hours. The two 15-minute breaks are provided mandatorily and considered part of work. The actual sitting and standing times were all calculated to reflect the actual ratio and remained constant throughout the given day.

Because this was a naturalistic study, the participants were left uninterrupted during work, except for necessary measurement periods as reported in Bao and Lin [20]. The participants were free to move around for personal or business needs. Due to the business nature, there was little day-to-day variation among the office works within individual as confirmed by observations and communications with the participants and the managers. During scheduling, every attempt was made to ensure that the four days were similar.

3 Results

There were no statistically significant sit-stand ratio effects on any daily computer usage indicators. While not statistically significant, the average time using computer, keystrokes, word count, and keyboard errors on a workday were the least when the standing time was the most (1 : 1 ratio) (Table 1).

No obvious relationship could be discerned between standing time and productivity measures. One noteworthy trend about the work performance for the three call center employees was on the time to finish a report (Table 2). The longest sitting duration resulted in the longest average time (4 minutes 27 seconds) to update or to finish a call report than the shortest sitting duration did (2 minutes and 21 seconds). The average duration per call was the shortest at 3 : 55 min on the longest standing schedule (1 : 1 ratio), compared to 4 : 06 on the longest sitting schedule (7 : 1 ratio).

Our results demonstrate that computer usage and job productivity were not affected by the amount of standing time during work. To the best of our knowledge, only three peer-reviewed studies examining the sit-stand workstation effect were conducted in the field collecting realistic work productivity and performance: Davis and Kotowski [4], Garrett et al. [19], and Chau et al. [23]. All of these studies reported participants’ self-reported standing times. In Davis and Kotowski [4], the sit-stand ratio ranged from 7 : 1 to 5 : 1 depending if a reminder software was installed in the participants’ computers. Garrett et al. [19], based on a pilot study, only yielded estimates of a sit-stand ratio at around 4 : 1 among their participants of call center employees. Chau et al. [23] observed the highest standing proportion (close to 2 : 1 sit-stand ratio) among call center employees after an intervention including training and daily email reminders. They did not stratify the sit-stand ratio, and found that the standing intervention, as a sole variable, did not impact productivity in terms of average time per call. Our study extended the sit-stand ratio from 7 : 1 to 1 : 1, and the findings support that standing as much as half of the time on a workday would not result in any performance decline.

Interestingly, several current metrics hint the trend that more standing could be associated with better productivity. Davis and Kotowski [4] reported a statistically insignificant 10% more calls completed among the call center sit-stand workstation users than their conventional sit-only counterparts. As compared earlier, the two levels of groups in their study fell in the equivalent range of the 7 : 1 and 3 : 1 ratios as in the our study. The calls made per hour of our three participants were 11 on 3 : 1 days and 10.3 on 7 : 1 days. That was about 7% increase when they stood longer. The effect was further enhanced to 14 calls on 2 : 1 days, equivalent to a 36% improvement. The other two performance metrics were clearer that more standing resulted in shorter call durations and faster report completions.

Comparable to field studies identified in the literature, work productivity relied on participating employers’ organizational metrics [4, 23]. These objective measures (calls per hour, duration per call, etc.) may not reflect productivity in other subjective or objective dimensions, for example, the difficulty of the calls could not be assessed. Similar to the three previous studies, our data showed variations in productivity in normal working days (standard deviations in Tables 1 and 2). The order of sit-stand ratio was assigned randomly, and the choice of days for a participant was also random. The variation should not affect our results. In a realistic office work environment, the fact that the uncontrolled and uncontrollable variations on the job could influence productivity more than the posture could. It is therefore reasonable to derive that productivity should not be a concerning factor for sit-stand workstation adoption.

One limitation of the current study was that there was not a sitting only “control treatment”. However, there were reasons it was not implemented. First, the highest sitting-standing ratio (7 : 1) resulted in about 105 minutes of sitting and 15 minutes of standing in a two-hour segment. In most of the segments the standing times were actually less. Consider that an eight-hour working day minus the two mandatory 15-minute breaks, the initial study setup, and end-of-the-day study wrap-up, which took about five to ten minutes each, the remaining working time was less than four two-hour segments (Fig. 2). Therefore, in the 7 : 1 schedule the participants had standing times often less than ten minutes per segment. Second, as well known by ergonomics practitioners and researchers that sitting all day is unhealthy, therefore we decided not to include an all-sitting treatment. Third, other activities (breaks, study preparation and wrap-ups) involving posture changes were not avoidable that there could never be a true all-sitting control. We believe that the 7 : 1 schedule was appropriate as a baseline condition for comparison, as also reflected in a population study [24].

We also acknowledge that the sample size of our pilot study participants was limited for the purpose of the present analysis, especially with the work productivity metrics, to draw conclusions with strong statistical evidences. The observations and analysis made in the current pilot study were aimed to provide information for a future larger scale study. In addition, we gained knowledge and identified additional opportunities that might be of use. For example, activity trackers could enhance our understanding of overall posture and activities during breaks and other posture variations not confined at the sit-stand workstations.

Our results indicate that amount of standing in a workday does not affect computer usage. It also did not impact productivity. It may even enhance several indicators of productivity, similar to the effect reported by Garrett et al. [19]. Unfortunately, up to date, there is no common ergonomics recommendation for sit-stand ratio [25]. As indicated in Bao and Lin [20], individual employee preferred different sit-stand ratios during their work hours. Collectively, office employees may wish to stand longer than they currently do [24]. Therefore, employers do not need to concern about employee performance with sit-stand workstation implementations. Providing the necessary furniture, administrative support, and proper training would be able to reduce the sitting time during work hours [26, 27].

Conflict of interest

None to report.