The workplace is a key health promotion setting [1,2], with now extensive evidence demonstrating the effectiveness of interventions targeting physical activity in this environment [3]. Generally, physical activity programs in workplaces have had a beneficial impact on health risk biomarkers, work attendance, and job stress [3]. However, a common criticism has been that workplace physical activity interventions typically only reach those who are already fit and motivated to be active [4], and have had negligible impact on reductions in sitting [5,6]. These two issues may be addressed through a shift in focus to organization-wide interventions that have the potential to reach a greater proportion of employees than conventional individually centered approaches, and to interventions that target change across the activity spectrum including: sedentary time (sitting or lying with low energy expenditure, ≤1.5 metabolic equivalents (METS) [7]); light intensity activities (such as standing or incidental movement, >1.5 to <3 METS [8]); as well as moderate- and vigorous-intensity physical activities (MVPA, ≥3 to <9 METS [8]).

Sedentary and light intensity behaviors occupy much of the waking day [8], more than 95% on average in adults. The distribution of time spent between these two behaviors is increasingly being recognized as having potentially important implications for health and well-being [9,10]. Higher levels of light intensity activity are associated with improved blood glucose levels [11], physical health and well-being [12], and decreased depression [13]. Although several behaviors fall within the light intensity spectrum, there is preliminary evidence to suggest that even just postural shifts to standing could have some metabolic benefit [14,15]. In contrast, high levels of daily sitting time have been detrimentally associated with outcomes such as all-cause mortality, cardiovascular disease incidence and mortality, type 2 diabetes incidence [16,17], and cancer incidence and mortality [17], as well as risk indicators for these [18]. In addition, there is growing evidence suggesting that the manner in which sitting is accumulated may be important, with more breaks (or interruptions: either with activity or standing) in sedentary time showing beneficial associations with cardiometabolic indicators including waist circumference [15,19,20], blood glucose [15,19-22], insulin [21,22], and triglycerides [15,18-20].

For office workers, workplace sedentary time is a large contributor to overall sedentary time [23,24], with studies showing that office workers spend, on average, three-quarters of their work hours sitting [23-27]. Given this pervasive nature of sitting in the office, and that office workers are the largest individual occupational sector [28], the office workplace has been identified as a key setting to target reductions in sitting time [29].

Office workers generally have the advantage of being colocated, which means a wide range of influences can be targeted within intervention approaches. Based on workplace health promotion frameworks [1,30] and behavioral models for understanding sitting and activity [31,32], common influences are organizational, environmental, and individual factors, with the notion that individual-level strategies alone are unlikely to be sufficient for sustained behavior change [32]. Organizational-level strategies are seen as particularly important for program implementation [33], to change the culture of an organization [30,34], and for programs to be institutionalized into the organization and sustained [30]. Key organizational-level strategies include having management support for programs [33,34] and implementing the program from within the organization via dedicated onsite staff or workplace ‘champions’ [35].

Despite these frameworks, workplace interventions targeting MVPA have typically targeted the individual and not the organizational- or environmental-level influences [36,37]. Similarly, there have been very few studies that have implemented organizational-level strategies in workplace sitting interventions. The studies that have addressed the organizational level have done so in combination with individual-level strategies (eg, health coaching) and/or physical environment (eg, sit-stand workstation) strategies [27,38-41]. Findings from these multicomponent interventions have shown significant and large reductions in sitting both at the workplace (eg, −125 minutes/8-hour workday [38], −89 minutes/8-hour workday [39]) and across the whole day (eg, −66 minutes/day [40]). While organizational elements were reported as important in these studies [38,39], and a lack of management support reported as a key issue for those studies that have reported less success [27,41], no study to date has identified how much change results from the organizational-level component alone. If found to be effective, an added benefit of this organizational-level approach is that the intervention elements are minimal both in terms of cost and employee burden, which may be beneficial for organizations that have limited employee time or funds.

While organizational-level strategies may be sufficient on their own, it is also important to consider individualized elements to possibly enhance the success of the intervention. Evidence suggests that individual, self-monitoring devices such as pedometers are a common element of successful workplace physical activity interventions [36,42] that can increase activity [43-48] and decrease sedentary time [43,44]. Recent advances in technology have seen the emergence of more sophisticated wearable activity trackers that go beyond just simple step counting to incorporate many of the strategies known to support behavior change [49]. Such strategies, including the provision of detailed, real-time feedback, long-term tracking, prompts, and goal setting, as well as the measurement of multiple behaviors, give activity trackers the potential to be effective behavior intervention tools [49,50]. Indeed, their potential as low-cost behavior change support tools has been recognized by several workplace wellness programs in the United States, where activity trackers are distributed to encourage employees to get healthy and reduce their insurance premiums [51,52].

There has been minimal research on the feasibility, acceptability, and effectiveness of activity trackers as intervention tools. A recent review highlighted the large heterogeneity in the small field of research studies and the mixed quality of the research [53]. However, there was some indication that activity trackers may lead to pre-post physical activity increases and are feasible to wear [53]. There was, however, very limited evidence for long-term physical activity increases [53]. In fact, only four studies out of the 11 in the review evaluated long-term outcomes (≥6 months), with only one study resulting in a significant change in physical activity [53]. There is even less evidence supporting the use of activity trackers to target sitting and standing: the activities that are most common in the office workplace setting [38,39]. As recently highlighted in a review [6], interventions that focus on increasing only physical activity do not necessarily result in changes in sitting; likewise, activity trackers that focus on steps and activity may not necessarily elicit changes in sitting time.

As such, the current study will pilot a waist-worn activity tracker (the LUMOback) that measures and notifies wearers of their sitting time, and also measures activities across the spectrum including number of steps and time spent standing, walking, and running [54]. Given the importance of organizational-level strategies [1,30], and the evidence suggesting self-monitoring tools are effective with additional strategies [42,53,55], the activity tracker will be implemented in conjunction with organizational-level support strategies. To evaluate whether the intervention strategies can also be sustained long-term, assessments will occur in the short (3 months) and long-term (12 months).

The primary aims of this pilot study are to assess two interventions (organizational-level strategies only; organizational-level strategies plus an activity tracker) in office workers regarding their feasibility, acceptability, and short-term (3 month) effectiveness for sitting reduction. The primary effectiveness outcomes are sitting time at work and overall. Secondary aims are to examine the short-term effectiveness of the interventions for other activities, and health- and work-related outcomes; the relative effectiveness of the two interventions for changes in sitting time and other activities; predictors of changes in sitting and activity; and the long-term (12 month) feasibility, acceptability, and effectiveness of the interventions.

A cluster-randomized design will be used, with data collection occurring at baseline, 3, and 12 months (see Figure 1). The cluster-randomized design (with work teams as the unit of clustering) was chosen to minimize contamination among participants from the two intervention groups. The trial has been approved by the University of Queensland Behavioural and Social Sciences Ethical Review Committee (approval number: 2014000089) and is prospectively registered with the Australian New Zealand Clinical Trials Registry [registration number: ACTRN12614000252617].

Participants will be office workers, recruited from work teams from two Australian capital cities (~1000-km apart) within the one organization, Lendlease, an international property and infrastructure group. Inclusion criteria are office-based workers, working at least 0.6 full-time equivalent (ie, at least 60% of full-time work hours) and ambulatory (able to walk at least 10 meters). Exclusion criteria are pregnancy at baseline, allergies to adhesive tape (Opsite and Hypafix required for assessments), and a planned absence from work for longer than 2 weeks during the first 3-month study period. Employees who have their own activity-permissive workstation at the baseline assessment will also be ineligible.

Following manager consent, and prior to the information session, each team will be numbered randomly, using a random number generator, and then listed in numeric order. Teams will then be randomized to either Group 1 (organizational-level strategies only) or Group 2 (organizational-level strategies plus activity tracker), across location strata (site A and B) and team size strata within site A (small/large) using a randomization website [56]. The randomization schedule will be created by a university staff member not involved in the study. A project staff member will then apply the randomization schedule to the list of teams. Neither project staff nor participants will be blinded to participants’ randomization condition.

All participants will receive the organizational-support intervention elements, which are based on the Stand Up Australia intervention [57]. The key intervention message is ‘Stand Up, Sit Less, Move More’ [57], which encourages staff to interrupt sitting at least every 30 minutes with a change in posture; replace sitting with standing or moving (working toward having equal amounts of sitting and upright activities in the day); increase physical activity of any intensity level; and to make these changes across the whole day, both during and outside of work hours.

Participants in Group 2 will be given the LUMOback activity tracker in addition to the organizational-level strategies. The LUMOback is worn around the waist as a belt, collecting information, and providing real-time feedback on sitting, standing, stepping, breaks from sitting, posture, and sleep. The LUMOback assesses activity by inertial sensors, which collect data at a constant 25 Hz [59], and is controlled through a mobile app via a Bluetooth connection that can be used by both iPhone operating system and Android platforms. Up to 3 weeks of data can be collected by the LUMOback before it must be synced with the app, with data transferred between the LUMOback and the app at 600 bytes/second. In the app, participants can view graphs, averages, and goals related to their sitting, standing, stepping, sitting breaks, posture (represented by an avatar, see Figure 2), and sleep. The device can monitor behavior, track attainment of the wearer’s goals, and provide real-time feedback. Participants can use the app to select the LUMOback to vibrate when they are sitting or standing in a ‘poor’ lumbar posture as identified by pelvic tilt angle [59], and to send a push notification to their mobile phone when they have been sitting too long. Vibrations can be chosen to be more or less intense, more or less exact about posture, or turned off; sitting notifications can be selected to range in time between 15 minutes and 2 hours of sitting, or turned off. The app also contains many learning videos about maintaining a good posture.

The LUMOback shows strong correlations and good agreement, in free-living conditions, with measures from the activPAL activity monitor in total time spent sitting (R ²=0.89 [60]; mean absolute percent error (MAPE)=9.5% [61]), standing (R ²=0.86 [60]), and number of steps (MAPE=0.4%, intraclass correlation coefficient (ICC)=0.99 [62]). In laboratory conditions, the LUMOback shows excellent agreement in step counts when tested against the Optogait treadmill test (MAPE=0.2%, ICC=0.99 [62]).

A LUMOback and a four-page instruction booklet that covers an introduction to the device, set-up instructions, and frequently asked questions will be distributed to Group 2 participants by the workplace champion following baseline assessments. On receipt of the device, participants will be asked to download the free app and sync the LUMOback with their phone. Participants will be asked to use their work email to set up their device as this will later be used to request the data from the company LUMO Bodytech. Participants can wear the device as much or as little as they like and their self-directed usage will be tracked. However, participants will be instructed by project staff to wear their LUMOback during the 3-month assessment period both to gather validity data and to estimate the intervention effectiveness while wearing the LUMOback. Participants will be allowed to keep the LUMOback.

Data collection will occur at baseline, 3, and 12 months, and has been approved by the organization to occur during work hours. At each assessment, participants will be asked to wear an activPAL3 activity monitor continuously (24 hours/day) for 7 days, and to complete a concurrent electronic work and sleep diary. They will also be emailed a link to a Web-based questionnaire, using LimeService [63] covering demographic information (baseline assessment only), health- and work-related information (all assessments), and feasibility and acceptability (3- and 12-month assessments). Height and weight will be measured by a trained Lendlease staff member (site A) or project staff member (site B) during a face-to-face baseline assessment session that is expected to take approximately 15 minutes.

Individual qualitative interviews on the feasibility and acceptability of both of the interventions will occur following the 3-month assessment via telephone. Focus groups covering the long-term feasibility and acceptability of the interventions will be conducted as part of the 12-month assessment. Project staff will request LUMOback usage data from Lumo Bodytech every 2 weeks during the initial 3-month intervention period and then periodically throughout the rest of the study.

An overview of the study feasibility, acceptability, and effectiveness outcome measures is provided in Table 1. In addition, a range of measures covering demographics, job, and health characteristics, psychosocial variables and technology confidence will also be collected.

For this pilot study, the sample size was selected based on what the workplace deemed feasible (18 teams). With a usual team size of 10 to 15, and two-thirds expected to be eligible and participate (just over eight per team), we anticipate approximately 150 participants in total, with 75 randomized to each group. This sample size will provide adequate power (≥80% power) with 5% two-tailed significance, to detect short term (3 month) changes within groups (primary effectiveness aim) of our minimum difference of interest (MDI) in our primary outcomes of work and overall sitting time (see Table 2). All calculations are based on no multiple comparison adjustment to significance (in line with the exploratory nature of the study), an anticipated 30% attrition, and strong clustering (ICC=0.1) [58] with an anticipated design effect of 1.48 (ie, 1+0.1×4.83, with an average 5.83 participants per team). Under these same assumptions, power to detect changes equal to the MDIs for the secondary activity outcomes with 5% two-tailed significance are presented in Table 2 , along with the minimum detectable differences (MDDs) between groups for relative effectiveness (secondary aim). This pilot did not power a priori on other research questions, such as health outcomes and long-term changes (12 months). The MDIs and MDDs all reflected modest effects for activity. The assumptions regarding standard deviations (SD), pre-post correlations and clustering were informed by published and unpublished data from previous workplace interventions [38,58] and Australian population data from the AusDiab study.

Statistical analyses will be conducted in SPSS Statistics version ≥22 and Stata version ≥13 with statistical significance set at P <.05, two-tailed. Within-group changes in activity, work and health outcomes (continuous) will be assessed, using linear mixed-models that account for repeated measures and clustering, to determine the effectiveness of each intervention for these outcomes in the short- and long-term. To compare the relative effectiveness of the combined organizational-level and activity tracker intervention to organizational support alone, mixed-models will be used, adjusting for baseline values and potential confounders; these address both repeated measures and clustering. Confounders will initially be chosen a priori from the literature and retained in models if they are associated with the outcome at P <.2. Models will be checked for linearity, normality, and heteroscedasticity. Analyses will follow intention-to-treat principles. Per-protocol analyses will also be conducted to evaluate what the efficacy of the intervention is specifically for those who use the activity tracker, because activity tracker usage is self-directed. Assumptions regarding missing data will be checked and sensitivity analyses will be conducted to evaluate the impact of missing data on findings. Predictors of changes in activity will be evaluated by linear regression, adjusting for baseline values, and correcting for clustering. Predictors will be considered separately and also mutually adjusted. Characteristics of individuals will be considered as potential predictors including demographics (eg, age), psychosocial (eg, preference for sitting at work), health (eg, musculoskeletal problems) and work-related characteristics (eg, perception of relationship with supervisor), and engagement with the intervention (eg, LUMOback usage).

Because we will work in partnership with the organization, it will not be possible to recruit a control group who will not receive the organizational intervention. As such, the effectiveness of each intervention can only be evaluated as per a single-group pre–post design. Accordingly, effectiveness findings will be considered in light of the usual findings within other studies’ control groups. For example, in prior Stand Up Australia research [38,39], there were no significant changes in sitting, standing, and stepping within control groups. Another consideration is contamination. Despite randomization occurring at the team level to reduce contamination, we will not know in advance the degree of interaction between the teams, and therefore the extent to which the participants randomized to Group 1 may receive visual cues to stand up by witnessing LUMOback wearers (Group 2) arise in response to device prompting. Focus groups will attempt to evaluate if any potential contamination occurred. In addition, another consideration is that the Head of Workplace Wellbeing will select the teams for participation. While this is not inconsistent with what might realistically occur in office wellness programs, it may introduce some selection bias, which may limit the generalizability of the results.

The interventions evaluated in this study have the potential to decrease sitting and increase standing and stepping in the office workplace. There has been minimal evaluation of organizational-level strategies alone, and whether these strategies can impact on sitting and activity behaviors when delivered as part of a worksite driven, ‘real-world’ intervention. Furthermore, there is minimal evidence on the feasibility, acceptability, and effectiveness of activity trackers for use in office workers and their effectiveness for reducing sitting and increasing standing and stepping. Despite the detrimental effects of sitting, very few activity trackers measure or target this behavior. If effective, the findings from this research may prompt developers to include sitting measures and prompts in their activity trackers. While only a pilot, this study aims to address these gaps and will provide information to guide future physical activity and sedentary behavior interventions and workplace health promotion programs.