Reducing office workers’ sitting time: rationale and study design for the Stand Up Victoria cluster randomized trial

This paper is a study protocol, meaning it outlines the design and methods of a planned research study, rather than presenting results. The researchers designed a study to test a multi-component intervention aimed at reducing sedentary behavior in office workers.

The intervention consisted of a 3-month program combining three levels of strategies:

1. Organizational strategies: These would involve workplace policies or culture changes to support less sitting. While not explicitly detailed in the abstract, such strategies often include management support, communication campaigns, and encouraging walking meetings or standing breaks.
2. Environmental strategies: The key environmental change was the provision of height-adjustable workstations (sit-stand desks) to allow workers to easily alternate between sitting and standing.
3. Individual behavioral strategies: These would likely involve education, coaching, or prompts to encourage workers to use their sit-stand desks and break up prolonged sitting.

The comparator was a "usual practice control group," meaning participants in this group would continue their normal work routines without the intervention.

The outcome measures were comprehensive, covering activity, health, and work-related aspects:

• Primary outcome: Workplace sitting time (total minutes/hours spent sitting at work).
• Secondary activity outcomes:
  • Prolonged sitting time (sitting accrued in unbroken bouts of 30 minutes or more).
  • Standing time at work.
  • Sit-to-stand transitions (how often workers changed posture).
  • Moving time at work.
  • Non-workplace activity (sitting, standing, and moving across the whole day, including outside of work).
• Cardio-metabolic biomarkers and health indicators:
  • Fasting glucose (blood sugar).
  • Lipids (cholesterol and triglycerides).
  • Insulin (hormone regulating blood sugar).
  • Anthropometric measures (body measurements like weight, height, waist circumference).
  • Blood pressure.
  • Musculoskeletal symptoms (e.g., back or neck pain).
• Work-related outcomes:
  • Presenteeism (being at work but not fully productive due to health issues).
  • Absenteeism (days missed from work).
  • Productivity.
  • Work performance.
• Other evaluations: The study also planned to assess the incremental cost-effectiveness of the intervention and identify factors (mediators and moderators) that influence the intervention's impact.

The study aimed to recruit a total of 320 participants (160 per arm) from 16 different worksites within a single large Australian government organization (Department of Human Services, DHS).

The target population was office-based workers aged 18–65 years who worked at least 0.6 full-time equivalent hours (FTE). Participants also needed to speak English and have designated access to a telephone, internet, and a desk within their workplace.

Exclusion criteria included:

• Being pregnant.
• Being non-ambulatory (unable to walk).
• Having a pre-existing musculoskeletal complaint that might be exacerbated by standing.
• Having a planned absence from work for more than 2 weeks or a planned relocation to another workplace during the 3-month intervention period.

The study was conducted in Victoria, Australia.

The researchers planned to use a combination of objective and self-reported measures:

• Objective activity measurement: The primary method for assessing workplace sitting, standing, and moving time was the activPAL3 activity monitor. This is a small, lightweight device typically worn on the thigh, which accurately distinguishes between sitting/lying, standing, and stepping. It provides precise data on total time spent in each posture, as well as the duration and frequency of bouts (e.g., prolonged sitting bouts of ≥30 minutes) and the number of sit-to-stand transitions. This device was to be worn at baseline, 3-months (post-intervention), and 12-months (follow-up).
• Cardio-metabolic biomarkers: These would be assessed through blood tests (for fasting glucose, lipids, and insulin) and physical measurements (for anthropometrics like weight, height, waist circumference, and blood pressure).
• Musculoskeletal symptoms: These would likely be assessed through self-report questionnaires or standardized symptom scales.
• Work-related outcomes: These would also be measured using self-report questionnaires or validated scales for presenteeism, absenteeism, productivity, and work performance.

The Stand Up Victoria study was designed as a two-arm cluster-randomized controlled trial (RCT). This means that instead of randomizing individual workers, entire worksites (clusters) were randomized to either the intervention group or the usual practice control group. There were 16 worksites in total, with an aim to recruit 160 participants per arm (intervention vs. control).

Randomization: The worksites were the unit of randomization. This approach is crucial when an intervention is delivered at a group or environmental level, as it prevents "contamination" that might occur if individuals within the same workplace were randomized separately (e.g., a control participant seeing their colleague get a sit-stand desk). By randomizing entire worksites, the researchers aimed to ensure that all participants within a given site received the same condition, thus maintaining the integrity of the intervention.

Blinding: The protocol does not explicitly state that participants or those delivering the intervention were blinded. Given that the intervention involved providing height-adjustable workstations and behavioral strategies, it would be impossible to blind participants to their group assignment. However, it is common practice in such trials to blind the outcome assessors (the people collecting and analyzing the data) to the participants' group assignments to prevent bias in measurement. The protocol does not specify if this was done. The lack of participant blinding means that participants in the intervention group would know they are receiving a special program, which could lead to a "Hawthorne effect" (where participants change their behavior simply because they are being studied).

Washout periods: Not applicable for this parallel-group RCT design, as participants remain in their assigned group throughout the study.

Duration: The intervention itself was planned for 3 months. Participants were assessed at three time points:

1. Baseline: Before the intervention began.
2. 3-months: Immediately after the intervention concluded.
3. 12-months: A follow-up assessment to determine if any changes were maintained 9 months post-intervention.

Statistical approach: The protocol mentions that the study would evaluate incremental cost-effectiveness and identify mediators and moderators of change. For a cluster RCT, specific statistical methods are required to account for the clustering of participants within worksites. Standard statistical tests that assume independent observations would be inappropriate and could lead to inflated p-values (making results seem more significant than they are). The researchers would need to use multi-level modeling or generalized estimating equations (GEE) to properly analyze the data, though the abstract does not detail the specific statistical models.

What this design can and cannot prove:

• Can prove: A cluster RCT, especially with a large sample size and objective measures, is a strong design for establishing a causal link between the multi-component intervention and changes in sedentary behavior and related health/work outcomes in a real-world office setting. By randomizing worksites, it helps ensure that, on average, the intervention and control groups are similar at baseline in terms of unmeasured factors, strengthening the ability to attribute any observed differences to the intervention. The 12-month follow-up is crucial for assessing the sustainability of the intervention's effects.
• Cannot prove: While it can show the effectiveness of the overall multi-component intervention, it cannot isolate the effects of individual components (e.g., whether the sit-stand desk alone was more impactful than the behavioral strategies or organizational support). This design also cannot definitively prove that reducing sitting time causes improvements in cardio-metabolic biomarkers, as other factors could be at play, though it aims to control for many. The lack of participant blinding means that some observed effects could be partly due to participants' awareness of being in the intervention group.

Major methodological strengths (as highlighted by the authors):

• This was planned as the first cluster-RCT to evaluate a multi-component intervention for reducing prolonged workplace sitting in office workers.
• Objective measurement of activity using the activPAL3 monitor is a significant strength, as it provides more accurate data than self-report, which is prone to recall bias.
• Assessment of the intervention's impact on **cardio-met