Participatory Workplace Interventions Can Reduce Sedentary Time for Office Workers—A Randomised Controlled Trial

The researchers tested three different participatory workplace interventions against each other (there was no true "do nothing" control group):

**Intervention A ("Active Office Work"):** Workers designed strategies to incorporate movement into productive work tasks. They had access to an "Active Workstation"—an electronically height-adjustable desk with an integrated treadmill (or a stationary cycle ergometer at one site). The recommendation was to use the workstation for short periods, starting at 10 minutes and building up to 30 minutes per session, several times per day. Workers also promoted incidental office activity (e.g., walking to a colleague instead of emailing, standing during phone calls).

**Intervention B ("Traditional Physical Activity"):** Workers focused on increasing activity during non-work time—active transport (walking/cycling to work), lunchtime walks, and breaks between productive work. All participants in this group received a pedometer (Yamax Digi-walker SW700) as a motivational tool to track steps.

**Intervention C ("Office Ergonomics"):** Workers focused on computer workstation setup, "active sitting" (moving while seated), and breaking up computer tasks with micro-breaks. This served as a comparison condition—it was a participatory intervention but did not specifically target reducing overall sitting time or increasing standing/walking.

The primary outcomes were:

Total sedentary time (minutes/day) on work days and during work hours

Sustained sedentary time (bouts of 30+ minutes)

Frequency of breaks in sedentary time (breaks per sedentary hour)

Light intensity physical activity and moderate/vigorous physical activity (MVPA)

All outcomes were measured objectively using accelerometers (ActiGraph GT3X) worn for 7 days at baseline and again after the 12-week intervention period.

**Sample size:** 62 office workers (clerical, call centre, and data processing workers) from 3 large government organisations in Perth, Western Australia. The original target was 120 participants, but recruitment fell short.

**Age range:** 25–59 years

**Inclusion criteria:** Workers performing office-bound duties for 6 or more hours per day, working 4 or more days per week.

**Exclusion criteria:** Unable to wear an accelerometer due to disability or confined to a wheelchair.

**Group sizes:** Active Office (n=19), Traditional Physical Activity (n=14), Office Ergonomics (n=29). The groups were uneven because randomisation was done by clusters (physical proximity groups within each organisation), not by individuals.

**Setting:** Three large government organisations with different work cultures:

- Organisation 1: Data processing of complex files; workers had flexible hours and autonomy over breaks.

- Organisation 2: Call centre; highly controlled—meetings, breaks, and work hours were set by a national office; productivity and call volume were monitored weekly; very little autonomy.

- Organisation 3: Data processing with strict daily quotas; workers sometimes handled calls; breaks were scheduled and controlled; productivity was monitored.

**Primary instrument:** ActiGraph GT3X accelerometer—a waist-worn device that measures acceleration in three axes. It provides objective, time-stamped data on sedentary behaviour (sitting/lying), light activity, and moderate/vigorous activity.

**Wear protocol:** Participants wore the accelerometer for 7 consecutive days during waking hours, both at baseline and after the 12-week intervention. Only data from work days and work hours were analysed.

**Sedentary time definition:** Time spent in activities ≤1.5 METs (metabolic equivalents) while awake—essentially sitting or lying down.

**Sustained sedentary bouts:** Periods of 30+ consecutive minutes of sedentary behaviour.

**Break rate:** Number of times a participant transitioned from sedentary to active per sedentary hour.

**Work hours definition:** Self-reported start and end times of the work day, excluding commute time.

**Validity note:** Accelerometry is considered the gold standard for objective activity measurement in field studies. It avoids the recall bias and social desirability bias of self-reported sitting time. However, it cannot distinguish between sitting and standing still (both are classified as sedentary if movement is below the threshold), and it cannot capture posture (e.g., standing desk use might be misclassified as sedentary if the person stands still).

**Design:** This was a parallel-arm, cluster-randomised controlled trial (RCT). Within each of the 3 organisations, workers were grouped by physical proximity (e.g., same floor, same building, or same suburb). These clusters were then randomly assigned to one of the three interventions using simple randomisation (drawing sealed envelopes from a hat). The allocation ratio was 1:1:1.

**Randomisation:** Randomisation was done at the cluster level (groups of workers), not at the individual level. This is important because workers in the same physical area might influence each other, share resources (like the Active Workstation), or be exposed to the same organisational culture. Cluster randomisation reduces "contamination" between groups but reduces statistical power compared to individual randomisation.

**Blinding:** There was no blinding. Participants knew which intervention they were in (you cannot blind someone to whether they have a treadmill desk or a pedometer). Researchers who analysed the accelerometer data were not blinded to group allocation. This is a significant limitation—expectations and social desirability could influence behaviour beyond the intervention itself.

**Duration:** The intervention period was 12 weeks. Measurements were taken at baseline (before the intervention) and immediately after the 12-week period. There was no follow-up after the intervention ended, so we do not know if any changes were sustained.

**Statistical approach:** The researchers used mixed-effects linear regression models, which account for the clustering of participants within organisations and within groups. They reported percentage changes in sedentary time and absolute changes in break rate, with p-values. They did not report effect sizes (like Cohen's d) or confidence intervals for the primary outcomes in the abstract, but the full text provides some.

**What this design can prove:**

That participatory workplace interventions, as a package, can reduce sedentary time and increase breaks over 12 weeks compared to baseline.

That the effects are consistent across different types of office work (data processing, call centre) and different organisational cultures (flexible vs. highly controlled).

**What this design cannot prove:**

**Causality relative to a true control:** There was no "no intervention" control group. All three groups received a participatory intervention. The comparison is between different types of participatory interventions, not between intervention and doing nothing. So we cannot say that the interventions caused the changes—it could be that simply participating in any health-focused programme (the Hawthorne effect) drove the results.

**Which specific component worked:** Because the interventions were multi-component (meetings, goal-setting, equipment, peer support), we cannot isolate which element was responsible for the changes.

**Long-term sustainability:** No follow-up beyond 12 weeks.

**Health outcomes:** The study measured activity behaviour, not health markers (e.g., blood glucose, cholesterol, blood pressure). We cannot conclude that these activity changes improve health—only that they change behaviour in a direction that is plausibly beneficial.

**Generalisability beyond government office workers:** All participants were from large government organisations in one Australian city. Results may not apply to private sector, manufacturing, or non-office workers.

**Major methodological weaknesses:**

1. **No true control group**—the most critical flaw.

2. **Small sample size** (n=62) with uneven groups (14–29 per group), reducing statistical power and increasing the risk of Type II error (missing real effects) or Type I error (false positives due to small group comparisons).

3. **Cluster randomisation with only 3 clusters per arm**—statistically, this is very weak. With only 3 clusters per intervention, the randomisation may not have balanced confounders (e.g., one organisation might have had more motivated workers).

4. **No blinding**—participants and researchers knew the allocation.

5. **Self-reported work hours**—participants decided when their work day started and ended, which could introduce bias if they changed their reporting after the intervention.

6. **Accelerometer cannot detect standing**—if participants used standing desks but stood still, that time would be misclassified as sedentary.

**Primary outcome—Sedentary time (all participants combined):**

On work days: Sedentary time decreased by 1.6% (p = 0.006). This means that, on average, participants sat for about 1.6% less of their total waking time on work days after the intervention.

During work hours: Sedentary time decreased by 1.7% (p = 0.014). For a typical 8-hour work day, 1.7% is roughly 8 minutes less sitting per day.

**Secondary outcomes—Breaks and activity (all participants combined):**

Breaks per sedentary hour on work days: Increased by 0.64 breaks/hour (p = 0.005). This means participants stood up or moved around about once every 1.5 hours more often than before.

Breaks per sedentary hour during work hours: Increased by 0.72 breaks/hour (p = 0.015).

Light intensity activity during work hours: Increased by 1.5% of work time (p = 0.012). For an 8-hour day, that is about 7 minutes more light activity (e.g., slow walking, standing while moving).

MVPA on work days: Increased by 0.6% (p = 0.012). For a 16-hour waking day, that is about 6 minutes more moderate/vigorous activity per day.

**Comparison between interventions:**

The researchers reported that there were no statistically significant differences between the three intervention groups for any of the primary or secondary outcomes. In other words, the Active Office group did not outperform the Traditional Physical Activity group or the Office Ergonomics group. All three participatory approaches produced similar-sized changes.

**Organisational differences:**

The effects were consistent across the three organisations, despite their very different work cultures (flexible vs. highly controlled). This suggests that participatory interventions can work in diverse settings.

**Important note on what was NOT found:**

There was no significant reduction in sustained sedentary bouts (30+ minutes). The intervention increased the frequency of breaks but did not eliminate long periods of uninterrupted sitting.

There was no significant increase in MVPA during work hours specifically (only on total work days, which includes before/after work).

The changes were modest but meaningful in the context of occupational health:

**Sitting time:** The average reduction was about 1.6–1.7% of total time. For a typical 8-hour work day, that translates to roughly 8–10 minutes less sitting per day. This is equivalent to standing up for about 1 minute per hour across the work day.

**Break frequency:** Participants added about 0.6–0.7 breaks per sedentary hour. If you normally sit for 6 hours at work, that means you took about 4 more breaks across the day—roughly one extra break every 2 hours.

**Light activity:** The 1.5% increase in light activity during work hours is equivalent to about 7 minutes of slow walking or gentle movement per 8-hour day.

**MVPA:** The 0.6% increase on work days is about 6 minutes of brisk walking or stair climbing per day—roughly the equivalent of walking one extra flight of stairs or parking further from the office.

To put this in perspective: Laboratory studies have shown that interrupting sitting every 30 minutes with 2-minute walks can improve post-meal blood sugar regulation by about 30% in overweight adults. The break frequency achieved in this study (roughly one extra break every 1.5 hours) is less intensive than that, but it was achieved in real-world conditions without requiring strict protocols.

The effect is roughly equivalent to what you might get from a simple reminder app that prompts you to stand every 45–60 minutes—but delivered through a participatory process where workers themselves chose the strategies.

**Acknowledged by authors:**

Small sample size (n=62 vs. target of 120), reducing statistical power.

No true control group—all groups received an intervention, so the changes could be due to participation itself (Hawthorne effect) rather than the specific interventions.

Cluster randomisation with few clusters (3 per arm), which may not have balanced confounders.

No blinding of participants or researchers.

Accelerometer cannot distinguish sitting from standing still—standing desk use may have been misclassified as sedentary.

Short intervention period (12 weeks) with no follow-up to assess sustainability.

Self-reported work hours could introduce bias.

**Additional critical notes:**

**Uneven group sizes:** The Office Ergonomics group (n=29) was twice the size of the Traditional Physical Activity group (n=14). This imbalance reduces the reliability of between-group comparisons.

**Selection bias:** Participants volunteered for the study, so they were likely more motivated to change than the average office worker. Results may not generalise to less motivated populations.

**No health outcome data:** The study measured activity, not health. We cannot assume that 8 minutes less sitting per day translates to improved cardiometabolic health—the evidence for health benefits from such small changes is suggestive but not conclusive.

**Funding and conflicts:** The authors declared no funding or competing interests, which is a strength, but the study was small and likely underpowered for between-group comparisons.

**Single country, single sector:** All participants were Australian government employees. Office culture, building design, and social norms around sitting vary widely across countries and sectors.

For someone running their own n=1 experiment to reduce occupational sitting:

### What to test (specific intervention and dose)

**Option A (Active Office):** Use a sit-stand desk or treadmill desk for 10–30 minutes per session, 2–4 times per day. Alternatively, redesign your workflow to include standing tasks (e.g., standing while reading, taking phone calls standing, walking to a colleague instead of emailing).

**Option B (Traditional Activity):** Use a pedometer or step counter with a daily step goal (e.g., 10,000 steps). Focus on walking during breaks, lunch, and active commuting. Take a 5–10 minute walk every 2 hours.

**Option C (Ergonomics + Micro-breaks):** Set a timer to stand up for 1–2 minutes every 30–45 minutes. Use a reminder app (e.g., Stand Up!, Time Out, or a simple phone alarm). Adjust your chair and desk to allow "active sitting" (small movements while seated).

**Which to choose?** The study found no difference between approaches, so pick the one that fits your work style. If you have a sit-stand desk, Option A is natural. If you prefer walking, Option B. If you want the simplest change, Option C (micro-breaks) requires no equipment.

### Minimum meaningful duration

**12 weeks** is the tested duration. However, you might see changes in break frequency within 1–2 weeks. For habit formation, aim for at least 4–6 weeks.

**For health outcomes** (e.g., blood sugar, energy levels, back pain), you may need 8–12 weeks to detect changes.

### What to measure (specific metrics)

**Primary metric:** Total sitting time per work day (in minutes). Use a wearable device (e.g., Apple Watch, Fitbit, Garmin) that tracks standing time and sedentary minutes, or use a simple log: record every time you stand up.

**Secondary metric:** Break frequency—number of times you stand up per hour. Aim for at least 1 break per 30–45 minutes.

**Tertiary metric:** Light activity minutes (e.g., steps taken during work hours,