Stress and Work Performance Responses to a Multicomponent Intervention for Reducing and Breaking up Sitting in Office Workers: A Cluster Randomized Controlled Trial.

The researchers tested a multicomponent intervention called "Beat the Seat" designed to reduce and break up sitting time in office workers. The intervention combined three levels of strategies:

**Individual-level strategies:**

A health check report with personalised feedback

Written information about the risks of prolonged sitting and guidance on setting step goals

Sticky notes and prompt cards placed in workspaces to encourage standing and moving

Computer software and/or a phone app that delivered reminders to break up sitting

Prompt posters displayed around the workplace

Weekly one-to-one health coaching phone calls using motivational interviewing techniques

**Organisational-level strategies:**

An educational presentation and brainstorming workshop about health risks of excessive sitting and strategies for change

A step challenge where participants received pedometers and entered daily steps onto a virtual leaderboard, with weekly spot prizes

**Environmental-level strategies:**

Participants were asked to modify their workspaces based on ideas from the brainstorming session (e.g., identifying areas for non-computer standing work, removing or relocating personal printers and bins to create walking distance)

The comparator was a control group that continued their usual work practices with no intervention.

**Primary outcomes:**

Physiological stress measured by cortisol concentrations (including the cortisol awakening response and diurnal pattern)

Perceived stress measured by questionnaire

**Secondary outcomes:**

Work-related outcomes (vigour, cognitive liveliness, work engagement, occupational fatigue)

Quality of life

The study included 89 office workers (mean age approximately 45 years, range 18–70) recruited from two office buildings of a national construction, services, and property group organisation in Bedfordshire, UK. Participants were clustered into 12 worksite divisions (clusters), with 6 clusters randomised to the intervention group and 6 to the control group.

**Inclusion criteria:**

Aged 18–70 years

Self-reported sitting for ≥75% of their work day

Worked at the same desk for ≥3 days per week with designated phone and internet access

Able to stand and walk unassisted

**Exclusion criteria:**

Unable to communicate in English

Worked night shifts

Were pregnant

Had planned absence for >2 weeks during the study

Participants were predominantly white-collar office workers. The sample was approximately 60% female. All participants received a £5 shopping gift voucher at each measurement time point as an incentive.

**Physiological stress (cortisol):**

Saliva samples were collected using Salivette collection devices (Sarstedt, Germany)

Participants collected samples at home on two consecutive weekdays at baseline and again at 8 weeks

Sampling times: immediately upon waking, 30 minutes post-waking, 60 minutes post-waking, and then at 3, 6, 9, and 12 hours post-waking

Samples were stored in home freezers, transported to the lab on ice, centrifuged, and frozen at -80°C

Cortisol concentrations were analysed using a commercially available enzyme-linked immunosorbent assay (ELISA) kit

Outcomes derived: cortisol awakening response (CAR, calculated as the increase from waking to 30 minutes post-waking), total daily cortisol output (area under the curve with respect to ground, AUCg), and diurnal cortisol slope (rate of decline across the day)

**Perceived stress:**

Perceived Stress Scale (PSS-10): 10-item questionnaire, scores range 0–40, higher scores indicate greater perceived stress

**Work-related outcomes:**

Vigour: measured using the vigour subscale of the Profile of Mood States (POMS), higher scores indicate more vigour

Cognitive liveliness: measured using a subscale from the Work-related Quality of Life scale, higher scores indicate better cognitive function

Work engagement: measured using the Utrecht Work Engagement Scale (UWES-9), 9 items measuring vigour, dedication, and absorption

Occupational fatigue: measured using the Occupational Fatigue Exhaustion/Recovery Scale (OFER-15), which assesses chronic fatigue, acute fatigue, and inter-shift recovery

**Quality of life:**

Measured using the EQ-5D-5L, a standardised measure of health-related quality of life covering mobility, self-care, usual activities, pain/discomfort, and anxiety/depression

**Demographic and anthropometric measures:**

Age, sex, marital status, ethnicity, smoking status, education level via questionnaire

Height and weight measured using a portable stadiometer and Tanita BC-418 bioelectrical impedance device

**Study design:** This was a cluster randomised controlled trial (RCT). Cluster randomisation means that groups of people (in this case, worksite divisions) were randomised together rather than individual participants being randomised separately. This design was chosen to minimise contamination between groups—if individuals in the same division were randomised to different conditions, they might share strategies or influence each other's behaviour.

**Randomisation:** Following baseline measures, 12 clusters were randomised 1:1 to intervention or control using an online randomisation tool (www.randomization.com). A single block randomisation plan was generated, and clusters were matched to the plan using a list randomiser (www.random.org). The research team and participants were blinded to group allocation until baseline measures were completed. After allocation, only the team members coordinating intervention delivery knew group assignments; all other researchers remained blinded throughout the study.

**Duration:** The intervention lasted 8 weeks. Outcomes were measured at baseline (before randomisation) and at 8 weeks (post-intervention).

**Blinding:** Participants and researchers were blinded until baseline measures were completed. After that, participants knew their group assignment (the intervention involved active participation, so true blinding was impossible). However, researchers analysing the data remained blinded to group allocation.

**Statistical approach:** The primary analysis used linear mixed models with group × time interaction effects, which accounts for the clustering of participants within divisions. This is appropriate for cluster RCTs because it adjusts for the fact that people within the same division may be more similar to each other than to people in other divisions. The study was described as exploratory because the sample size was calculated based on the primary outcome of the overall trial (sitting time reduction), not on the stress and work performance outcomes reported here.

**What this design can and cannot prove:**

**Can prove:**

Causal effects of the intervention on outcomes, because randomisation balances known and unknown confounders between groups

Effects at the cluster level (i.e., the intervention works when delivered to entire work divisions)

Changes over time within and between groups

**Cannot prove:**

Individual-level causal effects (because randomisation was at the cluster level, individual-level inferences are weaker)

Long-term effects beyond 8 weeks (no follow-up after the intervention ended)

Which specific component of the multicomponent intervention was responsible for any effects (the design cannot isolate individual strategies)

Generalisability to other populations (e.g., different industries, countries, or job types)

**Major methodological weaknesses:**

The sample size was not powered for the stress and work performance outcomes, making these analyses exploratory

No blinding of participants after allocation (participants knew whether they were in the intervention or control group, which could influence self-reported outcomes through expectancy effects)

Cortisol data collection relied on participants self-collecting samples at home, which introduces potential variability in timing and compliance

The study had no long-term follow-up, so it is unknown whether effects persist or decay after the intervention stops

The control group received no active intervention, so any effects could be due to attention/engagement rather than the specific sitting reduction strategies

**Primary outcomes (physiological and perceived stress):**

**Total daily cortisol output (AUCg):** The intervention group showed lower cortisol concentrations over the day at 8 weeks compared to controls, with a large effect size but non-significant result (difference = -0.85 nmol·L⁻¹; 95% CI -1.70 to 0.03; p = 0.06; Cohen's d = 0.79). This means the intervention group had approximately 0.85 nmol/L lower average cortisol across the day, but this difference could have occurred by chance.

**Cortisol awakening response (CAR):** No significant group × time interaction was found for the CAR (the rise in cortisol from waking to 30 minutes post-waking). The intervention did not appear to alter this specific aspect of cortisol dynamics.

**Diurnal cortisol slope:** No significant group × time interaction was found for the rate of cortisol decline across the day. The intervention did not change the pattern of cortisol secretion throughout the day.

**Perceived stress (PSS-10):** No significant group × time interaction was found. The intervention did not significantly change how stressed participants felt compared to controls.

**Secondary outcomes (work performance and quality of life):**

**Vigour:** The intervention group reported significantly higher vigour at 8 weeks compared to controls (p ≤ 0.05). This means participants felt more energetic and lively.

**Cognitive liveliness:** The intervention group reported significantly higher cognitive liveliness at 8 weeks compared to controls (p ≤ 0.05). This means participants felt more mentally alert and engaged.

**Work engagement (UWES-9):** No significant group × time interaction was found. The intervention did not significantly change overall work engagement.

**Occupational fatigue (OFER-15):** No significant group × time interaction was found for chronic fatigue, acute fatigue, or inter-shift recovery.

**Quality of life (EQ-5D-5L):** No significant group × time interaction was found.

**Feasibility outcomes:**

Cortisol sample collection was feasible in the workplace setting, with good compliance rates. Participants were able to collect saliva samples at the required times, and samples were successfully analysed.

The most notable finding was the trend toward lower daily cortisol output in the intervention group. The effect size (Cohen's d = 0.79) is considered large by conventional standards (0.2 = small, 0.5 = medium, 0.8 = large). To put this in context:

A reduction of 0.85 nmol/L in average daily cortisol is roughly equivalent to the difference between a person with low chronic stress and someone with moderate chronic stress in some population studies

For comparison, a single session of moderate-intensity exercise can temporarily increase cortisol by 5–15 nmol/L, so the reduction seen here is modest but potentially meaningful if sustained over time

The effect is similar in magnitude to what some studies have found with mindfulness-based stress reduction programmes

For the work performance outcomes, the improvements in vigour and cognitive liveliness were statistically significant but the paper does not report exact effect sizes for these outcomes. Based on the p-values (≤0.05), these effects were moderate enough to be unlikely due to chance alone.

**Acknowledged by authors:**

The study was exploratory and not powered for stress and work performance outcomes

The sample size was small (89 participants across 12 clusters)

The study duration was only 8 weeks, which may be insufficient to see changes in cortisol dynamics

Cortisol data collection relied on participant compliance with home sampling protocols

The intervention was multicomponent, so individual components cannot be evaluated separately

**Critical reader observations:**

No blinding of participants after allocation, which could bias self-reported outcomes (vigour, cognitive liveliness) through expectancy effects

The control group received no active intervention, so the "attention effect" of weekly coaching calls and workshops cannot be separated from the specific sitting reduction strategies

The study was conducted in a single organisation (construction/services company), limiting generalisability to other industries or job types

The cortisol finding was non-significant (p = 0.06), meaning there is a 6% probability the result is due to chance—conventionally, results are considered significant only if p < 0.05

No correction for multiple comparisons was applied, increasing the risk of false-positive findings

The study did not measure whether participants actually reduced sitting time (this was reported in a separate paper), so the link between sitting reduction and stress outcomes is indirect

The £5 voucher incentive may have introduced selection bias (people more motivated by small rewards may differ from the general population)

No long-term follow-up to assess whether effects persist or decay

For someone running their own n=1 experiment:

**What to test:**

Test a multicomponent approach to reducing sitting time: use a combination of (1) self-monitoring (track sitting time with a timer or app), (2) environmental changes (move your bin, printer, or water bottle away from your desk so you have to walk to use them), (3) regular prompts (set a timer to stand or walk for 2–5 minutes every 30–60 minutes), and (4) social accountability (involve a colleague or use a step challenge)

The specific dose tested here was: reduce prolonged sitting by ~39 minutes per work shift, increase sit-upright transitions by ~8 per shift, and increase stepping time by ~12 minutes per shift

**Minimum meaningful duration:**

Run the experiment for at least 8 weeks. The study found effects at 8 weeks, but cortisol changes may take longer to emerge. Consider running for 12 weeks if possible

Measure outcomes at baseline (before starting), at 4 weeks (to check for early effects), and at 8 weeks (primary endpoint)

**What to measure:**

**Physiological stress:** If you have access to saliva cortisol testing kits (available online from companies like iHealth or through your doctor), collect samples at waking, 30 minutes post-waking, and at 3, 6, 9, and 12 hours post-waking on two consecutive days at baseline and at 8 weeks. Track the total daily output and the morning rise

**Perceived stress:** Use the Perceived Stress Scale (PSS-10), a free, validated 10-item questionnaire. Score range 0–40, lower is better

**Work performance:** Track daily ratings of vigour (energy level on a 1–10 scale) and cognitive liveliness (mental clarity on a 1–10 scale) using a simple diary or app

**Sitting time:** Use a timer or app to track how many minutes per day you spend sitting, and how often you break up sitting (stand or walk for at least 1 minute every 30 minutes)

**Confounders to track:** Sleep quality (hours and quality rating), caffeine intake (cups per day), alcohol consumption (units per day), exercise outside work (minutes per day), and major life stressors (e.g., deadlines, personal events)

**Key confounds to control for:**

**Cortisol has a strong diurnal rhythm:** Always collect samples at the same times of day, relative to waking. Waking time itself affects cortisol, so keep waking time consistent (±30 minutes) across measurement days

**Caffeine and food:** Avoid caffeine and food for at least 1 hour before cortisol sampling, as both can affect cortisol levels

**Exercise:** Avoid vigorous exercise in the 24 hours before cortisol sampling, as exercise can elevate cortisol for hours

**Menstrual cycle:** If you are a woman, cortisol varies across the menstrual cycle. Try to measure at the same phase of your cycle at baseline and follow-up

**Seasonal effects:** If possible, run the experiment during a period with consistent daylight hours, as seasonal changes can affect cortisol

**Expectancy effects:** You know you're trying to reduce sitting, which could bias your self-reported outcomes. Consider using objective measures where possible (e.g., a step counter or sitting timer app that logs data automatically)

**What a positive result would look like:**

A reduction in total daily cortisol output of approximately 0.8–1.0 nmol/L (about 10–15% decrease from your baseline average)

A decrease in perceived stress score