A school-based physical activity program to improve health and fitness in children aged 6–13 years ("Kinder-Sportstudie KISS"): study design of a randomized controlled trial [ISRCTN15360785]

The intervention was a multi-component school-based physical activity program called "Kinder-Sportstudie KISS" (Children's Sports Study). The intervention group received four specific additions to their normal school routine:

1. **Two extra physical education (PE) classes per week** — taught by trained physical education teachers, bringing the total to five PE classes per week (instead of the usual three).

2. **Short physical activity breaks (2–5 minutes each)** — inserted during academic lessons (e.g., math, language classes) to break up sitting time.

3. **Physical activity homework** — children were given assignments to be active outside school hours.

4. **Adaptation of recreational areas around the school** — playgrounds and outdoor spaces were modified to encourage more active play during breaks.

The control group continued with their usual school curriculum (standard PE classes, no extra breaks, no homework, no playground changes).

The primary outcomes were:

Total physical activity (measured by accelerometry)

Aerobic fitness (measured by the 20-metre shuttle run test)

Percent body fat (derived from skinfold measurements)

Quality of life (assessed by the Child Health Questionnaire)

Secondary outcomes included:

Overall fitness (broader fitness battery)

Body composition including fat distribution

Cardiovascular risk factors (blood pressure, blood lipids, glucose)

Psychosocial health

Bone mineral content and density (femur, lumbar spine, total body)

Food intake

The study aimed to recruit children from 15 schools in Switzerland, stratified by geographic region (urban vs. rural) and by age group (1st grade, approximately age 6–7 years; and 5th grade, approximately age 10–13 years). The total target sample was approximately 1,200–1,500 children, though exact enrolment numbers were not finalised at the time of publication. All children in the intervention schools participated because the intervention was integrated into the normal school curriculum — no opt-in consent was required for the intervention itself, though individual consent was needed for blood sampling and DXA scans.

Key inclusion criteria:

Children attending 1st or 5th grade in participating schools

Both urban and rural settings

No specific exclusion criteria reported (the intervention was school-wide)

At least 70% of all children agreed to blood sampling and dual-energy X-ray absorptiometry (DXA) measurements for body composition and bone density.

The study used a comprehensive battery of measurements:

**Physical activity:**

**Accelerometry** — children wore an accelerometer (a small device that records movement intensity and duration) for at least 7 consecutive days at baseline and follow-up. This provided objective data on total physical activity, time spent in moderate-to-vigorous physical activity (MVPA), and sedentary time.

**Aerobic fitness:**

**20-metre shuttle run test** — a maximal running test where children run back and forth between two lines 20 metres apart, following a beep that gets progressively faster. The number of completed shuttles is recorded. This is a validated field test for estimating VO₂max (maximal oxygen uptake).

**Body composition:**

**Skinfold thickness** — measured at four sites (biceps, triceps, subscapular, suprailiac) using calipers. Percent body fat was calculated from these measurements using standard equations.

**Dual-energy X-ray absorptiometry (DXA)** — a low-dose X-ray scan that measures bone mineral content, bone mineral density, and body composition (fat mass, lean mass) for the total body, femur, and lumbar spine. This was optional (70% participation).

**Anthropometrics:**

Height, weight, waist circumference, hip circumference — measured using standard protocols.

Body mass index (BMI) calculated as kg/m².

**Cardiovascular risk factors:**

**Blood pressure** — measured using an automated oscillometric device, taken after 5 minutes of rest.

**Blood samples** — fasting blood draw to measure total cholesterol, HDL cholesterol, LDL cholesterol, triglycerides, glucose, insulin, and other markers.

**Quality of life:**

**Child Health Questionnaire (CHQ)** — a validated parent-reported questionnaire assessing physical and psychosocial well-being. The CHQ yields scores for physical functioning, general health, mental health, self-esteem, and family impact.

**Psychosocial health:**

Self-reported questionnaires on well-being, self-esteem, and peer relationships.

**Food intake:**

Food frequency questionnaire to estimate dietary patterns and energy intake.

**Study design:** This is a **randomized controlled trial (RCT)** — the gold standard for testing causal effects of an intervention. Fifteen schools were randomly assigned to either the intervention group (9 schools) or the control group (6 schools). Randomisation was stratified by geographic region (urban vs. rural) and by age group (1st grade vs. 5th grade) to ensure balance across these important factors.

**Why randomisation matters:** By randomly assigning entire schools (not individual children) to intervention or control, the design reduces the risk that pre-existing differences between groups (e.g., socioeconomic status, baseline fitness, school resources) could explain any observed differences at the end of the study. The stratification by region and age further reduces confounding.

**Blinding:** This was an **open-label (unblinded) study**. Teachers, children, parents, and researchers knew which schools were in the intervention group. This is a major limitation because:

Teachers in intervention schools may have been more motivated to encourage activity (expectancy effects).

Children in intervention schools may have altered their behaviour because they knew they were being studied (Hawthorne effect).

Outcome assessors were not blinded — those measuring fitness, body composition, and blood pressure knew which group children belonged to, which could introduce measurement bias.

**Duration:** The intervention lasted **one academic year** (approximately 9–10 months). Follow-up measurements were taken at the end of the school year. This is a reasonable duration to detect changes in fitness and body composition, but may be too short to see meaningful changes in bone density or long-term cardiovascular risk.

**Statistical approach:** The study used multilevel modelling (hierarchical linear models) to account for the fact that children were clustered within schools. This is critical because children in the same school are more similar to each other than to children in other schools (due to shared environment, teachers, and peer influences). Failing to account for clustering inflates the risk of false-positive findings. The primary analysis compared changes from baseline to follow-up between intervention and control groups, adjusting for baseline values and stratification factors.

**What this design can prove:**

If the intervention group shows greater improvements in physical activity, fitness, body composition, and quality of life compared to the control group, the design can establish **causality** — that the intervention caused those changes (assuming randomisation worked and no major confounds emerged).

The multi-component nature of the intervention means the study can test whether a "real-world" package of school-based changes is effective, but it **cannot** isolate which specific component (extra PE, activity breaks, homework, playground redesign) drove any effects.

**What this design cannot prove:**

It cannot determine the **optimal dose** of each component (e.g., is 2 extra PE classes better than 1? Are 5-minute breaks better than 2-minute breaks?).

It cannot rule out **contamination** — control schools may have adopted some of the intervention strategies if teachers heard about them.

It cannot assess **long-term sustainability** — the study only measured effects after one year, not whether changes persisted after the intervention stopped.

Because schools were the unit of randomisation but children were the unit of analysis, the study has **reduced statistical power** compared to individual-level randomisation (though the multilevel modelling partially addresses this).

**Major methodological weaknesses:**

**No blinding** — this is a significant source of potential bias, especially for subjective outcomes like quality of life and self-reported activity.

**No placebo control** — the control group received no alternative intervention, so any effect could be due to increased attention rather than the specific components.

**Cluster randomisation with only 15 schools** — with only 9 intervention and 6 control schools, there is a risk of imbalance on unmeasured confounds despite stratification.

**Protocol paper only** — no results are reported, so the actual effectiveness of the intervention is unknown from this publication.

**This is a protocol/design paper — no results are reported.** The authors state that "short-term results can be expected by the beginning of 2007." Therefore, there are no findings to summarise. However, the paper does provide preliminary baseline data showing that the enrolled children were representative of Swiss children in terms of sex, socio-demographic status, and BMI.

If you are looking for the actual results of the KISS study, you would need to consult the follow-up publications (e.g., Kriemler et al., 2010, *BMJ* — which did report positive effects on physical activity, fitness, and body fat).

Not applicable — no results are reported in this paper.

**Acknowledged by the authors:**

The study is a protocol paper; no results are yet available.

The authors note that the intervention is multi-component, making it impossible to isolate the effect of any single element.

They acknowledge that school-based interventions face challenges with implementation fidelity (teachers may not deliver all components as intended).

**Critical reader observations:**

**No blinding** — as discussed above, this is a major source of potential bias for all outcomes, especially subjective ones like quality of life.

**Cluster randomisation with few clusters** — 15 schools is a relatively small number for cluster randomisation, increasing the risk of baseline imbalance and reducing statistical power.

**Selection bias** — schools volunteered to participate, so they may be more motivated and have more supportive environments than the average school. This limits generalisability.

**No intention-to-treat analysis described** — the paper does not specify how they will handle children who drop out or do not adhere to the intervention.

**Measurement burden** — children had to undergo blood draws, DXA scans, and wear accelerometers for a week. This may have led to selective non-participation in the optional measurements (only 70% agreed), potentially biasing those sub-analyses.

**No long-term follow-up** — one academic year is too short to assess whether changes in physical activity or body composition are sustained.

**No control for school-level factors** — differences in school resources, teacher enthusiasm, or neighbourhood environment could confound results.

**Industry funding not disclosed** — the paper does not mention funding sources, so potential conflicts of interest are unknown.

For someone running their own n=1 experiment (e.g., a parent trying to increase their child's physical activity, or an adult trying to improve their own fitness):

**What to test:**

Test a **multi-component physical activity intervention** similar to KISS: add 2 extra structured exercise sessions per week (e.g., 45-minute sessions of moderate-to-vigorous activity), insert 2–5 minute movement breaks every 30–60 minutes during sedentary work or study, assign yourself "activity homework" (e.g., 10-minute walk after each meal), and modify your environment to make movement easier (e.g., set up a standing desk, keep resistance bands visible, park farther from destinations).

**Minimum meaningful duration:**

Run the experiment for **at least 8–12 weeks** to see measurable changes in aerobic fitness and body composition. For bone density changes, you would need 6–12 months. The KISS study used one academic year (~9 months), which is a good target for a thorough self-experiment.

**What to measure (specific metrics):**

**Physical activity:** Wear a step counter or accelerometer (e.g., Fitbit, Garmin, or a research-grade device if available). Measure total steps per day and minutes of moderate-to-vigorous activity. Record daily for at least 7 days at baseline and 7 days at follow-up.

**Aerobic fitness:** Perform a timed 1.5-mile (2.4 km) run or a 20-metre shuttle run test (you can do this in a gym or large hallway). Record the time or number of shuttles completed. Repeat at the same time of day under similar conditions.

**Body composition:** Measure waist circumference (at the narrowest point between ribs and hips) and hip circumference. Use skinfold calipers if available (measure triceps, subscapular, and abdominal sites). Alternatively, use a smart scale that estimates body fat percentage (though these are less accurate). Weigh yourself at the same time each morning after voiding.

**Quality of life:** Use a validated questionnaire like the SF-36 (for adults) or the Child Health Questionnaire (for children). Rate your energy levels, mood, and sleep quality on a 1–10 scale daily.

**Cardiovascular risk:** If you have access, measure resting heart rate (first thing in the morning, before getting out of bed) and blood pressure (at the same time each day, after 5 minutes of seated rest).

**Key confounds to control for:**

**Seasonal effects** — physical activity naturally varies with weather and daylight. Run your experiment at the same time of year, or at least control for season in your analysis.

**Diet** — if you increase activity, you may unconsciously eat more or less. Track your food intake (even roughly) to see if dietary changes confound your results.

**Sleep** — poor sleep reduces fitness and increases body fat. Track sleep duration and quality.

**Stress** — high stress can reduce motivation to exercise and increase cortisol, which promotes fat storage. Log your daily stress level.

**Injury or illness** — any missed sessions should be recorded and accounted for.

**Expectation effects** — if you know you're in the "intervention" phase, you may try harder. Consider a blinded design if possible (e.g., have someone else randomise your weeks without telling you which is which).

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

**Physical activity:** An increase of at least 2,000–3,000 steps per day (or 15–20 minutes more moderate-to-vigorous activity per day) compared to baseline.

**Aerobic fitness:** A 5–10% improvement in shuttle run performance (e.g., 2–4 more shuttles on the 20-metre test) or a 30–60 second improvement in a 1.5-mile run time.

**Body composition:** A 1–3% reduction in body fat percentage (if starting overweight) or a 1–3 cm reduction in waist circumference over 8–12 weeks.

**Quality of life:** A 5–10 point improvement on the SF-36 physical functioning or vitality subscales (0–100 scale), or a consistent 1–2 point increase on a 1–10 daily energy/mood rating.

**Cardiovascular risk:** A 3–5 bpm reduction in resting heart rate and a 2–5 mmHg reduction in systolic blood pressure.

**Important caveat for n=1 experiments:** Individual results are highly variable. A "positive result" in a single person does not prove the intervention works for everyone — but it can tell you what works for *you*. Run the experiment for at least two cycles (intervention vs. control) to increase confidence. Consider using a single-case experimental design (e.g., ABAB: baseline, intervention, return to baseline, intervention again) to strengthen causal inference.