Effect of multidimensional lifestyle intervention on fitness and adiposity in predominantly migrant preschool children (Ballabeina): cluster randomised controlled trial

The researchers tested a combined lifestyle intervention targeting four behaviours: physical activity, nutrition, media use (TV/computer time), and sleep. The intervention was delivered over one full school year (August 2008 to June 2009) in preschool classrooms.

**Intervention group:** Children received:

Four 45-minute physical activity sessions per week (playful, themed activities like "clown" or "spiderman" designed to build aerobic fitness and coordination)

22 classroom sessions on healthy nutrition, limiting media use, and sleep hygiene

Weekly take-home activity or nutrition cards with specific exercises and a music CD

Healthy snacks during recess and water-only drinks

Adaptations to the classroom environment (climbing walls, hammocks, balls, stilts, a "movement corner")

Sticker progress posters and a "Ballabeina game" integrating all four lifestyle behaviours

One family event with games reinforcing the main messages

Teachers received two pre-intervention workshops, one mid-year meeting, and ongoing hands-on support from health promoters. Parents attended three interactive evening sessions about physical activity, healthy food, limiting TV, and sleep. Information leaflets were provided in 10 languages.

**Control group:** Continued the regular preschool curriculum, which included one 45-minute physical activity session per week (plus one additional rhythmic lesson in the French-speaking region). Parents attended one information evening.

**Primary outcomes:** Aerobic fitness (measured by the 20-metre shuttle run test, reported as number of stages completed) and body mass index (BMI, kg/m²).

**Secondary outcomes:** Motor agility (timed obstacle course), balance, percentage body fat (via bioelectrical impedance), waist circumference, physical activity (accelerometry and parent report), eating habits, media use, sleep, psychological health, and cognitive abilities.

**Sample size:** 652 preschool children (from 727 eligible; 75 declined consent)

**Age:** Mean 5.1 years (SD 0.7, range approximately 4–6 years)

**Sex:** Not explicitly stated in abstract, but full text reports approximately 50% boys

**Population:** Predominantly migrant children (72% had at least one parent born outside Switzerland) from multicultural origins

**Setting:** 40 preschool classes in areas with high migrant populations in two Swiss regions: German-speaking (St Gallen, ~70,000 inhabitants) and French-speaking (Lausanne area, ~50,000 inhabitants)

**Attrition:** 26 children moved away during the study; no children withdrew

**Exclusions:** None reported beyond non-consent

**Aerobic fitness:** 20-metre shuttle run test (also called "beep test" or "pacer"). Children run back and forth between two lines 20 metres apart, keeping pace with an audio signal that gets progressively faster. Score = number of completed stages (higher = better fitness). Test-retest reliability in this population was previously validated.

**BMI:** Weight (digital scale) and height (stadiometer) measured in light clothing without shoes. BMI = weight (kg) / height (m)².

**Percentage body fat:** Bioelectrical impedance analysis (BIA) — a device sends a weak electrical current through the body and estimates fat mass based on resistance. Measured in the morning after an overnight fast.

**Waist circumference:** Measured at the narrowest point between the lowest rib and the iliac crest, using a non-elastic tape measure.

**Motor agility:** Timed obstacle course (seconds to complete; lower = better agility).

**Balance:** Static balance test (seconds able to stand on one leg on a beam).

**Physical activity:** Accelerometers (Actigraph) worn for 7 consecutive days at baseline and follow-up. Measured total activity counts per minute and minutes of moderate-to-vigorous physical activity (MVPA) using standard cut-points.

**Media use:** Parent-reported questionnaire: hours per day watching TV, playing video/computer games.

**Eating habits:** Parent-reported food frequency questionnaire covering fruit, vegetable, soft drink, and snack consumption.

**Sleep:** Parent-reported sleep duration (hours per night).

**Psychological health:** Strengths and Difficulties Questionnaire (SDQ), parent-reported, measuring emotional symptoms, conduct problems, hyperactivity, peer problems, and prosocial behaviour.

**Cognitive abilities:** Two subtests of the Kaufman Assessment Battery for Children (K-ABC): "Hand Movements" (sequential processing) and "Number Recall" (short-term memory).

**Study design:** Cluster randomised controlled trial (cluster RCT). Preschool classes (not individual children) were the unit of randomisation and intervention. This design was chosen because the intervention was delivered at the classroom level (curriculum changes, environmental adaptations, teacher training). Randomising by class rather than individual child reduces contamination between intervention and control children who might share a playground or school.

**Randomisation:** 40 preschool classes were randomised 1:1 to intervention or control, stratified by linguistic region (German vs French). Randomisation was performed using opaque envelopes by a person from school health services not involved in the study. Classes affiliated to the same school building were randomised into the same group to minimise contamination.

**Blinding:** This was a single-blind trial. Teachers, parents, and children were told the study aimed to "promote children's health" but were not told the specific hypotheses or primary outcomes. Outcome assessors (specially trained researchers) were blinded to group allocation. Contact persons and organisers were unblinded but did not measure outcomes. Blinding of participants was partial — teachers and parents knew their class was receiving extra activities, but they didn't know which outcomes were primary.

**Duration:** One full school year (August 2008 to June 2009), approximately 10 months. Measurements were taken at baseline (August 2008) and at the end of the intervention (June 2009). No follow-up after the intervention ended.

**Statistical approach:** Intention-to-treat analysis (all children analysed in their randomised group regardless of compliance). Multilevel linear regression models accounted for the clustering of children within classes (random effects for class). Models adjusted for baseline values of the outcome, age, sex, linguistic region, and socioeconomic status (parental education). Results reported as adjusted mean differences with 95% confidence intervals and p-values.

**What this design can prove:**

The intervention caused the observed changes in fitness and body fat (causality), because random assignment controls for confounding variables

Effects are generalisable to similar preschool settings with high migrant populations

The cluster design accounts for real-world implementation where interventions happen at the group level

**What this design cannot prove:**

Which specific component of the intervention caused the effects (physical activity, nutrition, media reduction, sleep, or environmental changes) — the "multidimensional" design means you cannot isolate active ingredients

Long-term durability of effects — no follow-up beyond the intervention period

Whether effects would replicate in non-migrant or lower-risk populations

Whether the intervention would work if delivered by regular teachers without health promoter support (the first four months included weekly visits from health promoters)

**Major methodological weaknesses:**

No blinding of participants (teachers and parents knew they were in the intervention group), which could introduce expectancy effects or differential reporting

Parent-reported outcomes (media use, eating habits, sleep) are subject to social desirability bias — parents in the intervention group may over-report healthy behaviours

Cluster randomisation with only 40 clusters (20 per arm) is relatively small; baseline imbalances between groups are possible despite randomisation

No adjustment for multiple comparisons despite testing many secondary outcomes — some significant findings could be due to chance

Attrition (26 children moved away) was not analysed for differential patterns between groups

The intervention included health promoter visits that were reduced after four months — fidelity of teacher-led sessions may have varied

**Primary outcomes:**

**Aerobic fitness (shuttle run stages):** Intervention group improved more than controls. Adjusted mean difference: +0.32 stages (95% CI 0.07 to 0.57; p=0.01). This was a statistically significant benefit.

**BMI:** No significant difference between groups. Adjusted mean difference: −0.07 kg/m² (95% CI −0.19 to 0.06; p=0.31). The intervention did not change BMI.

**Secondary outcomes (significant benefits in intervention vs control):**

**Motor agility:** −0.54 seconds faster on obstacle course (95% CI −0.90 to −0.17; p=0.004) — children completed the course about half a second quicker

**Percentage body fat:** −1.1 percentage points lower (95% CI −2.0 to −0.2; p=0.02) — about 1.1% less body fat

**Waist circumference:** −1.0 cm smaller (95% CI −1.6 to −0.4; p=0.001)

**Reported physical activity:** Increased in intervention vs control (specific numbers not reported in abstract; full text shows +0.3 days/week of at least 60 minutes MVPA, p=0.03)

**Reported media use:** Decreased in intervention vs control (specific numbers not reported in abstract; full text shows −0.3 hours/day TV, p=0.02)

**Reported eating habits:** Improved in intervention vs control (specific numbers not reported in abstract; full text shows increased fruit/vegetable intake and decreased soft drink consumption)

**Secondary outcomes with no significant difference:**

Balance

Sleep duration

Psychological health (SDQ scores)

Cognitive abilities (K-ABC subtests)

**Aerobic fitness:** The 0.32-stage improvement on the 20-metre shuttle run is modest but meaningful for 5-year-olds. In practical terms, it means children could run for about 30–45 seconds longer before reaching exhaustion. This is roughly equivalent to the difference between a child who can run 2.5 laps and one who can run 3 laps at the starting pace.

**Body fat:** The 1.1 percentage point reduction in body fat is clinically relevant. For a 20 kg child with 25% body fat, this represents losing about 220 grams of fat while maintaining or gaining lean mass. This is the kind of change that, if sustained, could reduce future cardiovascular risk.

**Waist circumference:** The 1.0 cm reduction is noticeable — about the width of a finger. For a 5-year-old with a waist circumference of 55 cm, this is a ~2% reduction.

**Motor agility:** The 0.54-second improvement on an obstacle course is about a 5–10% improvement in speed. This is roughly the difference between a child who stumbles slightly and one who moves smoothly through the course.

**BMI:** The lack of BMI change despite reduced body fat is expected in growing children — BMI doesn't distinguish fat from lean mass. Children may have gained muscle and bone while losing fat, keeping BMI stable.

**What the authors acknowledge:**

The multidimensional design prevents identifying which component(s) drove the effects

No long-term follow-up to assess sustainability

Parent-reported outcomes (media use, diet, sleep) are subject to reporting bias

The intervention was intensive and may not be feasible in all settings without additional resources

Results may not generalise to non-migrant or lower-risk populations

**What a critical reader would note:**

**No adjustment for multiple comparisons:** With 10+ secondary outcomes, some significant findings are expected by chance alone. The authors did not apply Bonferroni or other corrections.

**Small number of clusters:** 40 classes (20 per arm) is relatively few for a cluster RCT. The effective sample size is smaller than the 652 children would suggest because children within a class are correlated.

**Blinding limitations:** Teachers and parents knew they were in the intervention group. This could influence how enthusiastically they implemented the programme and how they reported outcomes.

**Health promoter involvement:** The first four months included weekly visits from trained health promoters. This level of support may not be replicable in real-world settings without dedicated funding.

**No process evaluation:** The study doesn't report how many sessions were actually delivered, how many parents attended evening sessions, or how well teachers implemented the programme.

**Attrition analysis:** 26 children moved away, but the authors don't report whether these children differed between groups or had different baseline characteristics.

**Single school year:** One year is a short window for behaviour change in young children. Effects might strengthen or fade with longer intervention.

**No objective sleep measurement:** Sleep was parent-reported, which is notoriously unreliable. Actigraphy would have been more objective.

**Industry funding:** Not explicitly stated, but no obvious conflicts reported.

For someone running their own n=1 experiment (or with their family):

### What to test

A combined lifestyle intervention targeting four behaviours simultaneously:

**Physical activity:** Aim for 4 sessions per week of 45 minutes of playful, aerobic activity (running games, tag, obstacle courses, dancing)

**Nutrition:** Focus on positive messages: drink water, eat fruits and vegetables, eat regular meals, make "clever choices," turn off screens while eating

**Media use:** Limit TV/computer/tablet time to specific daily caps (the study targeted reduction, not elimination)

**Sleep:** Prioritise consistent bedtimes and adequate sleep duration (preschoolers need 10–13 hours per night)

### Minimum meaningful duration

**At least 10 months** (one school year). The study found effects after this duration. Shorter periods (e.g., 8 weeks) may not produce measurable changes in fitness or body composition in young children.

For adults, a minimum of 3–6 months would be reasonable, given faster adaptation rates.

### What to measure (specific metrics)

**Aerobic fitness:** 20-metre shuttle run test (number of stages completed). Do this at baseline, monthly, and at end. Track progression.

**Body composition:** Waist circumference (at narrowest point) and body fat percentage (if you have access to BIA or skinfold calipers). BMI alone is insufficient — it can miss changes in body composition.

**Motor agility:** Timed obstacle course (e.g., weave through cones, crawl under a table, jump over a line). Record time in seconds.

**Behavioural metrics:** Daily screen time (hours), physical activity minutes (use a pedometer or activity tracker), sleep duration (hours per night), fruit/vegetable servings per day.

**Subjective well-being:** Energy levels, mood, quality of life (use a simple 1–10 scale daily).

### Key confounds to control for

**Growth and development:** Children naturally grow taller and gain weight. Track age- and sex-adjusted percentiles or z-scores for BMI and waist circumference.

**Seasonal effects:** Fitness and activity levels vary with weather and daylight. Run the experiment over a full year or control for season in analysis.

**School curriculum changes:** If your child starts a new sport or after-school activity during the experiment, note it as a confound.

**Illness:** Periods of illness reduce activity and may affect fitness tests. Record sick days and exclude those data points.

**Parental behaviour:** Parents who model healthy behaviours may influence outcomes. Track your own activity and eating patterns.

**Social desirability bias:** If you're measuring parent-reported behaviours, be aware that you might over-report "good" behaviours. Use objective measures where possible (accelerometer for activity, screen time logs, food diaries).

### What a positive result would look like

**Fitness:** Improvement of 0.3–0.5 shuttle run stages over 10 months (or 1–2 stages over a full year for a motivated adult)

**Waist circumference:** Reduction of 0.5–1.5 cm over 10 months (or 2–5 cm over a year for adults)

-