Employment, Social Affairs & Inclusion

Join the Healthy Boat – Primary School

Evidence level:
 
Evidence of effectiveness:
? - 0 + ++
Transferability:
? - 0 + ++
Enduring impact:
? - 0 +

Overview

‘Join the Healthy Boat – Primary School’ is a school-based intervention targeting children in grades 1 to 4 (ages 6-10). The intervention was designed, implemented, and evaluated by a research group at Ulm University from 2010 to 2011, in collaboration with schools across the region of Baden-Württemberg in Germany. It is delivered by trained teachers throughout the academic year as part of the existing school curriculum, and consists of face-to-face classroom modules, guided physical activities during recess, and media-based homework assignments to be completed with parents in the children’s homes. The intervention aims to prevent childhood obesity by educating children on the risks of unhealthy leisure and eating habits, as well as informing them of the various foods, drinks, and recreational activities that are consistent with a healthier and more active lifestyle.

Practice Category

  • Supporting Parenting and Assisting with Childcare
  • Helping Vulnerable Children
  • Facilitating Positive Transitions to Adulthood

Recommendation pillar

  • Improve education systems’ impact on equal opportunities
  • Enhance family support and the quality of alternative care settings
  • Support the participation of all children in play, recreation, sport and cultural activities

Countries that have implemented practice

  • Germany

Age groups

  • Middle Childhood (age 6 to 12)

Target groups

  • Children
  • Parents

Years in operation

  • 2009 - Still operating

Scope of practice

  • Regional level

Type of organisation implementing practice

  • State/district or other sub-national government
  • Public education organisation

The intervention was developed, implemented, and evaluated by a research team at Ulm university in Germany. It is now managed by the Baden-Württemberg Foundation in Germany, which is dedicated to promoting wellbeing for citizens in the state of Baden-Württemberg.

Rationale of practice

Through complete integration into the existing, year-long school curriculum and the involvement of parents in homework assignments, ‘Join the Healthy Boat’ aims to be an immersive program that makes lasting changes to children’s health-related knowledge and behaviours. The three main goals of the intervention are to increase physical activity, decrease consumption of sugar-sweetened beverages, and decrease time spent using screen media. School-based modules are used to educate children on the importance of healthy eating and exercise, how to distinguish healthy foods and drinks from unhealthy ones, and how to get adequate exercise through sports, games, and other activities. Homework modules are used to educate parents and children on the home-based changes that can be made to improve the health of all family members, such as eating breakfast, limiting the time spent using screen media, and ensuring that while the presence of unhealthy foods and drinks in the home is reduced, healthy alternatives are also made available.

Mode of delivery

  • Group sessions
  • Face-to-face

The intervention uses both face-to-face and media-based channels. Teachers deliver the intervention to children through education in classrooms and guided physical activities during recess, whereas media-based homework assignments are used to consolidate the intervention content in children’s homes and families outside of regular school hours.

Delivery dosage

  • Frequency: Between 2-6 times a week
  • Duration: More than 1 hour sessions

The intervention lasts throughout the school year and is delivered through a range of weekly and daily activities including classroom modules, twice-daily physical exercises during recess, and homework activities. As a result, the duration and frequency of the intervention varies depending on the components involved and is not pre-defined.

Location of practice

  • School based
  • Family / home based

The face-to-face components of the intervention are delivered by teachers in a classroom setting, while the media-based components are designed for the children to complete alongside their parents, thus extending the location of the intervention activities into the children’s families and homes.

Evidence of effectiveness

One large cluster-randomised trial of the ‘Join the Healthy Boat – Grades 1-2’ curriculum (DRKS-ID: DRKS00000494), referred to as the Baden-Württemberg trial, was conducted between September 2010 and November 2011 in 86 schools across the state of Baden-Württemberg in Germany. At baseline, there were 44 schools representing 80 classrooms in the intervention group (IG) and 42 schools representing 74 classrooms in the control group (CG). Baseline data was collected from 1,947 pupils in total. Post-intervention measures were assessed a year later. The number of pupils providing follow-up data varied depending on the outcome measure.

The intervention was assessed using a range of objective, expert-assessed, computer-assessed, and parent-reported outcome measures capturing children’s body measurements, cognitive abilities, motor abilities, physical activity levels, screen media usage, and dietary habits. This resulted in a large data set that was analysed for intervention effects in six separate articles.

These analyses found that at one-year follow-up, children in the IG showed a statistically significant improvement in conditional skills and a significantly smaller reduction in flexibility compared to children in the CG. Some positive intervention effects were also found when looking at specific sub-groups. Primary 1 children in the IG, for example, took sick days less frequently than Primary 1 children in the CG. Significantly fewer girls, children with non-migrant backgrounds, and children of parents who had lower educational attainment in the IG consumed an hour or more of screen media per day than children in the CG. A significantly smaller proportion of Primary 2 children in the IG reported skipping breakfast often/always than Primary 2 children in the CG. When looking specifically at children with migration backgrounds, girls in the IG were found to have better flexibility than girls in the CG, whereas boys in the IG were found to have better conditional skills than boys in the CG.

Transferability

  • Practice has not been evaluated in multiple populations

Implementation costs

  • Implementation cost information is available

A cost-effectiveness evaluation by Kesztyüs et al (2017) judged the total cost of training 81 teachers to be €36,506.41 (2011 rates). The cost of the intervention per child was €25.04, whereas the cost per case of abdominal obesity avoided during the study period was estimated to be between €1,514.92 and €1,992.77.

Practice Materials

  • Practice materials available online

‘Join the Healthy Boat – Primary School’ is supported by a ready-to-use curriculum covering grades 1-4. There is one manual available for grades 1-2, and another manual available for grades 3-4. These manuals and other intervention materials can be found online at https://www.gesundes-boot.de/grundschule/materialien/#c4148.

The cost of the materials is €38.90 for curriculum guides/manuals. Supplementary materials range in cost from €4.70 to €24.70  

Enduring impact

  • No follow-ups have been conducted

Evaluation details

Evaluation 1 

Reference: Kobel et al. (2014). Effects of a School-Based Health Promotion Programme on Obesity Related Behavioural Outcomes. Journal of Obesity, Volume 2014.  

Kobel and colleagues’ (2014) analysis of the Baden-Württemberg trial data focused on the intervention’s impact upon parents’ reports of the children’s time spent engaging in physical activity per week, time spent using screen media per day, amount of sugar-sweetened beverages consumed per week, and frequency of breakfast-skipping. Of the 1,947 pupils who provided baseline data, one-year follow-up data was available for 1,736 pupils (IG=954, CG=782). For this sample, the only significant baseline difference between the CG and IG was with respect to the percentage of children with a migrant background, which was significantly higher (P<0.01) in the IG (34.2%) than in the CG (27.2%).  

The authors used logistic regressions adjusting for all baseline measures to calculate odds ratios for the four health outcomes. When looking at the total sample of first- and second-grade children, exposure to the intervention was not found to have a significant impact upon any of the four outcomes under study. However, sub-group analyses found that exposure to the intervention was associated with improved outcomes for some specific populations. Parents of second graders in the IG were significantly less likely to report that their children often/always skip breakfast, while parents of girls, parents of children with non-migrant backgrounds, and parents with low educational attainment were significantly less likely to report that their children consume an hour or more of screen time per day in the IG than in the CG.

Summary of Results for Evaluation 1

Outcome

Intervention Group

Control Group

Outcomes improved at one-year follow-up (statistically significant)

 

 

Percentage of children skipping breakfast often/always (2nd graders only), n (%)

Odds ratio, IG vs CG

42 (10.8%)*

53 (16.6%)*

OR = 0.52, P=0.024, 95% CI [0.30; 0.92]

Percentage of children consuming greater or equal to 1 hour of screen media/day (girls only) n (%)

Odds ratio, IG vs CG

40 (9.8%)*

47 (14.2%)*

(OR=0.58, P=0.04, 95% CI [0.35; 0.96])

Percentage of children consuming greater or equal to 1 hour of screen media/day (children with non-migrant background only) n (%)

Odds ratio, IG vs CG

 

49 (9.3%)*

162 (12.8%)*

 

OR=0.61, P=0.043, 95% CI [0.38; 0.98]

Percentage of children consuming greater or equal to 1 hour of screen media/day (children whose parents have low educational attainment only), n (%)

Odds ratio, IG vs CG

70 (13.9%)*

75 (17.3%)*

(OR=0.64, P=0.032, 95% CI [0.43; 0.96])

Outcomes with no effect (total sample, n=1736)

 

 

Percentage of children engaging in moderate-to-vigorous physical activity for ⊃3;60 minutes/day, ⊃3;4 days/week

Odds ratio, IG vs CG

29.1%

26.5%

OR=1.18, P=0.19, 95% CI [0.92; 1.52]

Percentage of children consuming greater or equal to 1 hour of screen media/day

Odds ratio, IG vs CG

12.7%

14.6%

OR=0.75, P=0.10, 95% CI [0.53; 1.06]

Percentage of children consuming greater or equal to 1 sugar-sweetened beverage/week

Odds ratio, IG vs CG

21.8%

22.1%

OR=0.96, P=0.76, 95% CI [0.72; 1.28]

Percentage of children kipping breakfast often/always

Odds ratio, IG vs CG

12.4%

14.5%

OR=0.86, P=0.47, 95% CI [0.58; 1.29]

Note: *=statistically significant, p<0.05

 

Evaluation 2 

Reference: Lämmle et al. (2016). Intervention effects of a school-based health promotion program on children’s motor skills. Journal of Public Health, 24:185-192

Lämmle and colleagues’ (2016) analysis of the Baden-Württemberg trial data focused on the intervention’s impact upon children’s motor skills, and assessed the intervention’s impact using the standardised and validated Dordel-Koch-Test of flexibility, conditional skills, and coordinative skills (DKT). The exercises used include sit-and-reach (flexibility), lateral jumps and one-leg stand (coordinative skills), and sit-ups, push-ups, standing long jumps and 6-minute runs (conditional skills).

The DKT quantifies children’s skills in these areas through the following units:

  • Standing long-jump: how far (in centimetres) a child can jump while landing on both feet.
  • Six-minute run: the total distance (in meters) that a child can run in 6 minutes
  • Sit-ups: the number of sit-ups that a child can complete in 40 seconds
  • Push-ups: the number of push-ups that a child can complete in 40 seconds
  • One-leg stand: the number of times that children touch the floor with both feet while trying to balance on one foot for 60 seconds.
  • Lateral jumps: the number of times a child can jump back and forth over a line in two 15-second sessions
  • Sit-and-reach: the distance (in centimetres) that a child’s fingertips can reach past their toes when sitting with their legs fully extended in front of them.

Of the 1,943 pupils who provided baseline data, one-year follow-up data was available for 1,736 pupils (IG=957, CG=779). No significant baseline differences were detected between the intervention and control groups across the variables of gender, age, grade, height, weight, BMI, and proportion of overweight or obese pupils. However, there was a statistically significant difference between the IG and CG participants in the number of meters covered during the 6-minute run, with control group participants covering a larger distance on average than intervention group participants (IG mean = 839.7 (SD=122.9), CG mean = 855.3 (SD=120.1), p<0.01). There was also a statistically significant difference in the performance of IG and CG participants on the sit-and-reach test, with intervention group children reaching an average of 2.0cm beyond their toes (SD=5.8) and the control group children reaching an average of 1.4cm beyond their toes (SD=5.5) (p<0.02).

Between baseline and one-year follow-up, children in both the IG and the CG showed improvements in performance of the one-leg stand, lateral jump test, sit-up test, push-up test, standing long-jump test, and 6-minute run. There were no significant differences between the CG and the IG in the degree to which participants’ performance improved on any of these tests. However, for the construct of conditional skills as a whole, an Analysis of Covariance (ANCOVA) found that between baseline and follow-up, children in the IG improved to a significantly greater degree than children in the CG (F(1,1571) = 5.20, p ≤ 0.02).

Additionally, the intervention was found to have a positive impact on children’s flexibility. Between baseline and follow-up, the distance that children in the IG could reach beyond their toes in the sit-and-reach flexibility test increased slightly from 2.0 cm (SD=5.8) at baseline to 2.1 (SD=6.4) at follow-up. Meanwhile, the flexibility of children in the CG declined from 1.4 cm (SD=5.5) at baseline to 1.0 cm (SD=6.1) at follow-up. For the construct of flexibility as a whole, an ANCOVA found that children in the IG experienced a significantly smaller decline in flexibility than children in the CG (F(1, 1715) = 6.68, p<0.01).

Furthermore, the mean distance reached by girls in the IG in the sit-and-reach flexibility test increased from 2.9cm (SD=5.8) at baseline to 3.5cm (SD=6.5) at follow-up, whereas the mean distance reached by control group girls declined slightly from 2.4cm (SD=5.3) at baseline to 2.3cm (SD=6.0) at follow up. An ANCOVA showed that this improvement in flexibility for IG girls compared to CG girls was statistically significant: F(1,839) =100.88, p ≤ 0.02.

Summary of Results for Evaluation 2

Outcome

Intervention Group

Control Group

Outcomes improved (statistically significant)

 

 

Conditional skills overall, IG vs CG [ANCOVA results]

F(1,1571) = 5.20, p ≤ 0.02)

Adjusted R2=0.376

Flexibility

Overall flexibility, IG vs CG

[ANCOVA results]

(F(1, 1715) = 6.68, p<0.01)

Adjusted R2=0.540

Sit-and-Reach overall T2-T1 change, cm [mean(SD)]

-0.1 (4.3)***

-0.4 (4.6)***

Sit-and-Reach (girls only) T2-T1 change, cm [m(SD)]

0.3 (4.3)**

-0.2 (4.6)**

Outcomes with no effect

 

 

Coordinative skills

 

 

One-leg stand T2-T1 change, n [m(SD)]

-2.3 (5.1)

-2.1 (4.5)

Lateral jumps, T2-T1 change, n [m(SD)]

11.4 (9.4)

11.1 (8.9)

Conditional skills

 

 

Standing long jump T2-T1 change, cm [mean (SD)]

11.0 (18.4)

10.2 (18.1)

Sit-ups T2-T1 change, n [m(SD)]

2.8 (6.1)

2.5 (6)

Push-ups T2-T1 change, n [m(SD)]

2.1 (5.1)

1.5 (5.1)

6-minute run T2-T1 change, m [m(SD)]

70.5 (127.9)

58.9 (106.2)

Note: *significant difference p £ 0.05; **significant difference p £ 0.02, ***significant difference p £ 0.01

Evaluation 3

Reference: Kesztyüs et al. (2016). Effects of a statewide health promotion in primary schools on children’s sick days, visits to a physician, and parental absence from work: a cluster-randomized trial. BMC public Health, 16:1244

Kesztyüs and colleagues’ (2016) analysis of the Baden-Württemberg trial data focused on the intervention’s impact upon the frequency of children’s sick days, visits to a physician, and days of absence of mothers and fathers from work. Of the 1,968 pupils whose parents agreed to take part, only 1,714 of them provided baseline data (87%). At follow up, 1,379 parents (76.5%) provided data on children’s sick days, 1,253 parents (73.1%) provided follow-up data on visits to a physician, 732 parents (42.7%) provided data on mother’s days off of work, and 473 parents (27%) provided data on father’s days off of work.

In this sample, there were a number of statistically significant differences between IG and CG children at baseline. The proportion of children with a migration background was significantly higher in the IG than in the CG, and the average number of workdays missed was higher in the IG than in the CG for both mothers and fathers. The proportion of children who spent over 60 minutes playing outside per day was significantly lower in the IG than in the CG, and the average distance ran during a 6-minute running test was also significantly lower in the IG than in the CG.

Mann-Whitney U-tests, t-tests, and Welch tests were used to assess differences between groups across continuous variables. These tests found that sick days taken by children and days of work missed by mothers fell to a significantly greater degree in the IG than in the CG, but when controlling for baseline variables in linear regression models, the statistical significance of the intervention effect upon these outcomes was lost.

When looking only at children in Primary 1 for whom there was available data (n=654), however, the number of sick days taken by children was found to have fallen to a significantly greater degree in the IG than in the CG. This intervention effect remained significant (p=0.003) when controlling for baseline values in a linear regression model, meaning that for a small sub-group of Primary 1 students representing 38.2% of the original 1,714 Primary 1 and Primary 2 students who provided baseline data, participation in the intervention was associated with a significant reduction in the number of sick days taken by children.

For the pupils for whom follow-up data was available on the number of children’s visits to a physician (n=1,253) and the number of workdays missed by their fathers (n=473), no significant differences were found between the intervention group and the control group.

Summary of Results for Evaluation 3

Outcome

Intervention Group

Control Group

Outcomes improved (statistically significant) – Primary 1

 

 

Number of children’s sick days (primary 1 follow-up n=654) T2-T1 [m(SD]

Linear model – impact of intervention on sick days, controlling for grade, gender, & baseline values

-5.5(NR)***

-3.64(NR)***

B(SE) = -0.83(0.28), b = -0.06, p-value = 0.003

Outcomes with no effect – Overall sample

 

 

Number of workdays missed by fathers (n=473) T2-T1 [m(SD)]

-0.25(2.21)

0.06(1.56)

Number of workdays missed by mothers (n=732) T2-T1 [m(SD)]

Linear model – impact of intervention on mother’s days off work, controlling for baseline values & school clustering effects

-1.11(3,89)**

-0.52(2.84)**

Statistical significance lost (full data not reported)

Number of visits to a physician (n=1,253) [m(SD)]

-0.85(2.41)

-0.93(2.70)

Number of children’s sick days (overall sample follow-up n=1,268) T2-T1 [m(SD]

Linear model – impact of intervention on sick days, controlling for grade, gender, & baseline values

-3.18(7.08)*

-2.31(5.56)*

B(SE) = -0.30(0.22), b = -0.02, p-value = 0.182

Note: NR=Not Reported; * p £ 0.02;  **p-value £ 0.019; ***p-value £ 0.013

Evaluation 4 

Reference: Kobel et al. (2017). Effects of a Randomized Controlled School-Based Health Promotion Intervention on Obesity Related Behavioural Outcomes of Children with Migration Background. Journal of Immigrant Minority Health, 19:254-262

Kobel and colleagues’ (2017) analysis of the Baden-Württemberg trial data focused on a sub-group of children with migration backgrounds (n=525, 318 in IG and 207 in CG), and assessed the intervention’s impact upon parents’ reports of the children’s time spent engaging in physical activity per week, time spent using screen media per day, amount of sugar-sweetened beverages consumed per week, and intake of fresh fruit and vegetables. In this sample, the only baseline difference detected was for the proportion of children who used screen media for an hour or more each day, which was significantly higher in the IG than in the CG.

The authors used logistic regression analyses to compare the proportion of children in the IG and the CG between baseline and follow-up who 1) engaged in moderate-to-vigorous exercise for at least an hour at a time, four times per week, 2) consumed one hour or more of screen media per day, 3) consumed one or more sugar-sweetened beverage per week, or 4) consumed fresh fruit and vegetables at school ‘often/always’. These logistic regression analyses were adjusted to take account of children’s ages, weight status, parental education level, and baseline characteristics. At one-year follow-up, there were no statistically significant differences between the IG and the CG in the proportion of children who reported engaging in any of the four outcomes, suggesting that the intervention did not have a positive effect upon the intervention group.

A related-samples marginal homogeneity test was conducted to assess changes from T1 to T2 in the proportion of children who ‘never’/’rarely’ ate fruits and vegetables. The authors concluded that between T1 and T2, the proportion of children who ‘rarely’/’never’ ate fruits and vegetables reduced to a significantly greater degree in the IG than in the CG (p £ 0.035). However, the numbers upon which the related-samples marginal homogeneity test was conducted do not match the data reported in the publication, and the pre-planned logistic regression analyses conducted by the authors showed no significant intervention effects for any of the outcomes measured. Therefore, this positive intervention result should be treated with caution.

Summary of Results for Evaluation 4

Outcome

Intervention Group

Control Group

Outcomes improved (statistically significant)

 

 

The change in proportion of children who ‘never’/’rarely’ ate fruit and vegetables in school**

Fall from 25.4% at T1 to 18.0% at T2*

Fall from 21.4% at T1 to 21.1% at T2*

Outcomes with no effect***

 

 

Percentage of children engaging in moderate-to-vigorous physical activity for ⊃3;60 minutes/day, ⊃3;4 days/week

Odds ratio, IG vs CG

26.2%

27.1%

OR 1.085 [0.622; 1.892], p = 0.775

Percentage of children consuming greater or equal to 1 hour of screen media/day

Odds ratio, IG vs CG

19.4%

18.2%

OR 0.932 [0.497; 1.747], p = 0.826

Percentage of children consuming greater or equal to 1 sugar-sweetened beverage/week

Odds ratio, IG vs CG

80.2%

79.2%

Odds ratio not reported

Percentage of children consuming fresh fruit and vegetables at school often/always

Odds ratio, IG vs CG

80.2%

79.2%

OR 1.663 [0.895; 3.090]; p=0.108

Note: *p £ 0.035; **Please note that the numbers reported for this ad-hoc analysis do not match up with the data shown in tables 1 and 2, so the finding should be treated with caution; ***The odds ratios were calculated controlling for baseline characteristics, age, weight status, and parental education level;.

Evaluation 5 

Reference: Lämmle et al. (2017). Intervention Effects of the School-Based Health Promotion Program ‘Join the Healthy Boat’ on Motor Abilities of Children with Migration Background. Health, 9:520-533  

Lämmle and colleagues’ (2017) analysis of the Baden-Württemberg trial data focused on a sub-group of children with migration backgrounds (n=525, 318 in IG and 207 in CG), and assessed the intervention’s impact upon their motor abilities as assessed by the standardised and validated Dordel-Koch-Test of flexibility, conditional skills, and coordinative skills (DKT). The exercises used include sit-and-reach (flexibility), lateral jumps and one-leg stand (coordinative skills), and sit-ups, push-ups, standing long jumps and 6-minute runs (conditional skills).

The DKT quantifies children’s skills in these areas through the following units:

  • Standing long-jump: how far (in centimetres) a child can jump while landing on both feet.
  • Six-minute run: the total distance (in meters) that a child can run in 6 minutes
  • Sit-ups: the number of sit-ups that a child can complete in 40 seconds
  • Push-ups: the number of push-ups that a child can complete in 40 seconds
  • One-leg stand: the number of times that children touch the floor with both feet while trying to balance on one foot for 60 seconds.
  • Lateral jumps: the number of times a child can jump back and forth over a line in two 15-second sessions
  • Sit-and-reach: the distance (in centimetres) that a child’s fingertips can reach past their toes when sitting with their legs fully extended in front of them.

For this sample, no significant differences were found between children in the IG and the CG at baseline in terms of their grade, gender age, weight, height, BMI, or proportion of overweight or obese children.

The only outcomes for which a statistically significant and positive intervention effect was detected were the construct of boys’ overall conditional skills and the flexibility (sit-and-reach) test for girls. An ANCOVA found that boy’s conditional skills increased to a significantly greater degree in the IG than the CG between T1 and T2, while the mean change in cm that girls could reach during the flexibility test increased significantly for pupils in the IG (+1.0 (SD=4.2)) and fell significantly for pupils in the CG (-0.7 (SD=4.6)) (p<0.01).

Summary of Results for Evaluation 5

Outcome

Intervention Group

Control Group

Outcomes improved (statistically significant)

 

 

Improvement in overall conditional skills between T1 and T2 (boys), IG vs CG

ANCOVA

 

 

F(1,201) = 8.02, p ≤ 0.005)

Sit-and-Reach (girls only) T2-T1 change, cm [m(SD)]

ANCOVA

1.0 (4.2)*

-0.7 (4.6)*

F(1,226) = 10.72, p<0.01

Outcomes with no effect

 

 

Flexibility

 

 

Sit-and-Reach overall T2-T1 change, cm [mean(SD)]

0.2(4.4)

-0.64(4,8)

Coordinative skills

 

 

One-leg stand T2-T1 change, n [m(SD)]

-2.4 (5.7)

-1.9 (4.6)

Lateral jumps, T2-T1 change, n [m(SD)]

10.4 (9.7)

11.0 (9.1)

Conditional skills

 

 

Standing long jump T2-T1 change, cm [mean (SD)]

1.4 (20.1)

8.6 (18.0)

Push-ups T2-T1 change, n [m(SD)]

2.3 (4.8)

1.2 (5.2)

Sit-ups completed in 40 seconds T2-T1 change, n [m(SD)]

3.1 (6.2)

2.5 (5.8)

6-minute run, T2-T1 change, m [m(SD)]

75.0 (116.4)

52.7 (105.7)

Note: *significant difference p £ 0.05

Evaluation 6 

Reference: Kestyüs et al. (2017). Costs and Effects of a State-Wide Health Promotion Program in Primary Schools in Germany – The Baden-Württemberg Study: A Cluster-Randomized, Controlled Trial. PLOS One, 12(2): e0172332. doi:10.1371/journal.pone.0172332

Kesztyüs and colleagues’ (2017) analysis of the Baden-Württemberg trial data focused on the intervention’s impact upon children’s waist circumferences, waist-to-height ratios, and rates of overweight children, obesity and abdominal obesity. Of the 1,947 pupils who provided baseline data, follow-up data was available for 1,733 pupils (IG=955, CG=778). For this population, baseline analyses found that the proportion of children with migrant backgrounds was significantly higher in the IG than in the CG, and that a significantly greater proportion of mothers were smokers in the IG than in the CG. The proportion of children who played outside for at least 60 minutes per day was also significantly higher in the CG than in the IG.

T-tests, Mann-Whitney U tests, and Welch-tests were used to assess differences between the intervention and control group for variables using continuous data. These tests found that BMI percentiles changed to a significantly greater degree in the IG than in the CG (p=0.038), however when controlling for baseline values of BMI percentiles in a regression analysis, the significance of the intervention effect upon mean BMI percentiles per group was lost (p=0.403). Logistic regressions were run to assess the impact of the intervention upon abdominal obesity rates. The authors reported that for the sub-group of children who had complete follow-up data on breakfast-skipping habits, children in the intervention group had less than half the odds of developing abdominal obesity compared to children in the intervention group (OR 0.48 [0.25; 0.95]). However, the p-values were not reported and the adjusted R2 of the logistic regression is only 0.14, indicating that their analysis of the impact of the intervention only explains 14% of the total variance in abdominal obesity between IG and CG children. As a result, these results should not be treated as evidence of the effectiveness of the intervention in preventing abdominal obesity.

Summary of Results for Evaluation 5

Outcome

Intervention Group

Control Group

Outcomes improved (statistically significant)

 

 

BMI Percentile change, T2-T1 [m(sd)]

Regression model controlling for baseline BMI percentiles

0.67 (10.34)*

0.17(10.18)*

Statistical significance lost (p=4.03) (full regression results not shown)

Outcomes with no effect

 

 

Waist-to-Height Ratio change, T2-T1 [m(sd)]

-0.007 (0.022)

-0.008 (0.022)

Incidence overweight, change, T2-T1 n(%)

29 (3.1)

19 (2.4)

Remission overweight, change, T2-T1 n(%)

12 (1.3)

11 (1.4)

Incidence obesity, change, T2-T1 n(%)

10 (1.1)

5 (0.6)

Remission obesity, change, T2-T1 n(%)

5 (0.5)

4 (0.5)

Incidence abdominal obesity, n(%)

Logistic regression controlling for grade, gender, baseline waist-to-height ratio, and breakfast-skipping behaviour.

21 (2.2)

27 (3.5)

(OR 0.48 [0.25; 0.95]; p-value not reported)**

Remission abdominal obesity, n(%)

10 (1.0)

14 (1.8)

Note: *significant difference p = 0.038; ** This logistic regression model has a very low adjusted R2 value of 0.14, meaning that it only explains 14% of the variance in abdominal obesity rates between the IG and the CG.

Issues to consider

The research group that developed the intervention is currently the only group that has evaluated it. In studies assessing changes to children’s health-related behaviours at home, only parent-report measures are used, meaning that the results may be inaccurate due to social desirability bias. Some of the reports showed that there were significant differences between the intervention and control groups at baseline, which may have impacted the validity of the results. Only one of the 6 evaluations performed the calculations needed to manage the risk of clustering effects due to the cluster-randomised design. Many of the statistically significant intervention effects were only found in ad-hoc analyses of sub-groups. Standardised effect sizes were not always reported, and some ‘positive’ intervention effects were reported without p-values. For example, the cost-benefit analysis was based on an intervention effect for which no p-value was reported, meaning that the statistical intervention effect could have been entirely due to chance. Finally, some analyses were based on variables for which very little data was available (high attrition or high missing values).

Contact Information

Organisation:

The Baden-Württemberg Foundation

Address:

Kriegsbergstraße 42, Stuttgart, Baden-Württemberg 70174, DE

     Website: 

https://www.gesundes-boot.de/kontakt/ 
 

Available resources

Materials can be purchased at: https://www.auer-verlag.de/catalogsearch/result/?q=komm+mit+in+das+gesunde+boot.

These include:

  • Curriculum for grades 1-2, €38.90
  • Curriculum for grades 3-4, €38.90
  • Ideas for an active and healthy day, €24.70
  • 60 ideas for movement landscapes, €16.40
  • Reading stories with the Island Pirates, €4.70

Bibliography

Dreyhaupt et al. (2012). Evaluation of a health promotion program in children: Study protocol and design of the cluster-randomized Baden-Württemberg primary school study. BMC Public Health, 12:157; http://www.biomedcentral.com/1471-2458/12/157

Kesztyüs et al. (2016). Effects of a statewide health promotion in primary schools on children’s sick days, visits to a physician, and parental absence from work: a cluster-randomized trial. BMC public Health, 16:1244

Kestyüs et al. (2017). Costs and Effects of a State-Wide Health Promotion Program in Primary Schools in Germany – The Baden-Württemberg Study: A Cluster-Randomized, Controlled Trial. PLOS One, 12(2): e0172332. doi:10.1371/journal.pone.0172332

Kobel et al. (2014). Effects of a School-Based Health Promotion Programme on Obesity Related Behavioural Outcomes. Journal of Obesity, Volume 2014.

Kobel et al. (2017). Effects of a Randomized Controlled School-Based Health Promotion Intervention on Obesity Related Behavioural Outcomes of Children with Migration Background. Journal of Immigrant Minority Health, 19:254-262

Lämmle et al. (2017). Intervention Effects of the School-Based Health Promotion Program ‘Join the Healthy Boat’ on Motor Abilities of Children with Migration Background. Health, 9:520-533

Lämmle et al. (2016). Intervention effects of a school-based health promotion program on children’s motor skills. Journal of Public Health, 24:185-192

Wartha O, Brandstetter S. (2009). Join the Healthy Boat. Primary School. Physical Activity, Nutrition and Recreational Activities with the Little Pirates. 1./2. Grade Donauwörth: Auer Verlag GmbH.

Wartha O, Kobel S, Weber S. (2010). Join the Healthy Boat. Primary School. Physical Activity, Nutrition and Recreational Activities with the Little Pirates. 3./4. Grade Donauwörth: Auer Verlag GmbH.

Last updated

October 2019

Find a practice (within the 'Evidence based practices'):

The practices can be also searched along the three policy pillars of the Recommendation for Investing in Children:


Share this page