Long-term effects of daily physical education throughout compulsory school on duration of physical activity in young adulthood: an 11-year prospective controlled study

Abstract

Objectives

We examined whether daily physical activity (PA) during compulsory school encourages children to be more physically active during the intervention and 4 years after termination of the programme.

Methods

This prospective controlled intervention study followed the same 124 children (81 children in an intervention group and 43 controls) aged 7.7±0.6 (mean±SD) during a 7-year PA intervention and 4 years after the intervention when the children were 18.7±0.3 years old. The intervention included daily school physical education (PE) (200 min/week), whereas the controls continued with the Swedish standard of 60 min/week. Using a questionnaire, we gathered data about total PA, leisure time PA and sedentary activities (SA). Group comparisons are adjusted for age and gender, and data are provided as means with 95% CIs.

Results

At baseline, we found similar duration of PA and SA between groups. After a mean of 7 years with intervention, the intervention group was more physically active than the controls (+4.5 (2.9 to 6.0) hours/week), whereas SA was similar (+0.6 (−2.5 to 3.9) hours/week). Four years beyond the intervention, the intervention group was still more physically active than the controls (2.7 (0.8 to 4.7) hours/week), and SA was still similar (−3.9 (−9.7 to 1.7) hours/week).

Conclusions

Intervention with daily school PE throughout compulsory school is associated with higher duration of PA not only during the intervention but also 4 years after termination of the programme.

What are the new findings?

• An intervention programme with daily physical education throughout compulsory school is followed by higher duration of physical activity during the programme but also 4 years beyond termination of the intervention (11 years after baseline).

• These results were found with an intervention that was given within the compulsory national school curriculum, without extra resources and extra personnel or incremental costs, with guidance by the regular teachers, making it possible to apply to all schools.

• Contradictory to the Activity Stat Theory, children do not compensate increased duration of physical activity with increased sedentary behaviour.

• These findings are of importance since it seems possible to teach children an active lifestyle with maintained effects in young adulthood, and that this method could possibly be effective to prevent inactivity-related diseases later in life.

Introduction

Physical activity (PA) is in children associated with health benefits,¹ including outcomes related to bone and muscle,² body weight,³ academic achievement,⁴ and reduced risk of developing obesity, type II diabetes and hypertension.⁵ The guidelines of the WHO recommend children aged 5-17 years to take part in at least 60 min of PA per day.⁶ However, only 20% of global 13-15 year olds meet these recommendations⁷ and physical education (PE) in schools is decreasing.⁸ PE interventions have the potential to reverse these trends since schools reach almost all children.⁸

Most previous PA intervention studies have shown effective results on PA levels during the programme.^9–11 But few have focused on what happens with the activity levels after the programme, and the results from these studies are conflicting.^12–17 Further, studies that suggest that there is an association between PA levels in childhood and adulthood have only evaluated voluntary PA levels,^{18 19} but not if increased PA levels by interventions are maintained in young adulthood or not. Even more, the Activity Stat Theory suggests that the level of PA in children is centrally regulated to a predefined level, specific for every child,^20–22 and that all attempts to increase PA would be counteracted by a compensatory decline in PA during other parts of the day.

We therefore aimed to examine whether a school-based PA intervention throughout compulsory schoolinduces a more physically active lifestyle that remains 4 years after the intervention, when the children reached young adulthood.

Materials and methods

The Pediatric Osteoporosis Prevention study is a prospective controlled intervention study where four neighbouring, government-funded elementary schools were invited to participate. One school was chosen as intervention school and three as control schools. Children were allocated to schools according to their residential address. PE is a mandatory school subject in Sweden, all children had to participate. Before we initiated intervention, all schools had the Swedish standard PE curriculum of 60 min in one or two classes per week. The intervention school increased duration of PE from 60 min/week to 200 min/week (40 min PE per school day). The intervention included activities of the regular curriculum and were led by the same teachers as before the start of the study. The control schools continued with the Swedish standard PE curriculum.

At baseline we invited all children who started first or second grade between 1999 and 2000 to participate in the study. There were 192/302 boys and 157/262 girls who accepted participation (figure 1). At baseline the participants were 7.7±0.6 years old (mean±SD) and 98% were of Caucasian ethnicity. The children were annually re-evaluated with questionnaire, physical performance tests and physical measurements during the intervention and 3.7±0.5 years after termination. In this report, we present the re-evaluations after 5.0±0.1 and 7.4±0.6 years during the intervention period and the follow-up visit 4 years beyond the programme. In Sweden, compulsory school last for 9 years but since the children started the intervention in first or second grade, the mean duration of intervention was 7.4±0.6 in our study-cohort. At the last follow-up 3.7±0.5 years after termination of the intervention, 66 boys and 58 girls attended and were then aged 18.7±0.3 years. To be included in this report, children had to have participated in the baseline and the last follow-up exams (figure 1).

Figure 1.

A flow-chart of the study-population

We measured height (cm) and weight (kg) by standard equipment (Holtain Stadiometer and Avery Berkel HL 120 electric scale). Our research nurses assessed Tanner stage^{23 24} at baseline, while self-assessment was used thereafter and reported in the questionnaire. We registered organised PA during leisure time (ie, mean value of weekly summer and winter leisure time PA) and duration of sedentary activities (SA), defined as screen time activities by a questionnaire.^{25 26} SA was in the questionnaire answered in hours/day but converted to hours/week in this study. Total PA was estimated as the summarized duration of school PE and leisure time PA.

At the last follow-up the questionnaire was modified and the total PA was assessed as (1) weekly duration of PA (except walking) and (2) weekly duration of walking as exercise, also in summer and winter separately. If PA was only reported in one season, we used this as the annual duration of PA. Missing values in both seasons were excluded from the analysis. Reported weekly kilometres of walking as exercise were translated to duration by use of a walking speed of 6 km/hour. SA at the last follow-up also included book reading, crossword puzzle solving and time spent in vehicles. Annual mean PA was estimated as previously. Even if the questionnaires are not validated, they have been used in previous published reports.^{2 9 25 26} Questions regarding PA and SA from the questionnaires are found in online supplementary appendix 1.

Dropout analysis revealed no significant differences in anthropometrics, PA and SA at baseline and at the last intervention evaluation between children included in this report and those who did not fulfil the inclusion criteria (data not shown). We have previously found similar anthropometrics in the children who attended the baseline exam with those who did not by use of the Swedish compulsory first-grade school health examinations.²⁵

We used IBM SPSS (V.23); descriptive statistics are presented as numbers (n), proportions (%) and mean±SD, and inferential statistics as means with 95% CIs. Shapiro-Wilk test was used to test normality distribution of the data. Differences between groups were tested by Student’s t-test, χ² test or univariate bootstrapped 10 000 samples analysis of covariance (adjusted for age and gender). P<0.05 was considered a statistically significant difference.

Results

Descriptive statistics of the study cohort are presented in table 1. Throughout the 7-year intervention, the children in the intervention group were more physically active than their counterparts in the control group (p<0.001 for all evaluated periods). After 7 years of intervention, the children also participated in more leisure time PA (p<0.01). No significant difference in duration of SA was found between the intervention and control children throughout the intervention period (table 2).

Table 1.

Anthropometric measurements and lifestyle factors in the intervention group and control group

	Intervention		Control
	Boys	Girls	Boys	Girls
Baseline
n (%)	45 (56)	36 (44)	21 (49)	22 (51)
Caucasian (%)	45 (100)	36 (100)	19 (90.5)	22 (100)
Age (years)	7.5±0.6	7.5±0.6	8.0±0.7	7.9±0.6
Height (cm)	128.6±7.3	126.4±6.4	130.0±7.0	128.2±7.6
Weight (kg)	27.8±5.5	26.5±5.3	27.4±4.5	27.0±6.1
BMI (kg/m2)	16.7±2.5	16.5±2.5	16.1±1.5	16.2±2.1
Mean 5.0±0.6 (mean±SD) years after baseline
Age (years)	12.6±0.6	12.5±0.6	12.9±0.7	12.9±0.6
Height (cm)	158.0±10.5	158.6±8.6	159.7±9.5	158.5±8.1
Weight (kg)	48.1±11.4	48.4±9.8	47.0±9.6	46.7±9.7
BMI (kg/m2)	19.1±3.4	19.1±2.7	18.2±2.2	18.4±2.4
Tanner stage 1/2/3/4/5 (%)	2/16/36/41/5	6/14/34/40/6	5/10/52/29/4	14/0/38/34/14
Mean 7.4±0.6 (mean±SD) years after baseline
Age (years)	15.0±0.5	15.1±0.6	15.3±0.5	15.0±0.8
Height (cm)	175.0±9.1	166.9±5.6	175.1±9.3	164.9±5.5
Weight (kg)	64.8±13.8	57.9±7.7	62.7±11.3	54.3±9.7
BMI (kg/m2)	21.1±3.9	20.8±2.4	20.4±2.8	19.9±2.5
Tanner stage 1/2/3/4/5 (%)	0/0/3/23/74	0/0/12/39/49	0/0/0/18/82	0/0/0/54/46
Mean 3.7±0.5 (mean±SD) years after termination of intervention
Age (years)	18.7±0.2	18.7±0.4	18.8±0.3	18.7±0.3
Height (cm)	180.4±7.3	168.0±5.2	180.3±7.9	166.4±5.5
Weight (kg)	77.0±15.9	64.0±9.8	74.3±11.7	61.2±11.4
BMI (kg/m2)	23.7±4.8	22.7±3.3	22.9±3.6	22.0±3.3

BMI, body mass index.

Table 2.

Duration of PA and sedentary behaviour in hours/week during mean 7.4 years with daily school PA in the intervention group and 1–2 days/week in the control group and mean 3.7 years after termination of the intervention

	Intervention (unadjusted values) (n=81)	Control (unadjusted values) (n=43)	Mean difference (group comparison adjusted for gender and age)	P values (group comparison adjusted for gender and age)
Just before baseline
PA in leisure time	2.7 (2.0 to 3.3)	3.0 (2.1 to 3.8)	−0.2 (−1.2 to 0.8)	0.69
Sedentary activity	12.5 (11.1 to 13.9)	11.4 (10.1 to 12.8)	1.7 (−0.4 to 3.9)	0.13
Baseline
PE in school	3.3	1.0	Not applicable	Not applicable
PA in leisure time	2.7 (2.0 to 3.3)	3.0 (2.1 to 3.8)	−0.2 (−1.2 to 0.8)	0.69
Total PA	6.0 (5.3 to 6.6)	4.0 (3.1 to 4.8)	2.1 (1.1 to 3.1)	< 0.001
Sedentary activity	12.5 (11.1 to 13.9)	11.4 (10.1 to 12.8)	1.7 (−0.4 to 3.9)	0.13
5.0±0.6 (mean±SD) years after baseline
PE in school	3.3	1.0	Not applicable	Not applicable
PA in leisure time	6.1 (5.1 to 7.1)	5.0 (4.0 to 6.0)	1.2 (−0.2 to 2.6)	0.10
Total PA	9.4 (8.4 to 10.4)	6.0 (5.0 to 7.0)	3.5 (2.1 to 4.9)	< 0.001
Sedentary activity	15.9 (14.5 to 17.3)	18.7 (16.0 to 21.3)	−2.3 (−5.1 to 0.5)	0.11
7.4±0.6 (mean±SD) years after baseline
PE in school	3.3	1.0	Not applicable	Not applicable
PA in leisure time	6.7 (5.5 to 7.9)	4.4 (3.3 to 5.6)	2.1 (0.6 to 3.6)	< 0.01
Total PA	10.0 (8.8 to 11.2)	5.4 (4.3 to 6.5)	4.5 (2.9 to 6.0)	< 0.001
Sedentary activity	19.5 (17.2 to 21.7)	18.6 (16.0 to 21.1)	0.6 (−2.5 to 3.9)	0.71
3.7±0.5 (mean±SD) years after termination of intervention
PA besides walking	5.8 (4.5 to 7.1)	3.7 (2.7 to 4.6)	2.1 (0.5 to 3.7)	< 0.05
Duration of walking	1.6 (1.0 to 2.2)	1.0 (0.2 to 1.7)	0.7 (−0.2 to 1.5)	0.15
Total PA	7.4 (5.9 to 8.9)	4.7 (3.4 to 5.9)	2.7 (0.8 to 4.7)	< 0.01
Sedentary activity	23.5 (20.6 to 26.4)	27.1 (22.0 to 32.2)	−3.9 (−9.7 to 1.7)	0.19

Data are presented as mean and 95% CI (bootstrapped CI) within brackets.

Statistically significant differences are highlighted in bold.

PA, physical activity; PE, physical education.

Four years after termination of the programme, children who previously received intervention were more physically active than those who did not (p<0.01) (table 2), while SA was still non-significant (table 2).

Discussion

In this prospective controlled study, we found that daily PE intervention throughout compulsory school encourages a more physically active lifestyle in children who remain after the programme into young adulthood. Therefore, extra school PE seems to be a possible strategy to increase PA in children with maintained effects beyond the programme. With these results, we speculate that daily school PE may be one strategy to prevent inactivity-related diseases later in life.

Most previous studies have focused on short-term effects of PA interventions (ie, 1–3 years of an intervention),^9–11 and the few studies that have focused on possible long-term effects (ie, 1–2 years after the intervention) are conflicting.^12–17 Further, the studies that suggest that there is an association between PA levels in childhood and adulthood have only evaluated voluntary PA levels. Whether an intervention programme in young years could influence the ‘tracking’ has, to our knowledge, never been evaluated in studies with a prospective controlled study design. This study increases our knowledge when reporting that a PE intervention programme during growth has benefits in PA activity level that remain at least 4 years after the intervention.

Contradictory to the Activity Stat Theory, we found that children’s PA habits are modifiable and not set at an intrinsic level.^20–22 Furthermore, we found that children exposed to daily school PE had almost mean 4 hours less of SA per week than children in the control group. This difference was not statistically significant, but still of such interest that it should be explored in future studies, sinceinactivity, irrespective of the amount of PA, is associated to adverse health outcomes.²⁷

Study strengths include the long follow-up and the prospective controlled study design within a defined geographical area with a homogeneous population. Another benefit is that the intervention was given within the compulsory national school curriculum, without extra resources, extra personnel or incremental costs, with guidance by the regular teachers. It is also important to mention that even if the questionnaire was modified at the last follow-up, it is still possible to make a cross-sectional comparison between the two groups at each separate evaluation, even if it limits the possibility to evaluate changes in PA and SA longitudinally.

Weaknesses include the non-randomised design (which makes causal inferences impossible) and the relatively small sample size. There is a risk of confounding since other factors (eg, influence from parents and friends, living conditions and the child’s internal beliefs) are also associated with duration of PA in children.²⁸ We would also have preferred registration of quality and intensity of PA. The well-described difficulties to estimate PA in children are another concern,²⁹ and we would have preferred to use a validated PA questionnaire and objective measurements of PA. Further, our findings must be verified in other ethnic and socioeconomic groups.

In summary, an intervention with daily PE throughout compulsory school is associated with higher duration of PA also beyond the programme in young adulthood. In the same way that we teach our children to brush their teeth which they continue with throughout life, we can also teach them a physically active lifestyle that follows to young adulthood by implementing daily PA in school.