Public Health Nutrition: 26(7), 1488–1500 doi:10.

1017/S136898002300040X

Ecological system theory and community participation

to promote healthy food environments for obesity and
non-communicable diseases prevention among
school-age children
Pennapa Ritwong Suwannawong1, Naruemon Auemaneekul1,*, Arpaporn Powwattana1
and Rewadee Chongsuwat2
1
Department of Public Health Nursing, Faculty of Public Health, Mahidol University, 420/1 Ratchawithi Road, Thung
Phaya Thai, Ratchathewi District, Bangkok 10400, Thailand: 2Department of Nutrition, Faculty of Public Health,
Mahidol University, Bangkok, Thailand

Submitted 7 March 2022: Final revision received 20 September 2022: Accepted 15 November 2022: First published online 27 February 2023

Abstract
Objectives: To implement and evaluate the effectiveness of the community partici-
patory program between school and family based on ecological system theory and
participatory action research. The intervention covers three levels at the individual,
family and school levels and involves educating students and parents by using
technology, reducing sedentary behaviours, increasing exercise and changing to
healthy food environments at school and at home.
Design: A quasi-experimental design was used in this study.
Setting: Public primary school in Thailand.
Subjects: The participants in the study included 138 school-age children in grades
2–6 with their parents/guardians. The control group consisted of 134 school-age
children at a school of the same size with their parents/guardians.
Results: Results show that nutritional status was significantly improved within the
experimental group (P value = 0·000) and between groups during follow-up
(P value = 0·032). Students’ knowledge about obesity and non-communicable
chronic diseases (NCD) prevention, as well as physical activity and exercise behav-
iours, in the experimental group was significantly higher than that in the control
group (P value = 0·000 and 0·044, respectively). Parents’ perceptions of child
obesity and family modelling behaviours in the experimental group were also
significantly higher than that in the control group; P value = 0·013 and 0·000,
Keywords
respectively). Obesity prevention
Conclusion: The community participation program was found to be successful. Participation
Not only students, families and schools improved health behaviours and healthy Ecological system theory
food environments at home and school, but the students’ long-term nutritional School-age children
status also improved. Healthy foods

School-age obesity is a serious problem leading to a variety However, the sustainability of solving the problem of obesity
of chronic diseases. The percentage of children affected by among school-age children to prevent future risk factors for
obesity has more than tripled since the 1970s(1). Data from NCD by promoting a healthy environment would be a key
2015–2016 show that nearly one in five school-age children factor that should not to be overlooked.
aged 6–12 years in developing countries is obese(2). Previous studies(7–9) have clearly indicated that the
Empirical evidence demonstrates that obesity during the causes of obesity are multifactorial and complex. The inci-
school years strongly leads to a variety of chronic diseases dence of obesity among school-age children cannot be
in adulthood with consistently significant risk factors for explained based on the behavioural factors of students
non-communicable chronic diseases (NCD) such as CVD, alone; rather, the environmental conditions surrounding
metabolic syndrome and type 2 diabetes mellitus(3–6). students could also influence students’ health-related

*Corresponding author: Naruemon Auemaneekul, email naruemon.aue@mahidol.ac.th

© The Author(s), 2023. Published by Cambridge University Press on behalf of The Nutrition Society. This is an Open Access article, distributed under
the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use,
distribution and reproduction, provided the original article is properly cited.
Community participate for obesity prevention 1489
(10–12)
behaviours . This finding is supported by the ecology influence school-age obesity, but family environments
of human development by Bronfenbrenner (1979), who do, as well. Family factors are concealed with tremendous
developed an ecological system theory (EST) and illus- influence in building sustainability in school-age obesity
trated that children’s environments affect them in the form prevention(18). Previous studies have strongly recom-
of constant reciprocity between individuals and the envi- mended that family environmental factors must be
ronment(13,14). In other words, the family environment is taken under consideration in fostering health behaviour
a key setting for fostering healthy food intake and exercise among students so they can grow up as healthy adults(19).
for school-age children in the long run(10,11). At the same Successful promotion of healthy food environments for
time, the school acts as a vital platform for improving future long-term obesity and NCD prevention requires solutions
health through learning processes and school policy(6). covering all three levels at the individual, family and school
EST can help researchers explain interactions between levels(5,15,20,21). In particular, families need to seriously
school-age children and their environments in developing participate in solving the problem from the start until the
obesity(8) within complex layers of the environment among end of the process. Participation leads to all parties gaining
a set of nested and intertwined structures(13,14). the same understanding of the problem and setting the
According to EST by Bronfenbenner (1979), the envi- same goals(13,14).
ronment is clarified as five systems, beginning with the Thus, collaboration in problem-solving guidelines is
environment closest to the individual and extending to a sought in line with the real problems with an enhanced
broad distance from the individual. The microsystem sense of ownership of the problem and solutions(13,21).
describes settings in which there are direct interactions with This collaboration will pave the way for school and
children. This level is considered to be nearest to the indi- family to prevent school-age obesity in the same direction,
vidual and includes such settings as family environments. while contributing to sustainable and policy-driven
The home is the starting point for fostering children’s health programmes(5,22,23).
behaviours because home environments support family Currently, research on childhood obesity prevention
members in their efforts to have good health by providing places greater focus on family participation, but only as sup-
healthy foods, purchasing nutritious and accessible foods plementary efforts under school-based programmes(24,25).
for family members, increasing vegetable and fruit Unfortunately, parents in previous studies typically were
consumption and reducing sedentary activity(15). engaged by passive methods, such as newsletters(26), in
The mesosystem describes settings in which children one-way communication(25). Not surprisingly, these methods
are immersed and involves connections between settings, failed to engage parents and improve the nutritional status of
such as links between school and family. The exosystem children in the long run(24). On the contrary, some studies
describes settings in which there are no direct interactions have tried to promote collaborative efforts between school
with children, but indirect influences instead, such as and family to prevent childhood obesity by emphasising
school environments. Students spend at least eight hours participation principles through identifying problems and
per day at school(16). Therefore, school could be counted finding solutions by involving all stakeholders with
as an important environment to improve knowledge bottom-up approaches(17,21). The findings highlighted the
through classroom lessons, promotion of nutritious foods, development of participatory programmes by school and
reduction of unhealthy foods such as sweets and increases family collaboration to promote healthy food environments
of healthy foods such as vegetable and fruit menus during for obesity prevention by means of participatory action
school lunches, together with increases in daily exercise research (PAR), which is more likely to reduce limitations
through school policy(17). The macrosystem involves the regarding sustainability(5,20,21,27).
cultural and political values of a society, economic models Development of the programme based on the principle
and social conditions, and the chronosystem involves the of participatory action is likely to yield positive results(20,21).
historical moment in which the individual lives. School- Unfortunately, few studies have evaluated the effectiveness
age children spend their lives in two main communities, of programmes that were developed with this method(20,21).
namely home and school(15). Therefore, schools and family Therefore, the researcher’s interest is in developing an
members are considered communities for school-age chil- obesity prevention programme based on EST by applying
dren(15). In-depth understanding of the root causes of PAR with school and families for school-age obesity
school-age obesity might lead to long-term and appropriate prevention. Kemmis and Mc Taggart (1990) stated that
solutions in a factual context, which could result in a the PAR process consists of the following four steps: plan-
sustainable impact on the prevalence school-age obesity ning (discovering the cause of the problem, knowing the
at the community level(18). needs of all the stakeholders, solving problems, coping
Unfortunately, studies focused on individual-level with barriers and applying problem-solving guidelines),
prevention have been found to be unsuccessful and fail action (experimentation in real situations), observation
to create sustainability(16,17). Obviously, during school (observing possibilities) and reflection (reviewing and
term breaks, the incidence of obesity is higher and analysing processes and results)(28). We provide an over-
continues to rise. Therefore, not only school factors view of our methodological approach here with more
1490 PR Suwannawong et al.
detail in a forthcoming article. This paper highlights an experimental and control groups, respectively, who were
action step regarding the effectiveness of school and family studying in Grades 2–6 during the first semester of the
participatory programs to promote healthy food environ- 2019 academic year, together with their parents/guardians
ments for obesity and NCD prevention among school- in the same family.
age children.
Inclusion criteria
Students with normal and overweight nutritional status
Methods were included.

Design Exclusion criteria

A quasi-experimental design was used to test the effective- Students who transferred schools or dropped out during
ness of a community participatory program. Pretest, the study were excluded.
posttest and follow-up scores were assessed among experi- Students with underweight and obese nutritional status
ment and control groups. were excluded.

Participants Intervention
The purposive sampling technique was used in the study in Seven activities, as shown in Fig. 1 and Table 1, were the
areas among the top five provinces where child obesity is result of mutual participation in assessment, analysis and
prevalent, one of which was Aong Karuk, Nakhon Nayok decisions made together between family and school.
Province. Two public primary schools were selected as the Representatives were appointed from both sides to take
experimental and control groups. The two schools are responsibility for the planned project. These representa-
located in Health Region Area 4 with high statistics for chil- tives were called the core working group, which was made
dren with obesity in Thailand. The schools were also of ten willing volunteers composed of six persons from
willing to participate. Both schools met the health- school (school director, two teachers from Grades 4 and
promoting school criteria at the gold level under the 5, two physical education teachers and a school nurse)
health-promoting school recommendations of the WHO. and 4 persons from a family (of whom one was a public
There were 138 and 134 school-age children in the health nurse). Four persons from the core working group

Activity 1: Modification to a
healthy food environment at
school.
School level

Activity 2: Modification of the

Exosystem exercise environment at school.

School level Activity 3: Driving school health

policies between school and
family

Mesosystem Activity 1: Linking practices at

Linkage

home and school to create

Links between school and family consistency and sustainability.

Activity 1: Promoting a healthy

Microsystem food environment and exercise
Evaluation
Family level

environment at home by using

Developed Programme Family level technologies to raise awareness
Based on EST and reduce family obstacles.
Applying PAR

Individual level Activity 1: Promoting and

Individual level

educating students on obesity

and NCD prevention.

Activity 2: Reducing sedentary

behaviours or video game
addictions and increase exercise
PAR Process : planning, action, observation and reflection

Planning Action: 4 months for implementation Observation: 2 months Reflection

Pre-Test Post-Test Follow-up

(Before intervention) (Immediately after intervention) (Two months after intervention)

Fig. 1 The process of promoting healthy food environments for obesity and non-communicable chronic diseases (NCD) prevention
among school-age children
Community participate for obesity prevention 1491
Table 1 Activities and operational plans based on ecological system theory by teachers and parents (June–September 2019; 4 months)

Level Activities
Individual level activities Activity 1: Promoting and educating students on obesity and NCD prevention by teachers
- Contents on healthy eating and physical activity from teachers who integrated the content with daily
lessons
- Creating health knowledge boards along with entertaining games for students at school
- Monitoring the nutritional status of the students and teaching students to do self-assessments of their
nutritional status
Activity 2: Reducing sedentary behaviours or video game addictions and increasing exercise
- Encouraging students to make beneficial use of free time at school and home
Family level activities Activity 1: Promoting of a Healthy Food Environment and Exercise Environment at Home by Use of
(Microsystem) Technology
- Providing knowledge for parents/guardians by using an online course through three online platform
channels (YouTube, Facebook, LINE) to raise awareness and reducing family obstacles (co-produce,
co-design and co-create informational media via online platform together linking school and family)
- Learning from an online course for parents (Healthy-food eating and physical activity, family modelling
and parenting practices)
Mesosystem Activity 1: Linking practices at home and school to create consistency and sustainability
- Meeting to discuss health behaviours and healthy food environments of the students at home and school
between parents and teachers through the following two activities: 1. Recording daily behaviours at school
and home in terms of dietary intake and physical activities; 2. homework on healthy eating and physical
activity, family modelling and parenting practices where families put knowledge into practice at home to
monitor and control students’ behaviours in line with school activities
School level activities Activity 1: Modification to a healthy food environment at school
(Exosystem) - Changing of school lunches for students from the past when students could make their own purchases to
having the school organise a school lunch project for all students (providing highly nutritious foods and
vegetable menu items at school for students)
- Selling nutritious foods, such as fruits, to provide students with access only to healthy foods at their
schools; setting of rules for vendors and holding meetings once per term between teachers and
parents/guardians to discuss improvements in healthy menu items together
Activity 2: Modification of the exercise environment at school
- Organising 15-minute aerobic dance activities to a set rhythm in the mornings after lining up for the
national anthem
- Encouraging students to exercise more and making sports equipment available for students to play with
during their free time
Activity 3: Driving School Health Policies Together: Create an obesity and NCD prevention policy at school.

NCD, non-communicable chronic diseases.

later became the research assistants. Student representa- to join the programme. Normal-weight children ranged
tives also participated. The role of the researcher was to from normal weight, at risk of underweight and at risk of
catalyse and support-related theoretical knowledge. overweight, which ranged from –2 SD to þ2 SD, while over-
weight children ranged from þ2 SD to þ3 SD.

Outcomes
Output
Measures
To guarantee measurement quality and minimise error, Instrument development and quality testing
weight and height measurements were double checked The researcher and the core working group developed a
before recording. To ensure reliability, the researcher set of questionnaires together, and the researcher tested
assistants were trained for data collection prior to the study. the questionnaires for quality, as follows:
This study used the Thai Standard Growth Chart weight-for- For questionnaires, the content validity index (CVI) is
height ratios for Thai students(29). The Bureau of Nutrition, the proportion of items with expert consensus and should
Ministry of Public Health has developed a standard chart not be < 0·80. In the current study, testing was done by
adjusted by weight, height and sex specific to the Thai presenting the forms to three qualified experts to test for
population from 6 to 19 years of age to define underweight, content validity. Next, the questionnaires were pilot tested
at risk of underweight, normal weight, at risk of overweight, with thirty school-age children (fifteen normal-weight
overweight and obesity conditions(29). However, the students and fifteen overweight students) and their families
current study focused on the primary prevention of obesity at a school with a similar context. Cronbach’s alpha was
for school-age children. Children with obesity (defined as calculated to test internal consistency in the pilot study
school-age children over þ3 SD) and underweight children and had to be no < 0·75.
(defined as school-age children under –2 SD) who had In-depth interview guidelines were followed by the
malnutritional status were excluded from the study. Students school director, who steered the policy and participant-
with normal and overweight nutritional status were invited observation guidelines for the primary researcher to
1492 PR Suwannawong et al.
observe monitoring healthy environment factors at school. Family Perception of Child Obesity Questionnaire:
The advisory professors were asked to examine content There were twelve items that asked about family view-
accuracy and completeness in addition to making correc- points on the nutritional status of their children. The total
tions. Next, three qualified experts consisting of a nutrition of twelve items was scored on a four-point scale with
therapist, a paediatric nutrition expert and a paediatrician (1) totally disagree, (2) disagree, (3) agree and (4) totally
expert examined the content validity. agree. The CVI was 0·83, and the Cronbach’s alpha coeffi-
cient was 0·80.
Individual-level variables Family-Modeling Behavior Questionnaire: There were
For individual-level variables, the questionnaires were thirteen items that asked about individual practice as a
divided into the following two parts: Part 1 was a question- good role model of healthy eating and exercise(11). The
naire asking students about their obesity knowledge and total of twelve items was scored on a five-point scale with
perception; and Part 2 was a questionnaire asking students (1) never, (2) seldom, (3) sometimes, (4) often and (5)
about their physical activities, exercise behaviour and food daily. The CVI was 0·85, and Cronbach’s alpha coefficient
consumption behaviours. The respondents were Grades was 0·80.
2–6 students aged 8–12 years. The questions were read Parenting Practices Questionnaire: There were twenty
to the students by the researcher and well-trained research items that asked about the performance of parents in
assistants. Nevertheless, behavioural questionnaires were monitoring their children’s food consumption, physical
sent to parents/guardians for confirmation. activity and exercise behaviours at home(30). The total
Children’s Perception Questionnaires: six items of twenty items was scored on a five-point scale with
were asked about children with obesity and prevention, (1) never, (2) seldom, (3) sometimes, (4) often and (5)
including social norm attitudes on preferable appearance. daily. The CVI was 1·0, and the Cronbach’s alpha coeffi-
The six items were scored on a three-point scale with (1) cient was 0·81.
disagree, (2) uncertain and (3) agree. The scale applied
cartoon faces to represent the scale range from 1 to 3. School-level variables
This method fit the children’s age group by making the For school-level variables, the data were collected
questions interesting and easier to answer. The CVI was by qualitative data collection techniques with in-depth
1·0, and Cronbach’s alpha coefficient was 0·79. interview guidelines and participant observation about
Physical Activity and Exercise Behavior Questionnaires: school environment factors promoting the prevention of
seven items with cartoon pictures of activities asked about obesity.
physical activities and exercise behaviour during and An in-depth interview was conducted with one school
after school at home. The seven items were scored on a director as a rich, informative case. The primary researcher
five-point scale with (1) never, (2) seldom, (3) sometimes, conducted the interview by using open-ended questions
(4) often and (5) daily. The CVI was 0·88, and Cronbach’s following interviewing guidelines. The interview lasted
alpha coefficient was 0·75. approximately 1 h and 30 min. All data were audio-
Obesity Related Knowledge Questionnaire: There were recorded, and field notes were taken.
ten items that asked about causes of obesity, health effects Participant observation was conducted to observe
of obesity, healthy foods and unhealthy foods, sedentary school environment factors comprising the foods provided
lifestyle and physical activity. The ten items were scored from the school canteen and foods sold by vendors around
on a true-false scale with (0) false, (0) uncertain and the school. The observation was not limited to school heath
(1) true. The CVI was 0·9, and Kuder–Richardson Formula activities deployed as a result of school health policy and
20 (KR20) was 0·75. curriculum.
Food Consumption Questionnaire: twenty items with
pictures of various types of snacks to be chosen covered
unhealthy and healthy products during and after school Statistical analysis
at home. The twenty items were scored according to the The researcher and the core working group worked
frequency of dietary intake on a five-point scale with (1) never, together in collecting the quantitative data while the
(2) seldom, (3) sometimes, (4) often and (5) daily. The CVI was researcher proceeded with the statistically analysis. The
1·0, and the Cronbach’s alpha coefficient was 0·81. quantitative data collection included pretest, posttest
(immediately after intervention) and follow-up (2 months
Family-level variables after intervention). Quantitative data analysis involved
For family-level variables, the questionnaires were descriptive analysis, independent-sample t test, pair t test,
divided into the following three parts: family perception repeated measures ANOVA and χ2 test statistics. Child
of child obesity, family modelling and parenting practices. weight and height measurements were assessed and clas-
The respondents were the same parents/guardians or sified according to the standard sex-specific growth chart
same family members in all families participating in on weight and height for Thai citizens for nutritional
the study. status(29).
Community participate for obesity prevention 1493
Results caused some obesity to develop in in the students.
Ultimately, the obese students from the control group were
Individual level higher in number when compared with those of the exper-
imental group.
Nutritional status
The comparison points at the three times (pre-test at first Physical activity and exercise behaviours
week, post-test at the fourth month right after intervention, The scores for physical activity and exercise behaviours in
follow-up at 6 months later in both the experimental and the experimental group were significantly higher than that
control groups) showed that the nutritional status of the for the control group (F(1,270) = 28·388; P value = 0·000).
students in the control group did not improve, while the The results showed no differences at pretest, but significant
comparison points at the three times in the experimental differences between posttest and follow-up (P-value = 0·456,
group showed a nutritional-status progression of the 0·000 and 0·018, respectively). However, the results found no
students. As shown in Table 2, the results show that 0 % significant differences within the experimental group
of students in the experimental group were obese from (F(1,1·907) = 2·693; P-value = 0·071). (See Tables 3 and 4)
pretest to posttest with a slight increase to 0·7 % at Students’ Knowledge about Obesity and NCD
follow-up, while the prevalence rate of obese students in prevention – The scores for students’ knowledge about
the control group from pretest to posttest was 0 % to obesity and NCD prevention in the experimental group
1·5 %. Furthermore, this rate remained constant (1·5 %) at were significantly higher than in the control group
follow-up. After testing differences between the experi- (F(1,270) = 4·090; P value = 0·044). The results showed no
mental and control groups, no significant differences differences between pretest and posttest, but significant
were found between the groups at pretest and posttest differences at follow-up (P value = 0·947, 0·362 and
(P-value = 0·728 and 0·144, respectively), while significant 0·005, respectively). In addition, the results found
differences between the groups were found at follow-up significant differences within the experimental group
(P value = 0·032). Observations were made of the growing (F(1,1·829) = 7·806; P value = 0·001). Scores were signifi-
negative effects. Even at the beginning, obese students had cantly higher from pretest to posttest, posttest to
already been excluded from the study. It seems that during follow-up and pretest to follow-up (P value =< 0·007,
the six-month long intervention, certain factors could have 0·001, 0·001, respectively) (see Tables 5–7).

Table 2 Description of students’ nutritional status compared at three time points in the experimental (n 138) and control groups (n 134)

Pretest Posttest Follow-up

Nutritional status n % n % n %
Experimental group (n 138)
Underweight 0 0 6 4·3 8 5·8
Normal 120 87·0 123 89·1 122 88·4
Overweight 18 13·0 9 6·5 7 5·1
Obese 0 0 0 0 1 0·7
Mean SD Mean SD Mean SD

Weight 32·00 9·25 33·45 9·47 34·65 9·64

Height 135·88 11·05 137·75 11·03 140·23 11·46
Within group χ Pre × Post (df = 2) = 0·000* χ Pre × Follow up
χ Post × Follow up
(df = 3) = 0·000* (df = 6) = 0·000*
Control group (n 134)
Underweight 0 0 4 3·0 6 4·5
Normal 114 85·1 111 82·8 105 78·4
Overweight 20 14·9 17 12·7 21 15·7
Obese 0 0 2 1·5 2 1·5
Mean SD Mean SD Mean SD

Weight 35·40 11·43 36·34 12·28 37·76 12·42

Height 136·82 10·95 139·37 11·56 140·81 11·59
Within group χ Pre × Post (df = 2) = 0·000* χ Pre × Follow up
χ Post × Follow up
(df = 3) = 0·000* (df = 9) = 0·000*
Between groups χ† Pretest = 0·728 χ Posttest
(df = 3) = 0·144 χ Follow up (df = 3) = 0·032*

df, degree of freedom.

Pre × Post = compare differences between pretest and posttest.
Pre × Follow-up = compare differences between pretest and follow-up.
Post × Follow-up = compare differences between posttest and follow-up.
*P value < 0·05.
†Fisher’s exact test.
1494 PR Suwannawong et al.
Table 3 Comparison of physical activity and exercise behaviours of students between groups and within groups

Source of variables SS df MS F P value*,†

Between groups
Groups 10·475 1 10·475 28·388 0·000*
Between-group error 99·630 270 0·369
Within groups
Time 1·926 1·907 1·010 2·693 0·071
Time × Groups 5·372 1·907 2·817 7·512 0·001*
Within-group error 193·063 514·835 0·375

SS, sum of square; df , degree of freedom; MS, mean of square.

*P value < 0·05.
†Greenhouse-Geisser.

Table 4 Mean scores for the physical activity and exercise behaviours of students at three time points in the experimental and control groups†

Experimental Control group

group (138) (134)

Physical activity and exercise behaviours Mean SD Mean SD t df P value*

Pretest 3·10 0·61 3·05 0·59 0·747 270 0·456
Posttest 3·23 0·61 2·79 0·53 6·419 270 0·000*
Follow-up 3·22 0·64 3·04 0·62 2·372 270 0·018*

*P value < 0·05.

†A mean score of 1–3 points means ‘behaviour needs improvement’, a mean score of 3·1–3·95 points means ‘moderate level of behaviour’ and a mean score of 4·0–5·0 points
means ‘good level of behaviour’.

Table 5 Comparison of students’ knowledge about obesity and non-communicable chronic diseases (NCD) prevention between groups and
within groups

Source of variables SS df MS F P value*,†

Between groups
Groups 19·094 1 19·094 4·090 0·044*
Between-group error 1260·360 270 4·668
Within groups
Time 43·971 1·829 24·041 7·806 0·001*
Time × Groups 20·057 1·829 10·966 3·560 0·033*
Within-group error 1520·784 493·761 3·080

SS, sum of square; df, degree of freedom; MS, mean of square.

*P value < 0·05.
†Greenhouse-Geisser.

Table 6 Mean scores for knowledge about obesity and non-communicable chronic diseases (NCD) prevention at three time points between
the experimental and control groups†

Experimental
group (138) Control group (134)

Knowledge of students Mean SD Mean SD t df P value*

Pretest 7·69 1·65 7·70 1·57 −0·067 270 0·947
Posttest 8·34 1·94 8·14 1·63 0·913 270 0·362
Follow-up 8·47 1·92 7·74 2·30 2·849 270 0·005*

*P value < 0·05.

†A mean score of 8–10 points was interpreted as ‘high’ (> 80 %), a mean score of 6–7 points was interpreted as ‘moderate’ (60–79 %) and a mean score of 0–5 points was
interpreted as ‘low’ (< 60 %).

Family level higher than for the control group (F(1,270) = 6·191; P value
= 0·013). The results showed that there were no differences
Parents’ perception at pretest and posttest, but significant differences at follow-
The scores for parents’ perceptions about obesity and NCD up (P value = 0·760, 0·157 and 0·000, respectively). In addi-
prevention in the experimental group were significantly tion, the results found significant differences within the
Community participate for obesity prevention 1495
Table 7 Comparison of students’ knowledge about obesity and non-communicable chronic diseases (NCD) prevention
at three time points between the experimental and control groups

Students’ knowledge Mean difference Std. error P value*,†

Experimental group
Posttest – Pretest 0·652 0·211 <0·007*
Follow-up – Posttest 0·130 0·034 <0·001*
Follow-up – Pretest 0·783 0·208 <0·001*
Control group
Posttest – Pretest 0·440 0·189 <0·064
Follow-up – Posttest −0·403 0·251 <0·331
Follow-up – Pretest 0·037 0·252 <1·000

*P value < 0·05.

†Bonferroni’s method.

Table 8 Comparison of the parents’ perceptions about obesity in their children between groups and within groups

Source of variables SS df MS F P value*,†

Between groups
Groups 153·517 1 153·517 6·191 0·013*
Between-group error 6694·650 270 24·795
Within groups
Time 510·758 1·662 307·315 19·460 0·000*
Time × Groups 180·425 1·662 108·559 6·874 0·002*
Within-group error 7084·433 448·609 15·792

SS, sum of square; df, degree of freedom; MS, mean of square.

*P value < 0·05.
†Greenhouse–Geisser.

Table 9 The average mean scores of the parents’ perceptions about obesity in their children at three time points in the experimental and
control groups†

Experimental group
(138) Control group (134)

Parents’ perception score Mean SD Mean SD t df P value*

Pretest 34·23 4·07 34·39 4·35 −0·306 270 0·760
Posttest 36·17 3·73 35·52 3·75 1·420 270 0·157
Follow-up 37·16 4·54 35·04 4·24 3·967 270 0·000*

*P value < 0·05.

†A mean score of 12–23 points was interpreted as ‘poor’, a mean score of 24–35 points was interpreted as ‘moderate’ and a mean score of 36–45 points was interpreted as
‘good’.

experimental group (F(1,1·662) = 19·460; P value = 0·000). Table 10 Comparison of the parents’ perceptions about obesity in
Scores were significantly higher from pretest to posttest, their children at three time points between the experimental and
control groups
from posttest to follow-up and from pretest to follow-up
(P value =< 0·000, 0·065, 0·000, respectively) (see Parents’ perceptions Mean difference Std. error P value*,†
Tables 8–10). Experimental group
Posttest – Pretest 1·935 0·221 <0·000*
Family modelling Follow-up – Posttest 0·993 0·427 <0·065
The scores for family modelling in the experimental group Follow-up – Pretest 2·928 0·506 <0·000*
were significantly higher than the control group Control group
Posttest – Pretest 1·134 0·423 <0·025*
(F(1,270) = 13·858; P value = 0·000). The results showed Follow-up – Posttest −0·478 0·480 0·964
no difference at pretest but significant differences between Follow-up – Pretest 0·657 0·517 0·618
posttest and follow-up (P value = 0·108, 0·000 and 0·003, *P value < 0·05.
respectively). In addition, the results found significant †Bonferroni’s method.
1496 PR Suwannawong et al.
Table 11 Comparison of family modelling between groups and within groups

Source of variables SS df MS F P value*,†

Between groups
Groups 4·961 1 4·961 13·858 0·000*
Between-group error 96·660 270 0·358
Within groups
Time 7·539 1·941 3·884 12·610 0·000*
Time × Groups 8·880 1·941 4·575 14·854 0·000*
Within-Group error 161·396 524·014 0·308

SS, sum of square; df, degree of freedom; MS, mean of square.

*P value < 0·05.
†Greenhouse–Geisser.

Table 12 Mean scores of family modelling at three time points in the experimental and control groups†

Experimental group
(138) Control group (134)

Family godeling Mean SD Mean SD t df P value*

Pretest 3·50 0·66 3·63 0·60 −1·610 270 0·108
Posttest 3·61 0·54 3·23 0·36 6·822 239·173 0·000*
Follow-up 3·76 0·60 3·54 0·57 2·973 270 0·003*

*P value < 0·05.

†A mean score of 1–3 points means ‘behaviour needs improvement’, a mean score of 3·1–3·95 points means ‘moderate level of behaviour’ and a mean score of 4·0–5·0 points
means ‘good level of behaviour’.

Table 13 Comparison of family modelling at three time points Parents was appointed to work in school health
between the experimental and control groups networking. The parent team was very strong and helpful
Parents’ perceptions Mean difference Std. error P value*,† in taking policy into in the long term. Unhealthy food
vendors were banned both from school itself and from
Experimental group
Posttest – Pretest 0·102 0·072 <0·474 the areas surrounding the school. The school launched
Follow-up – Posttest 0·149 0·062 <0·050 the school lunch project suggested by the Ministry of
Follow-up – Pretest 0·251 0·072 <0·002* Public Health so that students would receive highly nutri-
Control group
Posttest – Pretest −0·400 0·059 <0·000* tious foods, avoid junk food and add vegetable and fruit
Follow-up – Posttest 0·316 0·061 <0·000* items to their menus. Regular exercise was scheduled
Follow-up – Pretest −0·084 0·070 <0·691 15 min every morning before class. The school improved
*P value < 0·05. facilities and sports areas for recreation to increase space
†Bonferroni’s method. for physical activities and exercise. The facilities for
physical activities and exercise were also improved.
differences within the experimental group (F(1,1·941) =
12·610; P value = 0·000). However, there were no signifi-
cant differences between pretest and posttest, but scores Discussion
were significantly higher from posttest to follow-up and
from pretest to follow up (P value =< 0·474, 0·050, 0·002, The programme was found to be successful in terms of
respectively) (see Tables 11–13). improving health behaviours and environments both at
home and at school, as suggested by EST. Students’ nutri-
School level tional status showed the progression of students with
From in-depth interviews and participant observation, it normal weight with likelihood of future sustainability as
was found that school health policy became more active. a result of true collaboration in the PAR process.
Health promotion and policies for preventing obesity In terms of EST at the individual level, the results
among children were integrated with school lessons in showed improvement in physical activity, exercise behav-
subjects such as physical education, health education iours and knowledge about obesity and NCD prevention
and science, all of which involve long-term activities. among students. This finding corresponds with the findings
Furthermore, there was an initiative in team setting for of previous studies that were focused on promoting exer-
school projects on preventing obesity in children. The cise in school-age children(31) and programmes that were
initiative team typically came from the research core focused on knowledge about food and nutrition and which
working group. Moreover, the Association of Leading were combined with promoting afterschool exercise(32).
Community participate for obesity prevention 1497
The findings indicate that promoting exercise at school is Previous research has illustrated interesting issues
efficient in improving the overall physical fitness of concerning health programmes aimed at preventing
students(31). In addition, it helps with improvements in four obesity in school-age children and how real participation
healthy eating behaviours: fruits, vegetables, sugar-free between family and school to make changes in both home
beverages and avoidance of unhealthy snack foods(32). and school environments together can lead to individual
For the microsystem, the results revealed that the behaviour modification(10,21). These factors are the key to
families had changed in terms of their perceptions of success, as they lead to satisfaction on all sides and result
obesity in children and family modelling behaviours in their in cooperative problem solving that is sustainable(5,38).
efforts to promote healthy children. Although the results Previous research has also found that focusing on only
show no significant difference in parenting practices one side (home or school) or only on individual behaviour
between the experimental and control groups, the qualita- modification results in less likelihood of long-term behav-
tive data show that parents are prone to be more concerned iour modification, as observed in the absence of improve-
about healthy food selection, sedentary reduction and ment in the students’ nutritional status(32,39). The reason for
regular exercise. This study was supported by empirical this finding may result from the fact that the activities
evidence indicating that parents as role models are focused only on school practice and did not include contin-
associated with young children’s weight, dietary intake uing practice at home(22,38,39).
and physical activity(33). The outstanding features of the current study are as
For the exosystem, the qualitative results revealed that follows: (1) the study was based on EST with intervention
there were noticeable improvements in school environ- efforts covering individual, family and school levels; The
ments in the aspects of health-promoting policy and problem of obesity in school-age children is complex,
curriculum integrated at school. In particular, the improve- and previous findings have yielded vague problem-solving
ment in the healthy food environments included physical guidelines, whereas sustainable solutions are required for
activity and exercise facilities at school. Three significant preventing obesity in school-age children. It is anticipated,
changes were made in the school environment: (1) integra- therefore, that this task would remain difficult if it continued
tion of obesity and NCD prevention knowledge in the towards separate solutions between school and family or
course curriculum and (2) creation of healthy policies opted to solve the problem of obesity in school-age chil-
related to obesity prevention such as modification of school dren only at school or at home because one factor may
lunch projects and healthy food vendor management. The influence others(40). Weight control within a normal range
school’s chef was trained in nutritionally based healthy for children must rely on promotion by family and school.
cooking under the recommendations of the Ministry of This need for family and school involvement means that
Public Health. This policy of healthy cooking is consistent interaction at the family and school levels can have tremen-
with the findings of Li et al.(34), who affirmed that issuing dous influence over long-term changes in behaviour(9,21).
rules for food sales in school cafeterias was correlated with This finding corresponds with the findings of previous
reducing the intake of energy-dense foods and the preva- studies in which it has been stated that preventing child-
lence of overweight conditions and obesity among school- hood obesity cannot be aimed at only one aspect.
age children. Ensuring that the school environment is filled Adjusting both home and school environments together
with more healthy than unhealthy foods led to modifica- can result in consistency and continuity of practice, both
tions in student consumer behaviours with access to at home and at school. Making this adjustment can lead
healthy foods.(35) (3) Adding an hour for exercise and facili- to sustainability in preventing obesity in school-age chil-
tating access to sports equipment and playground areas in dren(12) and strongly suggests that disconnected practice
this study correlated well with a study done by Cheng et between school and home can only adversely affect the
al.(36), who found school physical activity environments sustainability of solutions(41).
to be associated with students’ obesity with statistical (2) The programme was developed by applying the PAR
significance, including duration of physical activity during process from beginning to end. The results yielded by
school for at least 1 h/d and the provision of playground application of the PAR process in this study both support
equipment such as climbing structures and swings. and confirm the statements of previous researches that
Adequate daily exercise, access to sports and playground applied the PAR process to create numerous benefits
equipment may, therefore, facilitate school-age obesity for the participants: (1) every party had the same goal
prevention(36,37). of reducing the problem of misunderstanding(20,21);
Accordingly, times are always changing. The culture of a (2) empowerment was created among all stakeholders,
society and social conditions have changed from the tradi- which served as a key foundation for future extension(5)
tional Thai food culture. Western-style fast-food consump- and (3) participation was increased at every stage and
tion (saturated fat and sweet drink intake) plays an every party was instructed to feel love, attachment and
important role in Thai daily life. School-age children now ownership of the problem together(13,14). The PAR process
follow fashion/trends in food consumption because it adhered to the principle of participation by all parties and
represents modernity and such foods are easy to buy. promoted motivation for every party to participate in
1498 PR Suwannawong et al.
offering opinions or proposing ideas while sharing experi- be of greater benefit if only one platform or technologies
ences and knowledge together(23–41). This study also found was used rather than many. Fourth, it would be better to
that adhering to the principles of a participatory process remove some of the take-home message activities to make
based on EST from beginning to end of the process resulted the dissemination more practical. Although the results of
in work-performance clarity, reduced the problem of this study found positive impact in preventing obesity
confusion and built uniform understanding among all among school-age children, the PAR process in this study
parties. Nevertheless, the main barrier found during the included only one loop because of the time and resource
course of conducting the study was that families had no constraints of the research.
time to participate in all school activities because of time
constraints caused by working to earn a living(41–44).
Therefore, families proposed the use of technology to help Conclusions
reduce low participation and to improve the communica-
tion linkage between school and family. This study is This study reveals strategic methodologies for obesity
supported by previous research stating that applying these prevention among school-age children to meet the
technologies enhances strong participation and linkage need of encompassing environmental linkages from indi-
between school and family(45). Moreover, the collaboration vidual, family and school levels. The investigation that
between school and family to design child obesity preven- was conducted revealed the programme to be successful
tion video clips resulted in media that met the needs and and led to improvements in health behaviours and environ-
was more interesting to follow(46). ments. The students’ nutritional-status outcomes indicated
The purpose of this study was to formulate policies for successful progression with a likelihood of future sustain-
controlling obesity among school-age children. These poli- ability. Developing community participation programmes
cies were found to be more effective because (1) good rela- based on the EST and PAR process provides suitable guid-
tionships were built, which led to a mutual willingness to ance for future obesity prevention. Nevertheless, students
hear opinions and needs of each party(20–22,42); (2) oppor- served as excellent bridges as a linkage between home and
tunities were offered for every party to participate in the school. The outstanding strategies involved applying social
problem-solving process, which led to new ideas(22,44); media platforms as communication channels leading to
(3) parents/guardians and teachers were empowered; such increased participation.
empowerment gave every party confidence in solving their
problems(21) and (4) family and school felt a joint respon-
sibility, which can result in the creation of sustainable
policies in the future(21). In terms of context, this type of Acknowledgements
intervention is feasible in low(47), middle-income(20,21)
and high-income countries(22). Acknowledgements: The study was supported
Nevertheless, there were some limitations in this study. for publication by Faculty of Public Health, Mahidol
First, barriers might be encountered concerning access to University, Bangkok, Thailand. The authors would like to
technologies by elders and time constraints for family express deep appreciation to families for their constant
participation because of the need of parents/guardians to care. Special thanks go to all stakeholders at Anuban
work and support their families(20,21). Therefore, planning Aong Karuk School and Anuban Nakhon Nayok School,
strategies for family participation must be given primary Nakhon Nayok Province, Thailand, for their kind arrange-
consideration and not be overlooked. Second, during the ment, cooperation and participation in the research.
PAR process, there were some difficulties among stake- Financial support: This research was partially supported
holders in terms of making the exact dates for appoint- by the Graduate Studies Alumni Association and Faculty
ments. This situation might have caused them to some of Graduate Studies and Department of Public Health
extend the delays in their schedules(43). Third, the platforms Nursing, Faculty of Public Health, Mahidol University,
for disseminating information suggested by the stake- Bangkok, Thailand (No grant number). Graduate Studies
holders included Facebook groups, LINE groups and a Alumni Association and Faculty of Graduate Studies and
YouTube Channel, all of which were considered highly Department of Public Health Nursing had no role in the
beneficial communication channels. However, the limita- design, analysis or writing of this article. Conflict of interest:
tion in terms of the process of validation with the target There are no conflicts of interest. Authorship: PRS (study
audience could lead to a limitation of the study regarding design, data collection, data analysis, manuscript prepara-
its methodological aspects. In particular, families with tion), NA (study design, data collection, data analysis and
elders would be less likely to be familiar with such plat- manuscript preparation), AP (study design, data analysis
forms. This limitation involved the potential of exclusion and manuscript preparation), RJ (study design, data
and abandonment from their inability to use such plat- analysis, manuscript preparation). Ethics of human subject
forms. However, a reflection of results resulting during participation: This study was conducted according to the
the evaluation period verified the conclusion that it would guidelines laid down in the Declaration of Helsinki, and
Community participate for obesity prevention 1499
all procedures involving research study participants were s12889-016-3118-6.
approved by the Ethical Committee on Research in 12. Hung LS, Tidwell DK, Hall ME et al. (2015) A meta-analysis of
school-based obesity prevention programs demonstrates
Human Subjects (approval number MUPH 2018–012), limited efficacy of decreasing childhood obesity. Nutr Res
Faculty of Public Health, Mahidol University. Written 35, 229–240. doi: 10.1016/j.nutres.2015.01.002.
informed consent was obtained from all subjects/patients. 13. Bronfenbenner U (1979) The Ecology of Human
Development: Experiments by Nature and Design. Cambridge:
Harvard University Press.
14. Bronfenbenner U (1970) Two Worlds of Childhood. London:
Supplementary material Allen.
15. Davison KK, Jurkowski JM & Lawson HA (2013)
For supplementary material/s referred to in this article, Reframing family-centred obesity prevention using the
Family Ecological Model. Public Health Nutr 16,
please visit https://doi.org/10.1017/S136898002300040X 1861–1869. https://doi.org/10.1017/S1368980012004533
16. Feng L, Wei D-M, Lin S-T et al. (2017) Systematic review and
meta-analysis of school-based obesity interventions in main-
References land China. PLoS ONE 12, 1–19. doi: 10.1371/journal.pone.
0184704.
17. Pérez Solís D, Díaz Martín JJ, Álvarez Caro F et al.
1. Fryar CD, Carroll MD & Ogden CL (2018) Prevalence of
(2015) Effectiveness of a school-based program to prevent
Overweight, Obesity, and Severe Obesity among Children
obesity. An Pediatr 83, 19–25. doi: 10.1016/j.anpede.2015.
and Adolescents Aged 2–19 years: United States, 1963–1965
06.002.
through 2015–2016. NCHS Health E-Stats (Internet). https://
18. Hebestreit A, Intemann T, Siani A et al. (2017) Dietary
www.cdc.gov/nchs/data/hestat/obesity_child_15_16/obesity_
patterns of European children and their parents in associa-
child_15_16.htm (accessed January 2021).
tion with family food environment: results from the I.
2. Hales CM, Carroll MD, Fryar CD et al. (2017) Prevalence of
Family study. Nutrient 9, 126. doi: 10.3390/nu9020126.
Obesity among Adults and Youth: United States, 2015–2016.
19. Zhou N & Cheah CS (2015) Ecological risk model of
NCHS Data Br. (Internet). 288, 1–8. https://www.cdc.gov/
childhood obesity in Chinese immigrant children. Appetite
nchs/products/databriefs/db288.htm#Suggested_citation
90, 99–107. doi: 10.1016/j.appet.2015.02.028.
(accessed February 2022).
20. Sirikulchayanonta C, Pavadhgul P, Chongsuwat R et al.
3. Baird J, Jacob C, Barker M et al. (2017) Developmental
(2011) Participatory action project in reducing childhood
origins of health and disease: a lifecourse approach to the
obesity in Thai primary schools. Asia Pac J Public Health
prevention of non-communicable diseases. Healthcare 5,
23, 917–927. doi: 10.1177/1010539510361965.
14. doi: 10.3390/healthcare5010014.
21. Chotibang J, Fongkaew W, Mo-suwan L et al. (2010)
4. Rerksuppaphol S & Rerksuppaphol L (2015) Metabolic
Development of a Family and School Collaborative (FASC)
syndrome in obese Thai children: defined using modified
program to promote healthy eating and physical activity
‘The National Cholesterol Education Program/Adult
among school-age children. Thai J Nurs Res 13, 133–147.
Treatment Panel III’ criteria. J Med Assoc Thai 98, S88–S95.
22. Anselma M, Altenburg TM, Emke H et al. (2019)
5. Sirikulchayanonta C, Sirikulchayanonta V, Suriyaprom K
Co-designing obesity prevention interventions together with
et al. (2022) Changing trends of obesity and lipid profiles
children: intervention mapping meets youth-led participa-
among Bangkok school children after comprehensive
tory action research. Int J Behav Nutr Phys Act 16, 130.
management of the bright and healthy Thai kid project.
doi: 10.1186/s12966-019-0891-5.
BMC Public Health 22, 1–10. doi: 10.1186/s12889-022-
23. Coghlan D & Brydon-Miller M (2014) The SAGE Encyclopedia
13712-w.
of Action Research. Thousand Oaks: SAGE Publications.
6. Llewellyn A, Simmonds M, Owen CG et al. (2016) Childhood
24. Lambrinou C-P, Androutsos O, Karaglani E et al. (2020)
obesity as a predictor of morbidity in adulthood: a systematic
Effective strategies for childhood obesity prevention
review and meta-analysis. Obes Rev 17, 56–67. doi: 10.1111/
via school based, family involved interventions: a critical
obr.12316.
review for the development of the Feel4Diabetes-study
7. Walker M, Nixon S, Haines J et al. (2019) Examining
school-based component. BMC Endocr Disorders 20,
risk factors for overweight and obesity in children with
1–20. doi: 10.1186/s12902-020-0526-5.
disabilities: a commentary on Bronfenbrenner’s ecological
25. Chan CL, Tan PY & Gong YY (2022) Evaluating the impacts of
systems framework. Dev Neurorehabilitation 22, 359–364.
school garden-based programmes on diet and nutrition-
doi: 10.1080/17518423.2018.1523241.
related knowledge, attitudes and practices among the school
8. Wang X, Hu J, Huang S et al. (2022) Exploring overweight
children: a systematic review. BMC Public Health 22, 1251.
risk trajectories during childhood and their associations with
doi: 10.1186/s12889-022-13587-x.
elevated blood pressure at late adolescence: a retrospective
26. Taylor JC, Zidenberg-Cherr S, Linnell JD et al. (2018)
cohort study. Hypertension 79, 1605–1613. doi: https://doi.
Impact of a multicomponent, school-based nutrition inter-
org/10.1161/HYPERTENSIONAHA.121.18714
vention on students’ lunchtime fruit and vegetable avail-
9. Moraeus L, Lissner L, Yngve A et al. (2012) Multi-level
ability and intake: a pilot study evaluating the Shaping
influences on childhood obesity in Sweden: societal factors,
Healthy Choices Program. J Hunger Environ Nutr 13,
parental determinants and child’s lifestyle. Int J Obes 36,
415–428. doi: 10.1080/19320248.2017.1374899.
969–976. doi: 10.1038/ijo.2012.79.
27. McIntosh B, Daly A, Mâsse LC et al. (2015) Sustainable child-
10. Kothandan SK (2014) School based interventions v. family
hood obesity prevention through community engagement
based interventions in the treatment of childhood obesity –
(SCOPE) program: evaluation of the implementation
a systematic review. Arch Public Health 72, 3–3. doi: 10.
phase. Biochemistry Cell Biol 93, 472–478. doi: 10.1139/
1186/2049-3258-72-3.
bcb-2014-0127.
11. Cunningham-Sabo L, Lohse B, Smith S et al. (2016) Fuel for
28. Kemmis S & Mc Taggart R (1990) The Action Research
fun: a cluster-randomized controlled study of cooking skills,
Planner, 3rd ed. Victoria: Brown Prior Anderson National
eating behaviors, and physical activity of 4th graders and
Library of Australia Catalouging in Publication Data.
their families. BMC Public Health 16, 444. doi: 10.1186/
1500 PR Suwannawong et al.
29. Department of Health (2021) AnamaiMedia (Internet). 39. Hill JL, Zoellner JM, You W et al. (2019) Participatory
https://nutrition2.anamai.moph.go.th/th/cms-of-13/175117 development and pilot testing of iChoose: an adaptation of
(accessed September 2022). an evidence-based paediatric weight management program
30. Conlon BA, McGinn AP, Lounsbury DW et al. (2015) The for community implementation. BMC Public Health 19,
role of parenting practices in the home environment among 1–16. doi: 10.1186/s12889-019-6450-9.
underserved youth. Child Obes 11, 394–405. doi: 10.1089/ 40. Eisenmann J, Gentile D, Welk G et al. (2008) SWITCH:
chi.2014.0093. rationale, design, and implementation of a community,
31. Song JH, Song HH & Kim S (2021) Effects of school-based school, and family-based intervention to modify behaviors
exercise program on obesity and physical fitness of urban related to childhood obesity. BMC Public Health 8, 223.
youth: a quasi-experiment. Healthcare 9, 358. doi: 10. doi: 10.1186/1471-2458-8-223.
3390/healthcare9030358. 41. Phaitrakoon J, Powwattana A, Lagampan S et al. (2014)
32. Rieder J, Moon JY, Joels J et al. (2021) Trends in health Effects of an obesity control program for Thai elementary
behavior and weight outcomes following enhanced after- school children: a quasi-experimental study. Pac Rim Int J
school programming participation. BMC Public Health 21, Nurs Res Thail 18, 290.
672. doi: 10.1186/s12889-021-10700-4. 42. Stringer ET & Genat WJ (2004) Action Research in Health.
33. Kaiser L, Aguilera A, Horowitz M et al. (2015) Correlates New Jersey: Pearson Education.
of food patterns in young Latino children at high risk of 43. Øen G & Stormark KM (2013) Participatory action
obesity. Public Health Nutr 18, 3042–3050. doi: 10.1017/ research in the implementing process of evidence-based
S1368980014003309. intervention to prevent childhood obesity: project
34. Li M, Xue H, Wen M et al. (2017) Nutrition and physical design of the ‘Healthy Future’ study. J Obes 2013, 1–10.
activity related school environment/policy factors and child doi: 10·1155/2013/437206
obesity in China: a nationally representative study of 8573 44. Norman Å, Nyberg G & Berlin A (2019) School-based
students in 110 middle schools. Pediatr Obes 12, 485–493. obesity prevention for busy low-income families
doi: 10.1111/ijpo.12169. organisational and personal barriers and facilitators to imple-
35. Jia P, Li M, Xue H et al. (2017) School environment and mentation. PLoS ONE 14, 1–19. doi: 10.1371/journal.pone.
policies, child eating behavior and overweight/obesity in 0224512
urban China: the childhood obesity study in China megac- 45. Fonseca H, Prioste A, Sousa P et al. (2016) Effectiveness
ities. Int J Obes 41, 813–819. doi: 10.1038/ijo.2017.2. analysis of an internet- based intervention for overweight
36. Cheng L, Li Q, Hebestreit A et al. (2020) The associations adolescents: next steps for researchers and clinicians. BMC
of specific school- and individual-level characteristics Obes 3, 15. doi: 10.1186/s40608-016-0094-4.
with obesity among primary school children in Beijing, 46. Schober DJ, Sella AC, Fernandez C et al. (2016) Participatory
China. Public Health Nutr 23, 1838–1845. doi: 10.1017/ action research to develop nutrition education videos
S1368980019004592. for child care providers: The Omaha Nutrition Education
37. Sprengeler O, Wirsik N, Hebestreit A et al. (2017) Domain- Collaborative. Pedagogy Health Promot 2, 244–50.
specific self-reported and objectively measured physical doi: 10.1177/2373379915627669
activity in children. Int J Environ Res Public Health 14, 47. Xu F, Ware RS, Leslie E et al. (2015) Effectiveness of a
242. doi: 10.3390/ijerph14030242. randomized controlled lifestyle intervention to prevent
38. Bottorff JL, Huisken A, Hopkins M et al. (2020) A RE-AIM obesity among Chinese primary school students: CLICK-
evaluation of healthy together: a family-centred program to Obesity study. PLoS ONE 10, 1–12. doi: 10.1371/journal.
support children’s healthy weights. BMC Public Health 20, pone.0141421
1–12. doi: 10.1186/s12889-020-09737-8.