This is an open-access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.
Serious games have the potential to promote health behavior. Because overweight is still a major issue among secondary vocational education students in the Netherlands, this study piloted the effects of “Balance It,” a serious self-regulation game intervention targeting students’ overweight-related behaviors: dietary intake and physical activity (PA).
We aimed to pilot the effects of Balance It on secondary vocational education students’ dietary intake and PA.
In total, 501 secondary vocational education students participated at baseline (intervention: n=250; control: n=251) in this pre-post cluster randomized trial. After 4 weeks, at immediate posttest, 231 students filled in the posttest questionnaire (intervention: n=105; control: n=126). The sample had a mean age of 17.28 (SD 1.26, range 15-21) years, 62.8% (145/231) were female, and 26.8% (62/231) had a non-Dutch background. Body mass index (BMI kg/m2) ranged from 14.4 to 31.1 (mean 21.1, SD 3.3). The intervention and control groups were compared on the primary (behavioral) outcomes of dietary intake (fruit and vegetable consumption, snack consumption, and soft drink consumption) and PA (moderate and vigorous). Additionally, we explored (1) differences between the intervention and control groups in determinants of dietary intake and PA, including attitude, self-efficacy, intention, barrier identification, action planning, and action control, and (2) differences between active (intervention) users and the control group in dietary intake, PA, and associated determinants.
After corrections for multiple testing, we did not find significant differences between the intervention group and control group in terms of dietary intake, PA, and determinants of dietary intake and PA. Exploratory research indicated that only 27.6% (29/105) of the intervention group reported actual intervention use (ie, active users). For exploratory reasons, we compared the active users (n=29) with the control group (n=124) and corrected for multiple testing. Results showed that active users’ snack consumption decreased more strongly (active users: mean change=–0.20; control group: mean change=–0.08; beta=–0.36,
The Balance It intervention did not show favorable effects on dietary intake and PA compared to the control condition. However, only a small number of people in the intervention condition actually used Balance It (27.6%). Exploratory analyses did suggest that, if used as planned, Balance It could contribute to changing dietary intake and PA behaviors, albeit it remains debatable whether this would be sufficient to prevent overweight.
Overweight and obesity are related to various chronic health problems, including type 2 diabetes mellitus, cardiovascular disease, cancer, and also psychosocial problems [
Recent advances in technology enable researchers to tailor dietary intake and PA interventions to the needs of the target population. Moreover, it is possible to design a program that is cost effective, that has a wide reach, and that can function as a standalone program [
Balance It combines behavior change techniques derived from self-regulation theory [
A cluster randomized trial was conducted in 2014/2015 with measurements taken at baseline, immediately posttest (after 4 weeks of game play), and at a 4-week follow-up. Fifteen vocational education schools in the Netherlands were approached to participate in this study. In total, 4 schools agreed to participate and were randomly assigned to the intervention or waiting list control group. To counteract contamination effects between participant groups and to increase participants’ compliance with the study, random allocation to conditions took place at the level of schools [
The power calculation (alpha=.05, beta=.80) was based on a mean effect size of 31 for dietary intake and PA intervention [
One week before the study, participants received passive consent forms addressed to their parents or caregivers. At baseline, a research assistant went to the schools to introduce the study and to collect the survey data. In total, 238 students gave their consent to participate in this study. At baseline, participants filled out an online baseline questionnaire regarding their mean dietary intake and PA, social cognitive factors (ie, attitude, self-efficacy, and intention), perceived barriers, self-regulation skills, action planning, and action control. After they finished the questionnaire, participants received a link to the Balance It website and further instructions about downloading the Balance It app from the research assistant. All students received a posttest questionnaire 4 weeks after the baseline measure was taken. Participatory incentives of €20 vouchers were randomly distributed among participants. The chance of winning a voucher increased with the number of measures completed (one measure 1:8, two measures 2:8).
Balance It was designed as a tailored, interactive multimedia game in which each game could be played either individually or competitively with others, at any time and place desired. It was designed as an educational, strategic game that could be played on a daily basis for 4 continuing weeks or on a weekly basis for 6 continuing weeks. Within each game, players set their own graded tasks (eg, to eat two pieces of fruit per day), which were selected from a multiple-choice list (
Screenshots of task initiation in the Balance It app.
At baseline, the control group was instructed to fill in the baseline questionnaire and informed that the researcher would return in 4 weeks for a posttest measure. Between measures, no interventions were offered by the researchers. Immediately following the posttest, students were provided with information about Balance It and given the opportunity to play.
The assessment of dietary intake was derived from a validated food frequency questionnaire [
The PA measures were derived from the Injuries and Physical Activity in the Netherlands (“Ongevallen en Bewegen in Nederland”) questionnaire (validated; [
In addition to the behavioral outcomes, social cognitive factors were measured for healthy dietary intake (fruit and vegetable intake) and unhealthy dietary intake (snacks, sweets, and soft drink consumption). All measures of determinants were preceded by a stem, followed by the behavioral outcome measures as subcategories.
Attitudes toward dietary intake were assessed by three items using semantic differential response scales, such as “I think that eating two pieces of fruit a day is...” (1=very bad to 5=very good; 1=very unpleasant to 5=very pleasant; 1=very unhealthy to 5=very healthy) derived from [
Self-efficacy toward dietary intake was assessed by one item preceded by a question stem: “If I want to, I am capable of...” Items were derived from van der Horst et al [
Dietary intake intention was assessed with one item preceded by the stem: “I planned to...” derived from [
Barriers to healthy dietary intake were assessed separately from barriers to unhealthy dietary intake because different barriers influence fruit and vegetable intake and unhealthy dietary intake. Healthy dietary intake was assessed with five items using 5-point Likert scales: “I am capable of eating sufficient fruit and vegetables, also when I am...” Response options ranged from 1=definitely not to 5=definitely. Subcategories referred to when I am alone, during the weekend, when I am in a hurry, when I experience difficulties preparing fruits and vegetables, and when there is a lack of choice. Items were derived from previous measures [
Barriers to unhealthy dietary intake were assessed with 13 items using 5-point Likert scales: “I am capable of eating a limited amount of unhealthy snacks, also when I am...” Response options ranged from 1=definitely not to 5=definitely. Subcategories referred to physical settings (eg, when I am at home), sedentary activities (eg, when I am watching TV), social settings (eg, when I am at a party), and mood (eg, when I am sad). Items were derived from previous measures [
Action planning in terms of dietary intake was assessed by four items, such as “I have a clear plan for when I...” Response options ranged from 1=definitely not to 5=definitely. Subcategories referred to when, where, how, and how often participants planned to eat more healthy or less unhealthy foods (derived from [
Action control in terms of dietary intake was measured with four items using 5-point Likert scales, such as “During the last month, I have constantly monitored my...” Response options ranged from 1=definitely not to 5=definitely. Subcategories referred to self-monitoring of fruit and vegetable consumption, awareness of fruit and vegetable standards, self-regulatory effort to eat more healthy and less unhealthy foods, and self-regulatory effort to conform to norm behavior (eg, eat two pieces of fruits a day) (derived from [
Social cognitive factors were also measured for moderate PA (eg, walking and cycling) and vigorous PA (eg, exercising). All measures of PA determinants were preceded by a stem followed by the behavioral outcome measures subcategories.
Attitudes toward PA was assessed by three items using semantic differential response scales, such as “I think that exercising is...” (1=very bad to 5=very good; 1=very unpleasant to 5=very pleasant; 1=very unhealthy to 5=very healthy) (derived from [
Self-efficacy toward PA was assessed by one item preceded by a question stem: “If I want to, I am capable of...” Items were derived from Van der Horst et al [
Intention was assessed with one item preceded by the stem: “I planned to...” (derived from [
Barriers to PA were assessed with seven items using 5-point Likert scales, such as “I am capable of being more physically active, also when I am...” Response options ranged from 1=definitely not to 5=definitely. Subcategories referred to when I am busy, when I am stressed, if I failed last time, when I am tired, when it is raining, if I do not have the time, and if I do not get social support (derived from [
Action planning in terms of PA was assessed by four items using 5-point Likert scales, such as “I have a clear plan for when I...” Response options ranged from 1=definitely not to 5=definitely. Subcategories referred to when, where, how, and how often participants planned to be more physically active (derived from [
Action control in terms of PA was measured with four items using 5-point Likert scales, such as “During the last month, I have constantly monitored my...” Response options ranged from 1=definitely not to 5=definitely. Subcategories referred to self-monitoring of PA, awareness of PA standards, self-regulatory effort to be more physically active, and self-regulatory effort to conform to norm behavior (eg, to be moderately active) (derived from [
Items regarding gender, age, BMI, educational level, cultural background, accommodation, and living situation were included at the beginning of the baseline measure. Ethnicity was defined according to the procedures of Statistics Netherlands; individuals were considered to have a Dutch background if both parents were born in the Netherlands. If one of the parents was born outside the Netherlands, the student was considered to have a non-Dutch background [
To evaluate subjective experience of using the Balance It app or website, 19 items regarding the Balance It app in general were preceded by the stem: “What did you think of...” Response options ranged from 1 (very bad) to 5 (very good) (compare attitude measures [
Descriptive statistics were used to characterize both study groups at baseline (ie, gender, age, educational level, ethnicity, and and body mass index [BMI]). Chi-square tests and
In total, 501 students were invited to participate in this study (intervention: n=250; control: n=251;
Of all participants who remained in the intervention group at posttest (n=105), 27.6% (29/105) reported actual intervention use. Compared to the control group (n=200), self-reported active users were less likely to follow vocational education related to care and well-being (
Flow diagram of the enrollment and selection of study participants.
Demographic background of Balance It participants at baseline (N=231).
Demographic variables | Intervention group (n=105) | Control group (n=126) | χ2 (df) | |||
Age (years), mean (SD) | 16.96 (1.10) | 17.52 (1.36) | –3.34 | .003 | ||
Gender (male), n (%) | 39 (37.1) | 47 (37.3) | 0.0 (1) | .93 | ||
Ethnicity (Dutch), n (%) | 77 (73.3) | 92 (73.0) | 0.1 (1) | .77 | ||
Care and well-being | 25 (23.8) | 115 (91.3) | 108.4 (1) | .001 | ||
Economics | 71 (67.6) | 0 (0.0) | 128.9 (1) | .001 | ||
Technique | 1 (1.0) | 2 (1.6) | 0.2 (1) | .69 | ||
Social work | 0 (0.0) | 1 (1.0) | 0.8 (1) | .55 | ||
0.5 (1) | .49 | |||||
Level 3 | 5 (4.8) | 9 (7.1) | ||||
Level 4 | 95 (90.5) | 115 (91.3) | ||||
39.3 (1) | .001 | |||||
Year 1 | 100 (100.0) | 84 (67.7) | ||||
Year 2 | 0 (0.0) | 40 (34) | ||||
5.3 (4) | .26 | |||||
Both parents | 79 (76.7) | 100 (79.4) | ||||
One parent | 19 (18.4) | 16 (12.7) | ||||
Alone | 2 (1.9)) | 2 (1.6) | ||||
Other | 3 (2.9) | 8 (6.3) | ||||
4.7 (3) | .66 | |||||
Underweight (BMI <18.5) | 8 (12.7) | 10 (11.2) | ||||
Normal weight (BMI 18.5-25) | 47 (74.6) | 58 (65.1) | ||||
Overweight (BMI 25-30) | 8 (12.7) | 18 (20.2) | ||||
Obese (BMI >30) | 0 (0.0) | 3 (3.4) |
Effects of Balance It on behavioral outcomes and determinants.a
Outcome variable | T0, mean (SD) | T1, mean (SD) | T0-T1, ∆ mean | Difference test | |||||
Intervention |
Control |
Intervention |
Control |
Intervention | Control | B |
|||
Fruit intake |
0.81 |
0.80 |
1.05 |
0.81 |
0.14 | 0.01 | 0.21 |
.01 | |
Vegetable intake |
1.26 |
1.32 |
1.21 |
1.28 |
–0.05 | –0.04 | –0.03 |
.00 | |
Snack consumption |
0.91 |
0.98 |
0.86 |
0.90 |
–0.05 | –0.08 | 0.01 |
.00 | |
Soft drink consumption |
1.07 |
1.11 |
0.92 |
1.07 |
–0.15 | –0.04 | –0.25 |
.03 | |
Moderate PA |
4.30 |
3.82 |
3.91 |
3.31 |
–0.39 | –0.51 | 0.20 |
.00 | |
Vigorous PA |
5.21 |
5.25 |
4.74 |
4.78 |
–0.47 | –0.47 | 0.10 |
.00 | |
Active transport |
2.55 |
2.50 |
3.20 |
2.38 |
0.65 | –0.12 | 0.94 |
.02 | |
Attitude | 4.01 |
3.98 |
3.93 |
4.00 |
–0.08 | 0.02 | –0.26 |
.02 | |
Self-efficacy | 4.33 |
4.29 |
4.04 |
4.13 |
–0.29 | –0.16 | –0.44 |
.02 | |
Intention | 3.86 |
3.73 |
3.66 |
3.67 |
–0.20 | –0.06 | –0.32 |
.01 | |
Perceived barriers | 3.45 |
3.52 |
3.53 |
3.49 |
0.08 | –0.03 | 0.13 |
.00 | |
Action planning | 3.04 |
3.05 |
3.37 |
2.86 |
0.33 | –0.19 | 0.36 |
.01 | |
Action control | 2.89 |
2.80 |
3.40 |
2.86 |
0.51 | 0.06 | 0.53 |
.02 | |
Attitude | 3.63 |
3.42 |
3.72 |
3.63 |
0.09 | 0.19 | –0.23 |
.01 | |
Self-efficacy | 4.28 |
4.10 |
3.92 |
3.91 |
–0.36 | –0.19 | –0.37 |
.01 | |
Intention | 3.75 |
3.40 |
3.60 |
3.35 |
–0.15 | –0.05 | –0.14 |
.00 | |
Perceived barriers | 3.52 |
3.35 |
3.48 |
3.37 |
–0.04 | 0.02 | 0.21 |
.01 | |
Action planning | 3.11 |
2.99 |
3.37 |
3.06 |
0.27 | 0.07 | 0.33 |
.01 | |
Action control | 2.96 |
2.78 |
3.28 |
2.73 |
0.32 | –0.05 | 0.48 |
.02 | |
Attitude | 4.21 |
4.21 |
4.02 |
4.13 |
–0.20 | –0.09 | –0.18 |
.01 | |
Self-efficacy | 4.42 |
4.30 |
3.98 |
4.08 |
–0.44 | –0.22 | –0.39 |
.02 | |
Intention | 4.01 |
3.87 |
3.70 |
3.68 |
–0.31 | –0.19 | –0.44 |
.02 | |
Perceived barriers | 3.47 |
3.16 |
3.39 |
3.16 |
–0.08 | 0.00 | –0.03 |
.00 | |
Action planning | 3.26 |
3.21 |
3.43 |
3.16 |
0.17 | –0.05 | 0.27 |
.01 | |
Action control | 3.05 |
2.93 |
3.38 |
2.82 |
0.33 | –0.11 | 0.60 |
.03 |
a Differences between the intervention group and control group at posttest measurement are derived via linear regression analyses for linear variables (B and 95% CI are reported), correcting for the baseline score of Y, and demographic variables for which differences were found between groups at baseline (age, vocational education sector, year of education, and the use of active transport); corrected for multiple testing (based on false discovery rate).
b Except for action planning and action control for fruit and vegetable intake, snack and soft drink consumption, and PA (n=99).
c Except for action planning and action control for fruit and vegetable intake, snack and soft drink consumption, and PA (n=124).
d Except for fruit and vegetable intake and snack and soft drink consumption (n=126), moderate PA (n=124), and active transport (n=123) for behavioral outcomes, and as follows for fruit and vegetable intake, snack and soft drink consumption, and PA: attitude (n=99), self-efficacy and intention (n=96), perceived barriers (n=95), and action planning and action control (n=92).
e Except for fruit and vegetable intake and soft drink consumption (n=104), moderate PA (n=101), vigorous PA (n=98), and active transport (n=99) for behavioral outcomes, and action planning and action control for fruit and vegetable intake, snack and soft drink consumption, and PA (n=124).
f Physical Activity.
Change scores for the intervention group were compared with change scores for the control group for both behavioral (primary) outcome measures and determinants (secondary outcome measures). All findings of the linear regressions are presented in
After correcting for multiple testing, we did not find significant differences in change scores between the intervention group and the control group for dietary intake and PA (see
The same regression analyses performed to compare the intervention and control groups were also used to compare the groups “active users” in the intervention group (29/103, 28.2%) and the control group (n=124). Allocation to these groups was based on self-reported intervention use. There were no significant baseline differences between active and nonactive users in the intervention group.
Compared to the control group, active users were more likely to participate in the economics vocational education sector (χ21=90.4,
After correcting for multiple testing, we found that active users reported marginally stronger increases in fruit intake (active users: mean change=0.51; control group: mean change=0.01; beta=0.34,
User data (an objective measure) showed that Balance It was played 771 times in total. These games primarily consisted of daily tasks (671/771, 87.0%) and individual game play (632/771, 82.0%). Of all the goals set (ie, type of tasks), players chose to improve their fruit intake in 15.0% of all cases (116/771 of which 44.0%, 51/116 of the goals were accomplished), 3.0% opted to increase their vegetable intake (23/771 of which 39%, 9/23 of the goals were accomplished), 29.1% opted to decrease their snack consumption (224/771 of which 70.1%, 157/224 of the goals were accomplished), 8.9% opted to decrease their soft drink consumption (69/771 of which 63%, 44/69 of soft drink-related goals were accomplished), 31.0% opted to increase their moderate PA (239/771 of which 54.8%, 131/239 of moderate PA goals were accomplished), and 13.0% opted to increase their vigorous PA (100/771 of which 39%, 39/100 of vigorous PA goals were accomplished). Goal accomplishment was more likely when participants were motivated (OR 2.6, 95% CI 1.9-3.5), and less likely when they did not have the time (OR 0.6, 95% CI 0.4-0.9) or when they experienced the location as a barrier (OR 0.6, 95% CI 0.4-0.9).
At posttest, 50% (15/29) of the participants who used the intervention reported that they played Balance It because they wished to have a healthier lifestyle, and 50% (14/29) played the game because they were asked to for the purpose of our study (29/103). Of the participants who did not play Balance It, 24% (18/74) reported that they did not have the time to play. The participants who used the intervention were, on average, neutral to positive about the Balance It app. When asked whether they were planning to recommend Balance It to others, participants gave a mean score of 3.14 (SD 1.03) on a scale ranging from 1 (very bad) to 5 (very good); likewise, the mean rating for the tutorial (using the same scale) was 3.72 (SD 0.75). Also, the specific game elements were evaluated neutrally to positively, on average, ranging from 3.43 (SD 1.00) on a scale ranging from 1 (very stupid) to 5 (very nice) for the construction worker, to 3.62 (SD 0.90) for the option of using special powers on the tower of an opponent. The mean overall rating given for the Balance It app (on a scale of 1 to 10, 1=the lowest grade, 10=the highest grade) was 6.71 (SD 1.96). The mean overall rating for the website (using the same scale) was 6.50 (SD 1.40).
The aim of this study was to pilot the effects of Balance It, a serious game intervention targeting secondary vocational education students’ dietary intake and PA.
No significant differences between the intervention and control groups in terms of dietary intake and PA (the primary outcomes) were observed. Additional exploratory analyses did not reveal significant differences in change scores between the intervention and control group in terms of psychological determinants of dietary intake and PA, as targeted by Balance It.
The study also revealed that the number of people that used the Balance It intervention was less than expected because only 27.6% used it as intended. For exploratory purposes, we examined the potential of Balance It among active users by comparing participants in the intervention group who reported that they had used the intervention with the control group. We did find that active users increased their fruit consumption marginally and active transport significantly, and showed stronger decreases in snack consumption compared to the control group. Although we should acknowledge that other factors could explain these differences (ie, self-selection), the findings could indicate Balance It may contribute to changes in PA and dietary intake if used as planned.
Taking into account that a difference of 100 kcal in daily caloric intake/expenditure can contribute to overweight prevention [
Consistent with previous serious gaming studies targeting dietary intake and PA, our results suggest that the use of a self-regulation game intervention could improve dietary intake and active transport among youth [
Previous research shows that youth from low SES families are less engaged with health behaviors and not as successful in terms of translating their health intentions into behavior [
Some limitations of this study should be acknowledged. First, the study is a cluster randomized trial, which was chosen over a randomized controlled trial because of practical considerations (ie, school coordinators who wanted their students to be in the same condition), to prevent contamination effects and to enhance participant compliance [
Finally, it should be noted that despite the potential of the peer-support component as included in the Balance It website to increase intervention effects on self-regulation skills [
The Balance It intervention did not show favorable effects on dietary intake and PA compared to the control condition. However, only a small number of people in the intervention condition actually used Balance It (27.6%). Exploratory analyses did suggest that, if used as planned, Balance It could contribute to changing dietary intake and PA behaviors, albeit it remains debatable whether this would be sufficient to prevent overweight.
body mass index
odds ratio
physical activity
socioeconomic status
None declared.