Professional Documents
Culture Documents
DOI: 10.1111/sms.13539
REVIEW ARTICLE
1
Department of Physiotherapy, School of
Technology and Sciences, São Paulo State
Objective: To investigate the effectiveness of active video games (AVGs) on obe-
University (UNESP), Presidente Prudente, sity‐related outcomes and physical activity levels in children and adolescents.
Brazil Design: Systematic review with meta‐analysis.
2
Department of Physiotherapy, Universidade
Methods: Literature search was performed in five electronic databases and the main
Federal de Minas Gerais (UFMG), Belo
Horizonte, Brazil clinical trials registries. Randomized controlled trials investigating the effect of
3
Department of Physical Education, School AVGs compared with no/minimal intervention on obesity‐related outcomes (body
of Technology and Sciences, São Paulo mass index [BMI], body weight, body fat, and waist circumference) and physical
State University (UNESP), Presidente
Prudente, Brazil
activity levels of children and adolescents were eligible. Two independent reviewers
4
Centro Universitário UNA, Contagem, extracted the data of each included study. PEDro scale was used to assess risk of bias
Brazil and GRADE approach to evaluate overall quality of evidence. Pooled estimates were
5
Department of Education, School of obtained using random effect models.
Technology and Sciences, São Paulo State
Results: Twelve studies were considered eligible for this review. Included studies
University (UNESP), Presidente Prudente,
Brazil mostly reported outcome data at short‐term (less or equal than three months) and in-
termediate‐term follow‐up (more than 3 months, but <12 months). AVGs were more
Correspondence
Diego G. D. Christofaro, Department
effective than no/minimal intervention in reducing BMI/zBMI at short‐term (SMD
of Physical Education, São Paulo State = −0.34; 95% CI: −0.62 to −0.05) and intermediate‐term follow‐up (SMD = −0.36;
University (UNESP), School of Technology 95% CI: −0.01 to −0.71). In addition, AVGs were more effective in reducing body
and Sciences, Presidente Prudente, Brazil.
Email: diegochristofaro@yahoo.com.br weight compared with no/minimal intervention at intermediate‐term follow‐up
(SMD = −0.25; 95% CI: −0.46 to −0.04). Regarding physical activity levels, AVGs
Funding information
were not more effective compared with minimal intervention at short‐term and inter-
Fundação de Amparo à Pesquisa do Estado
de São Paulo, Grant/Award Number: mediate‐term follow‐up.
2016/03826‐5 Conclusions: Our review identified that AVGs were better than minimal interven-
tion in reducing BMI and body weight, but not for increasing physical activity in
young people.
KEYWORDS
active video game, adolescents, children, obesity, physical activity
|
4 ©
wileyonlinelibrary.com/journal/sms
2019 John Wiley & Sons A/S. Scand J Med Sci Sports. 2020;30:4–12.
Published by John Wiley & Sons Ltd
OLIVEIRA et al.
|
5
associated with reduced obesity‐related outcomes,9-11 the recommendations,24 and it was prospectively registered at
vast majority of children do not meet the recommended levels PROSPERO (CRD42017058483).
of physical activity.12 A comprehensive literature search was performed from
In addition to physical activity, other interventions have their inception to October/2018 in the following databases:
been proposed to reduce weight in children including behav- MEDLINE, EMBASE, CINAHL, SPORT Discus, and
ioral, dietary, and drug therapy approaches. However, current PEDro. The search strategy was conducted using a combi-
evidence reveals that none of these interventions are effective nation of search terms for children (“youth,” “childhood”),
for reducing weight in children.13,14 A systematic review15 AVGs (“exergame,” “active videogame,” “game‐based”) and
demonstrated that physical activity interventions have only randomized controlled trials. In addition, we searched in
a small effect on improving physical activity levels of chil- the main trial register databases (ie, ClinicalTrials.gov, the
dren which may explain, in part, why these interventions International Standard Randomized Controlled Trial Number
have limited success in reducing weight in this population. A Register, and the Australian New Zealand Clinical Trials
new generation of video games, known as active video games Registry) to identify ongoing and unpublished trials. The
(AVGs) or “exergames,” has emerged as an innovative inter- search was also aided by the World Health Organization—
vention to reduce childhood obesity.16 International Clinical Trials Registry Platform search por-
Several types of AVG are available, such as Nintendo tal. The searches were restricted to publications in English,
(Redmond, WA) Wii fit games, “Dance Dance Revolution” Spanish, and Portuguese. Appendix 1 contains the search
(Konami Digital Entertainment, El Segundo, CA), and strategy performed in MEDLINE. Citation tracking was
Kinect sports, and they might be considered a possible option performed to detect relevant studies missed by our search
given that 83% of young people have access to at least one strategy.
video game console.17 AVGs explore and integrate exercise
and gaming entertainment using a digital interface to pro-
2.1 | Study selection
mote physical activity and discourage sedentary behavior18
which may be delivered at home, in schools, and in laborato- Two reviewers independently screened the titles and ab-
ries. Recently, a growing number of studies have been con- stracts retrieved from the search strategy. Records selected
ducted to estimate the effectiveness of AVGs in children and as potentially eligible studies for the review were assessed
adolescents. through the full texts by two independent reviewers. Any
Previous systematic reviews19,20 investigated the effective- case of disagreement was resolved in consensus with a third
ness of AVGs on physical activity levels and obesity‐related reviewer.
outcomes. The first review19 conducted in 2014 investigat- Randomised Controlled Trials and quasi‐RCTs investi-
ing the effectiveness of AVGs on obesity‐related outcomes gating the efficacy of AVGs compared with minimal inter-
was unable to perform a meta‐analysis. The second review20 vention (eg, usual care, no intervention, education, advice,
conducted in 2015 reported a meta‐analysis showing favor- inactive videogames, and waiting list) in children aged
able results for the AVG group over sedentary behaviors (eg, from 2 to 19 years were included in this review. Quasi‐
sedentary video games, TV viewing) among young people.20 RCTs were considered eligible in this review because, in
However, this review also included non‐randomized con- this study design, participants are allocated using a method
trolled trials in the meta‐analyses which may overestimate which may not be considered strictly random (eg, date of
the results.21,22 Furthermore, AVGs delivered out of school birth).25 We adopted this approach in accordance with a
environment are more feasible due to the nature of the inter- previous review instead of including any non‐random-
vention and may increase the practice of physical activity in ized trial.26 AVGs were defined as any computer game
order to reduce weight and promote health improvement.23 using interaction between body movements and specific
Therefore, the aim of this systematic review of randomized feedback and goals. Therefore, interventions which in-
controlled trials was to investigate the effectiveness of AVGs cluded consoles, such as the Nintendo Wii, Playstation
performed out of school (ie, home‐based and laboratory‐ (Sony Computer Entertainment Inc), and Kinect TM for
based) compared with minimal intervention on obesity‐re- Microsoft's Xbox 360TM (Microsoft Inc), and game con-
lated outcomes and physical activity levels in children and trollers, such as GameCycle (Three Rivers Holdings) and
adolescents. Sony EyeToyVR (Sony Computer Entertainment Inc), were
considered eligible. Interventions using AVGs isolated or
in combination with other approaches were required to be
2 | M ET H OD S delivered at home or in the laboratory. Therefore, school‐
based interventions were excluded from this review.
This review followed the Preferred Reported Items Weight‐related outcomes, such as BMI and BMI z‐
for Systematic review and Meta‐analyses (PRISMA) score, body weight, body fat, and waist circumference, were
|
6 OLIVEIRA et al.
Regarding the risk of bias, one study (7.7%) had a PEDro body weight, there was moderate‐quality evidence
score of 7 out of 10, four studies (30.7%) scored 6 out of (downgraded for risk of bias) that AVGs were not more
10, and six studies scored 5 out of 10 (46.1%) (Appendix effective than no/minimal intervention at short‐term
4). Two studies (15.4%) blinded the assessors, five (38.5%) follow‐up (5 trials, SMD = −0.12; 95% CI: −0.30 to
performed intention‐to‐treat analysis, and ten (76.9%) studies 0.06) (Appendix 7). Nevertheless, there was high‐qual-
presented adequate follow‐up (ie, <15% of loss to follow‐up). ity evidence that AVGs were more effective in reducing
After performing the search in the base of records of ran- body weight compared with no/minimal intervention at
domized clinical trials, three ongoing studies were found. intermediate‐term follow‐up (2 trials, SMD = −0.25;
The estimated sample sizes of the studies ranged from 60 95% CI: −0.46 to −0.04) (Figure 3). For body fat and
to 116 children and adolescents with two trials investigating waist circumference (Appendix 8), there was high‐
exercise and one sports games. Two trials planned reporting quality evidence demonstrating non‐significant ef-
obesity‐related outcomes, and two trials report physical ac- fects for AVGs compared with minimal intervention at
tivity levels. Appendix 5 describes the characteristics of on- short‐term follow‐up (3 trials, SMD = −0.31; 95% CI:
going trials. −0.72 to 0.10/ 2 trials, SMD = −0.64; 95% CI −1.46
to 0.17).
3.1 | Weight‐related outcomes
3.2 | Physical activity levels
Appendix 6 summarizes the overall quality evidence
using the GRADE approach for primary and secondary Figure 3 shows pooled effects for physical activity levels
outcomes. A total of 11 trials reported weight‐related using the less conservative analysis. There was moder-
outcomes. For BMI/zBMI data, there was high‐quality ate‐quality evidence (downgraded for risk of bias) that the
evidence that AVGs were more effective than no/mini- AVGs were not more effective than the control group for
mal intervention at short‐term (6 trials, SMD = −0.34; increasing physical activity levels at the short‐term fol-
95% CI: −0.62 to −0.05) and intermediate‐term fol- low‐up (6 trials, SMD = 0.06; 95% CI: −0.19 to 0.31).
low‐up (2 trials, SMD = −0.36; 95% CI: −0.71 to There was moderate‐quality evidence that AVGs were
−0.01) (Figure 2). In fact, the effect sizes were small not more effective compared with minimal intervention
at short‐term and intermediate‐term follow‐ups. For at intermediate‐term follow‐up (4 trials, SMD = 0.22;
|
8 OLIVEIRA et al.
F I G U R E 2 Effectiveness of
active video game compared to minimal
intervention on BMI and zBMI in children
and adolescents. Abbreviations: AVGs,
active video game; CI, confidence interval;
SMD, standardized mean difference
F I G U R E 3 Effectiveness of
active video game compared to minimal
intervention on physical activity levels (ie,
less conservative analysis) in children and
adolescents. Abbreviations: AVGs, active
video game; CI, confidence interval; SMD,
standardized mean difference
95% CI: −0.09 to 0.52) (Figure 3). Considering the most intermediate‐term follow‐up. Of note, our search identified
conservative analysis (Appendix 9), there was a signifi- that at least three studies have already been recorded in the
cant difference favoring the minimal intervention group clinical trials database which could change the results for
at short‐term follow‐up (6 trials, SMD = −0.35; 95% the outcomes with moderate‐quality evidence, according to
CI: −0.64 to −0.05) and non‐significant differences be- GRADE.
tween groups at intermediate‐term follow‐up (4 trials, Considering the objectives of this review, three other re-
SMD = −0.04; 95% CI: −0.19 to 0.11). views were detected with a similar objective.19,20,45 Our find-
ings conflict with Bochner, Sorensen 45 which detected no
significant differences favoring the AVGs in a meta‐analysis
4 | D IS C U S S ION including seven studies. One of the possible differences be-
tween the reviews might be attributable to the data analysis.
The main findings of this review are that we found moder- Bochner, Sorensen45 used a combination of weight‐related
ate‐quality evidence that AVGs are more effective in reduc- outcomes to compare the intervention group and control
ing BMI and weight compared with no/minimal intervention. group, while we analyzed the outcomes separately in the cur-
Considering the physical activity levels, AVGS were not bet- rent review. Another factor is that two new RCTs were added
ter compared with minimal intervention at short‐term and in the current review.42,46 Meanwhile, the findings of the
OLIVEIRA et al.
9
|
current review align to those found by Gao, Chen 20 which included studies in this review; however, we could not exclude
also observed a small effect size on body composition in the possibility of missing studies. It is noteworthy that another
young people favoring the AVGs. aspect to be highlighted is the three unpublished articles found
Our review showed that AVGs may reduce BMI and body in the clinical trials database, which might increase the overall
weight of children and adolescents. Similarly, a previous re- quality of evidence in the near future. Another strength is that
view demonstrated that exercise can effectively reduce BMI we also investigated the effectiveness of AVGs on promoting
in obese and overweight children.47 Indeed, this review also additional benefits in addition to improving weight‐related
explored whether changes in exercise and/or physical activity outcomes including to increase physical activity levels of
levels beyond the exercise intervention influenced between‐ young people. One of the limitations of this review was the
group changes in BMI. However, they did not detect any sig- heterogeneity across individual studies of weight‐related out-
nificant association between these variables. One explanation come measurements and physical activity levels. To reduce the
is that the substitution of a sedentary activity (ie, inactive possible heterogeneity in the comparisons, we conducted the
videogame) to the practice of AVGs may stimulate children meta‐analyses were performed grouping the measures, for ex-
and adolescents to adopt a healthier lifestyle regardless of the ample, weight‐related outcome was stratified by the variables
increase in physical activity levels. For example, previous weight, BMI, body fat, and waist circumference. Furthermore,
studies have shown that an increased time spent in sedentary the included studies reported the outcome measures only at
activities is associated with food intake and poor eating hab- short‐term and intermediate‐term follow‐up. Thus, further
its.48,49 Therefore, we could argue that the practice of AVGs studies should be conducted to assess the long‐term effective-
may generate additional changes toward a healthy lifestyle ness of AVGs in young individuals.
in this population (eg, healthy eating habits) which, conse- In summary, there was high‐quality evidence to support
quently, would reduce BMI and body weight. Nevertheless, the use of AVGs among young people to reduce BMI and
future studies should test the effects of including AVGs as body weight, but there is limited evidence to support the ra-
part of a multidisciplinary intervention in order to improve tionale that AVGs can increase physical activity levels and re-
weight‐related outcomes in children and adolescents. duce body fat and waist circumference. Nevertheless, AVGs
In contrast, AVGs were not effective in reducing body should be performed in combination with sports activities,
fat and waist circumference of children and adolescents. but especially when adolescents are spending time on screen
Considering there was a significant reduction in body mass activities. Indeed, the practice of this type of intervention
index and body weight, it might have been a reduction in lean could be further implemented to promote a behavior change
mass of children and adolescents after the AVGs. However, toward a healthy lifestyle, since a large proportion of these
the pooled estimates of body fat and waist circumference young people spend a high number of hours per day playing
showed very low quality of evidence which indicates high inactive video game.53
uncertainty in these estimates, and therefore, further studies This review demonstrated that AVGs can contribute to re-
are required to reach a definitive conclusion. duce BMI and body weight, but should not replace traditional
Considering the effects of AVG use on physical activity, sports because they did not increase the physical activity levels
this meta‐analysis found no significant differences in increas- of children and adolescents. However, some questions remain
ing the physical activity levels in the AVGs group compared to be answered by future studies: (a) if the AVG is performed
with the control group. This finding aligns to the results of together with traditional sports activities, could it cause bet-
the Gao et al which also detected no significant differences ter effects on weight‐related outcomes and levels of physical
in promoting the physical activity levels favoring the AVGs. activity intensity of young people; (b) does the type of game
In fact, these findings show that despite AVGs contributing adopted in these interventions (usually dance and sports)
to a reduction in BMI and body weight they do not replace have the same effects on the health of these young people; (c)
the physical activity of traditional sports.50-52 Considering long‐term interventions (minimum 12 months) could provide
that most studies (7 out of 8) reported objectively measured changes in the lifestyle of young people; and (d) what types of
physical activity, one explanation for these findings is that games could be played in the competitions or cooperatively.
the varied amount of movements performed in the interven-
tion could not be entirely captured by the objective methods
due to the limited sensibility in specific activities (eg, arms 5 | CONCLUSION
movements). Future studies investigating the effects of AVGs
on physical activity levels should be conducted using the ob- The findings of the current review demonstrated that AVGs
jective measures with different choices of data collection (eg, were more effective in reducing BMI and body weight com-
multiple accelerometers) to clarify this issue. pared with minimal intervention of young people at short‐
The strength of this review may be considered the litera- term and intermediate follow‐up. In contrast, AVGs were not
ture searches conducted in five databases which identified the better than minimal intervention for increasing the physical
|
10 OLIVEIRA et al.