Mobile- and Web-Based Interventions for Promoting Healthy Diets, Preventing Obesity, and Improving Health Behaviors in Children and Adolescents: Systematic Review of Randomized Controlled Trials

Introduction

Background

The World Health Organization recognizes overweight and obesity as a global epidemic affecting both low- and middle-income countries and higher-income countries [1,2]. They pose a substantial threat and challenge to public health, as they are also linked to the efficiency of the health care system [1,3]. In particular, obesity and overweight are associated with health effects and diseases in childhood and adolescence. The literature indicates that such risks include metabolic syndrome, type 2 diabetes, and hypertension [1,4]. Global research data are alarming, as they show that as many as 39 million children younger than 5 years were living with overweight or obesity in 2020, and more than 340 million children and adolescents aged 5 to 19 years were diagnosed with overweight or obesity in 2016 [5].

In addition, children consume fruits and vegetables at levels well below the recommended values. Furthermore, more than 80% of adolescents worldwide do not engage in sufficient physical activity [6,7]. A lack of a well-balanced diet combined with little or no physical activity may contribute to the development of noncommunicable diseases such as cancer, diabetes, and cardiovascular disease [8,9].

Childhood obesity can also have severe psychological effects, as studies have shown that childhood obesity affects self-esteem and is associated with depression, lower quality-of-life outcomes, and emotional and behavioral difficulties [10]. In addition, diminished life satisfaction—an important dimension of subjective well-being—is associated with overweight and obesity [11] and low subjective health perceptions [12]. Obesity also contributes to poor functioning in the peer environment in children, as children with obesity report stigmatization and bullying by peers [10].

It is necessary to develop a coherent strategy to combat the obesity epidemic among children and adolescents, as obesity poses a threat to the health and well-being of individuals and has an impact on resources and economic costs [10]. Health education can foster and reinforce healthy eating habits and attitudes through various interventions aimed at children and adolescents.

In children and adolescents, specially developed games and apps may help develop and promote healthy lifestyles [8]. Gamification uses the process of teaching or learning with new technologies and apps, which are very attractive to this age group. The games are designed to promote motivation for various achievements and awards but also include elements of interaction between the participants [13,14]. Elements of gamification have been used to improve the uptake of healthy habits (such as increased physical activity or improved nutrition) [15,16], to reinforce changes in habits and behaviors [17,18], and to enhance positive emotions associated with subjective well-being [17,19]. There is some evidence that after game-based interventions among children and adolescents, fruit and vegetable consumption increase, but the improvement in nutrition-related knowledge [20-23] is inconsistent.

Objectives

The aim of this paper was to present conclusions from a systematic review (SR) of scientific literature on the effectiveness of mobile- and web-based interventions in promoting healthy diets, preventing obesity, promoting physical activity, and improving attitudes and knowledge toward nutrition in children and adolescents compared with traditional interventions and a lack of any intervention. The findings might constitute the basis for planning methodologically sound subsequent mobile- and web-based interventions for promoting health in children and adolescents.

Unlike previous SRs that focused on either dietary or physical activity outcomes, this review provided a comprehensive synthesis of digital interventions targeting multiple health behaviors in children and adolescents. Given the increasing integration of digital tools that address multiple lifestyle behaviors simultaneously, our broad inclusion criteria allowed us to capture real-world applications of these interventions. This approach provided valuable insights for designing future interventions that consider the combined impact of diet and physical activity on health outcomes.

Methods

Protocol and Registration

The review was registered with PROSPERO, with the registration number CRD42023423512. PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) 2020 statement was used for reporting this review [24]. There were no deviations from the registered protocol. A meta-analysis was not planned from the outset due to the anticipated heterogeneity of interventions.

Information Sources

A systematic literature search was conducted across 4 databases, including Web of Science, PubMed, Scopus, and Google Scholar, to identify relevant randomized controlled trials (RCTs) on mobile- and web-based interventions for promoting healthy diets, preventing obesity, and improving health behaviors in children and adolescents. The search encompassed records available up to September 30, 2024.

Search Strategy

The search strategy was developed collaboratively by the research team, with the guidance of an experienced librarian. The librarian provided expertise in constructing comprehensive and precise search strings. Keywords were identified based on an iterative review of the existing literature, discussions among team members, and previous SRs in related fields. The selection of terms reflected the study objectives, focusing on key elements such as population (eg, children and adolescents), interventions (eg, games and digital tools), and outcomes (eg, dietary habits and physical activity). Search terms were tested and refined to ensure they captured a broad yet relevant evidence base, with final search strings tailored for each database (PubMed, Web of Science, Scopus, and Google Scholar). The search words focused on the target population, intervention types, and desired outcomes, using Boolean operators for precise search queries. The search queries per database are presented in Multimedia Appendix 1, and the results from each database query are provided in Multimedia Appendix 2.

SRs identified during the database search were excluded from the primary analysis as they did not meet the inclusion criteria related to study type. However, to ensure comprehensiveness, all SRs identified were reviewed for their relevance to include original studies that aligned with our scope. Reference lists of SRs that were relevant in terms of our research questions were reviewed for relevant original articles. Duplicate records were identified and removed using the web-based SR software Covidence (Veritas Health Innovation Ltd). The process involved automatic detection of duplicates followed by manual verification to ensure accuracy. Any discrepancies in duplicate identification were resolved through discussion among the reviewers.

Eligibility Criteria

Articles were included when the outcomes of interest concerned dietary intake, diet quality, anthropometric measurements (such as BMI or weight), physical activity, or attitudes and knowledge related to diet or food. We included studies with a mean age of the participants from 6 months to 18 years and with participants without a preexisting severe chronic disease. Nonrandomized studies and case reports were excluded. Articles in languages other than English, Spanish, Finnish, or Polish were excluded because of limited resources for translation. However, no studies in other languages appeared in the search.

Screening

The first screening was carried out using the population, intervention, comparison, outcomes, study design, and time (PICOST) framework [25]. To ensure a systematic and comprehensive approach, the PICOST framework was used to inform both the development of the search query and the inclusion or exclusion criteria. The population focused on children and adolescents (eg, children, adolescents, kids, and family), targeting studies that investigated participants with a mean age ≤18 years. The intervention criteria included digital and mobile-based tools promoting health, such as game-based interventions (games for health, mobile phone, digital tools, and gamification). The comparison element was not explicitly restricted, allowing studies comparing digital interventions with traditional methods, no intervention, or other digital strategies. The outcomes criteria were centered on dietary behaviors, nutrition, and health-related habits, as reflected in the search terms dietary habits, food habits, dietary choices, healthy eating, eating behavior, and healthy lifestyle. These outcomes were aligned with public health objectives aiming to track improvements in eating behaviors and nutritional intake. The study design criteria focused on RCTs evaluating the effectiveness of these interventions. Finally, while the time criteria were flexible, preference was given to studies that reported both short-term and long-term outcomes. This structured approach guided our search, leading to the use of Boolean operators and specific keywords to capture relevant studies in the databases, as reflected in the search queries. A detailed list of inclusion and exclusion criteria used during the screening process is provided in Multimedia Appendix 3.

The titles and abstracts of the studies were independently screened by 5 authors (CT, NdQ, FAA, NZ, and PR), with each author reviewing a subset of the results. To assess interrater reliability, 20% (65/324) of the studies were screened in duplicate by 2 independent reviewers, and the κ coefficient was calculated to evaluate the consistency of the screening process. The agreement between the duplicate selections was found to be high (κ=0.86), and therefore, the rest of the selections were performed by 1 author. The conflicts in the 20% duplicate selection were resolved through discussion by the researchers concerned.

Following the title and abstract screening, full-text articles of studies that met the initial inclusion criteria were retrieved and assessed for eligibility. The full-text eligibility assessment was conducted by single reviewers for most articles, with 20% (18/89 full texts) assessed in duplicate to ensure consistency (κ=0.71 for interrater agreement). Discrepancies in duplicate assessments were resolved through discussion among reviewers.

Data Extraction

The data were extracted using the web-based SR software Covidence. To extract the data from the primary articles, we used a data extraction form (Multimedia Appendix 4) adapted from the one used by the Cochrane Collaboration [26]. Five authors (ESC, FAA, NZ, CT, and NdQ) extracted the data. After the first round of data extraction for approximately 19% (7/37) of the records, the data were independently reviewed by another author (ESC, FAA, NZ, CT, and NdQ). At this level of screening, the interrater reliability was κ=0.71. Disagreements were resolved through discussion, and if consensus could not be reached, a third reviewer was consulted.

The extracted data encompassed a comprehensive set of information, including but not limited to general characteristics (authors, publication year, and study country), study population or sample details, aims, intervention characteristics, study duration, and various types of outcome measures. The results were synthesized narratively because the interventions had too many differences compared with each other. The results were grouped for reporting according to the outcome. All types of effect measures (risk ratio, odds ratio, means, among others) were extracted.

Effectiveness of Interventions

The effectiveness of the interventions was evaluated based on the difference between the intervention and control groups as a result of the intervention in the outcomes of interest (Eligibility Criteria section).

Risk-of-Bias Assessment

The risk of bias was assessed using the Cochrane Collaboration Risk-of-Bias assessment tool [26], which evaluates 6 domains: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, and selective reporting. Each domain was rated as low, high, or unclear risk. The assessment was conducted in duplicate for 20% of the included studies to ensure reliability (κ=0.71). Disagreements were resolved through discussion and, if necessary, consultation with a third reviewer.

Results

Included Studies

The database search resulted in 325 records, and 228 after removing duplicates (Figure 1).

‎

Of the 36 SRs identified, 4 were considered relevant to our research questions, the rest were excluded. From the reference list of the 4 relevant SRs, 81 articles were identified as relevant.

After removing duplicates, 72 were screened [5,8,27,28]. Altogether 300 articles were screened from which 88 were rigorously assessed for eligibility. In total, 37 articles based on 34 intervention studies were found to be eligible. Of the eligible RCTs, 3 reported their results in 2 publications [29-34]. The number of excluded studies and reasons for exclusion are listed in Figure 1.

Characteristics of the Interventions

The included studies, published between 2003 and 2024, evaluated digital interventions targeting dietary behaviors, physical activity, and other lifestyle-related behaviors in children and adolescents. The interventions varied in design, duration, and behavioral focus, with some addressing multiple lifestyle behaviors simultaneously The characteristics of the included studies and a detailed description of the interventions are summarized in Table 1 and Table 2, respectively.

Of 34 studies, 5 (15%) focused on young children aged 2 to 5 years [31,33,35,51,57,60], while 20 (59%) targeted primary school-aged children between 6 and 12 years [12,22,30,32,34-39,41-44,46,48-50,53,55,62,65].

Of 34 studies, 2 (6%) focused on children aged 9 to 14 [43,58] and 7 (21%) were conducted among adolescents aged 13 to 19 years [32,39,40,42,43,56,59,63,64,66,67]. In 2 (6%) studies, young adults aged 16 to 21 years were included [52,64]. Family-based interventions, where both children and parents participated, were included in 2 (6%) studies [41,66].

Of 34 studies, 20 (59%) implemented school-based interventions, while home-based interventions were used in 8 (24%) studies [31,33,41,50-52,54,57,61]. In total, 2 (6%) studies used mixed settings, integrating both school and home components [48,55]. The remaining 4 (12%) studies were conducted in a food pantry [41], a study clinic [56], or an after-school program [58].

Of 34 studies, 9 (26%) exclusively targeted dietary intake, such as fruit and vegetable intake or healthy food choices [32,34,35,37,38,43,54,56,60,61,65]. In total, 8 (24%) studies exclusively targeted nutritional knowledge [22,36,42,44,48,49,62,63]. In total, 2 (6%) studies assessed only physical activity levels [41,59]. A total of 4 (12%) studies focused on both dietary intake and physical activity [39,45,46,66]. In total, 3 (1%) studies focused on dietary intake and knowledge [52,53,55], with 1 (3%) of these also including physical activity [55]. In addition, 5 (15%) studies incorporated other outcomes, such as attitudes or intrinsic motivation [40,47,50,57,64].

Anthropometric measurements were included in 9% (3/34) studies [31,33,51,58], while 3% (1/34) study included 4 outcome types: anthropometric measurements, dietary intake, physical activity, and attitude or intrinsic motivation [51].

Across the 34 studies, a primary outcome was clearly identified in 9 (26%; Multimedia Appendix 5) studies. These included outcomes were dietary intake of fruit, juice, and vegetables [37]; dietary intake of target vegetable preparations [43]; dietary intake of vegetables, fruit, or fruit and vegetables [32,34,65]; total energy intake [56]; knowledge of nutrition [48,52]; BMI [51]; and fat mass index [31,33]. Altogether, in 65% (22/34) of studies, it was unclear what the primary outcome was. In 53% (18/34) of those, there was no sample size calculation [22,35,36,40-42,44-47,50,53,57-60,62,63], and in 12% (4/34) studies, it was not clear for what outcome the sample size was calculated [39,49,54,66]. In total, 9% (3/34) of studies were referred to as pilot studies without a power calculation [22,41,58]. The status of the outcomes (primary, secondary, and unclear) is marked in Multimedia Appendix 5 and as relevant in Table 3.

Among the included RCTs, 35% (12/34; reported in 14 articles) were from the United States [32,34,35,37-40,43,57-60,63,65]. The European studies were from Spain [36], the United Kingdom [41], Italy [62], Sweden [31,33], Norway [61], the Netherlands [47,53,64], Germany [45,55,66], Finland [52], and Switzerland [50]. Studies from other continents included Brazil [42], Australia [44,51], New Zealand [56], India [54], the Philippines [49], Canada [48], and Mexico [22].

Effectiveness of the Interventions

Multimedia Appendix 5 shows the effects of all the interventions on all the various outcomes, and Table 3 summarizes the main findings.

Dietary Intake

Most positive results were observed for fruit intake: 5 of 11 (45%) studies succeeded in increasing intake, from which the effect was ascertained with an objective measure (skin carotenoid concentration) [65]. In total, 4 of 11 (36%) studies targeting the intake of fruit and vegetables have succeeded only in increasing fruit intake [32,34,37-39,65], which could indicate greater acceptability of fruits than vegetables among children. Nevertheless, 9% (1/11) of studies succeeded in increasing vegetable intake, but not fruit intake [61].

The most common behavioral change techniques (BCTs) were goal setting, instructions on how to perform a behavior, self-monitoring of behavior, and action planning (Multimedia Appendix 6). Only 33% (2/6) studies using feedback on behavior were successful [4,45]. However, all of them used a different combination of BCTs, in which there was only one feedback, and no clear pattern of BCTs could be seen in the successful versus unsuccessful intervention.

The 2 (50%) successful interventions out of a total of 4 interventions with the aim of decreasing sugar-sweetened beverage (SSB) use were nongame mobile apps [33,57]. They both included a mobile app for use by the parents of 4-year-old children [33] or preschool-aged children [57]. The main difference compared with the other interventions reporting SSB intake [6,32] was that they targeted much younger children (4 years and preschool age vs 9-11 years) and the app was used by the parents instead of the children. The primary aims of the interventions also differed—the studies by Nyström et al [33] and Nollen et al [58] were obesity prevention interventions, and Thompson et al [32] primarily targeted increased fruit and vegetable consumption, while the study by Nezami et al [57] targeted decreased SSB intake. A computer game intervention aiming to increase fruit, vegetables, and whole-grain intake and to decrease added sugar and fat intake succeeded in decreasing sugar intake but not in changing the other intakes. Given the small number of different types of studies, it is difficult to make generalized conclusions about what types of interventions would be successful in decreasing sugar intake in children.

Out of 34 studies, 4 (12%) studies have reported changes in the consumption of snacks, candy, or chocolate [31-33,46], of which 1 (25%) was successful [32], 1 (25%) had negative effects [46], and the rest had no effect. The successful intervention applied goal setting and planning of how to meet the goal, in addition to a game as the intervention. The one with negative effects used an advergame (a free web-based game integrating advertising messages, logos, and trade characteristics) as an intervention. Interestingly, cues of healthy and unhealthy foods in the advergame resulted in increased intake of unhealthy foods.

The persistence of increased fruit intake and carotenoid concentrations 3 months after the end of the FIT game intervention [65] and 3 months after the end of the game and action plan intervention in the studies by Thompson et al [32,34], and reduced discretionary food intake 6 months after the end of the eHealth intervention Time2bHealthy [51], contradicts the general pattern reported by others [61]. The lack of sustained results in the study by Nyström et al [31] aligns with this conclusion. The longer follow-up period (12 months) in the study by Nyström et al [31] could explain the difference in the nonpersistence of changed food intake compared with that in Thompson et al [32], Wengreen et al [51], and Hammersley et al [65].

Anthropometric Measurements

Regarding anthropometric measurements, of 34 studies, 1 (3%) study revealed an effect on fat-free mass [31,33]. The lack of effect in the other studies may be attributed to the fact that, on average, the study participants were not overweight [6,33] or the intervention was too short to have measurable results (12 weeks in the study by Nollen et al [58], <2 months in the study by Baranowski et al [39]). In addition, only the studies by Nyström et al [33] and Hammersley et al [51] reported that a sample size calculation was based on the chosen anthropometric outcome, whereas the study by Nollen et al [58] did not perform a power calculation, and the study by Baranowski et al [39] did not mention for what variable the sample size calculation was performed.

Physical Activity

Overall, the articles that evaluated any aspect associated with physical activity indicated no effect of the intervention [6,33,45,63,66]; of 34 studies, 1 (3%) reported that the sample size was too small [41], and the other had an attrition rate too large for statistical analysis [59].

Knowledge and Attitudes

In total, 2 (50%) out of 4 game-based studies reported beneficial effects on attitudes toward healthy eating or nutrition, and physical activity [40,52]. Both studies were rather small, which limits the interpretability of the results. Most of the studies reporting on knowledge about nutrition, food, or physical activity were successful. Interestingly, studies with no impact targeted older individuals—13 to 19 years [42] and 18 years [57]—compared with successful studies (7-13 years). In the study by Chagas et al [42], the control group did not receive any intervention. In the study by Heikkilä et al [52], the control group also participated in nutrition education sessions but without a mobile app intervention. Both the intervention and control groups increased their nutrition knowledge, but there was no difference between the groups, suggesting no additional benefit from the mobile app.

Risk-of-Bias Assessment

Table 4 shows the risk-of-bias assessment for each study. The risk of bias due to random sequence generation was considered low for all 34 included studies (37 publications) [22,31-44,46,48-51,53,55-60,62-67]. Concerns of either high or unclear risk of bias were raised in all other domains for some studies, ie, allocation concealment (17/34, 50% studies); blinding of participants and personnel (25/34, 74% studies); blinding of outcome assessment (28/34, 82% studies); incomplete outcome data (18/34, 53% studies); selective outcome reporting (3/34, 9% studies); and other biases (16/34, 47% studies). The interventions conducted in the studies by Nollen et al [58] and Thompson et al [34] were considered to have a low risk of bias in all domains, and those conducted by Folkvord et al [46] and Kato-Lin et al [54] were considered to have a low risk of bias in all but one domain.

Discussion

Dietary Intake

The interventions showed variable effects on dietary intake—many of the dietary outcomes related to food and energy intake remained unaltered by the interventions conducted. However, the reported outcomes were not always the main targets of the intervention, which probably explains some of the null results. Most positive results were observed for fruit intake, from which all but one used games as an intervention. The more frequent success in fruits compared with vegetables could indicate greater acceptability of fruits than vegetables among children. Other SRs have also shown that the most reported successful effect on dietary intake is fruit consumption [8,27,43,62]. Our findings also suggest that targeting parents instead of children may be more effective in reducing SSB consumption in children under school age.

This result is in accordance with previous studies targeting parents with educational content to reduce children’s SSB intake [68]. The negative implication of increased energy-dense food consumption after playing advergames, regardless of the type of advertised food (energy-dense foods or fruits) [46], calls for more research on the psychological mechanisms of games.

In our study, both maintained and not maintained long-term effects on fruit and vegetable intake were seen [31,32,34,65]. However, the general pattern reported by others seems to be a lack of long-term effects [61]. It may be that the use of digital tools needs to be recurrent to achieve long-term effects. In the future, it will be important to study the long-term effects of different digital tool interventions. Regarding other dietary behaviors, it is impossible to draw conclusions because of the small number of studies.

Anthropometric Measurements

It is difficult to draw firm conclusions about the potential of mobile apps or games to improve anthropometric measurements because of the shortcomings of the studies. However, the lack of effect found in this review is in accordance with other SRs in which no changes were found regarding anthropometric outcomes in children and adolescents [8,69]. Future studies should target overweight instead of normal weight children, be long enough to have an impact, and apply a sample size calculation.

Physical Activity

In this SR, studies involving physical activity interventions without dietary intervention were excluded, and our review focused only on combined dietary or physical activity interventions. This probably explains the differences compared with other SRs, which have also included physical activity interventions without dietary interventions [27,63,69]. Therefore, our results provide a limited view of the potential of digital tools for assessing children’s physical activity.

Knowledge and Attitudes

The results suggest that the interventions were successful in increasing young children’s knowledge about food and nutrition. Regarding attitudes, the small number of studies does not allow firm conclusions. In addition, the use of different types of control groups between interventions complicated comparisons of the effectiveness of the studies. Inconclusive results regarding adolescents’ knowledge, attitudes, and skills were also found in other reviews focused on the use of digital interventions [62].

Behavior Change Theories and Techniques

In several articles, researchers referred to psychological theories relevant to the creation of interventions aimed at improving health behaviors and attitudes in children. Most authors referred to well-known psychological concepts, such as social cognitive theory [31,37,39,42,61] and self-determination theory [52,65]. However, mentioning a theory provided little insight into what BCTs the theory led to. Describing the actual BCTs, as was done by many studies, was more informative. We compared the effectiveness of different BCTs by categorizing them based on the BCT taxonomy by Michie et al [70]. However, no clear pattern was seen. This could have been affected by (1) too different combinations of BCTs between the studies, (2) all interventions using the same core BCTs, making little difference between the interventions, or (3) us not capturing all BCTs in some interventions because of a lack of detailed description. While it is worth considering incorporating psychological concepts and well-founded theoretical frameworks into the design and development of interventions [47], a detailed enough description of the BCTs is necessary for appropriate comparison of successful and unsuccessful BCTs.

Some of the games were linked with changing real-life behaviors, for example, through setting goals for behaviors outside the game environment. Others included goal setting and behavior change only in the game environment. Although this could be meaningful regarding the success of the intervention, the frequency of these was not different between the successful and unsuccessful interventions.

Strengths and Limitations of Our SR

This review has several limitations that must be considered when interpreting the findings.

First, it is challenging to draw conclusions based on studies with substantial heterogeneity in intervention type, duration, and control group comparisons. For example, the appeal and engagement of the digital tools may have differed significantly across studies, but compliance data were not consistently reported.

Second, many studies lacked rigorous sample size justification: 18 (53%) out of 34 studies did not report a power calculation, and among those that did, several did not clarify the outcome for which the calculation was performed. This may have resulted in underpowered analyses and limited detection of statistically significant effects.

Third, we did not conduct hand-searching or include gray literature, which may have excluded potentially relevant but unindexed or unpublished studies. This could have reinforced the effect of publication bias in our synthesis.

Fourth, the use of single reviewers for most of the title or abstract and full-text screening stages is a methodological limitation. Although only 20% of the records were screened in duplicate, the interrater agreement (using Cohen κ) was “almost perfect” for title or abstract screening and “strong” for full-text eligibility assessment [71].

Fifth, we acknowledge that categorizing findings by outcome rather than by intervention type or study setting may limit insights into which characteristics of digital tools were most influential in promoting behavior change. Future reviews may benefit from structuring the synthesis accordingly. Although we mapped the BCTs used in the studies investigating the effect on fruit and vegetable intake, the variety in combinations of BCTs among the studies prevented us from drawing conclusions.

Sixth, we were unable to perform a Grading of Recommendations, Assessment, Development and Evaluations assessment due to limited resources. Although we assessed the risk of bias in detail, this precluded a formal evaluation of the certainty of the evidence, which would have further strengthened the interpretability of the findings.

Seventh, we did not conduct a meta-analysis due to the high degree of heterogeneity across interventions, outcomes, and measurement tools. As such, we are unable to comment on the pooled effect size or clinical significance of the interventions studied.

Eighth, additional subgroup and sensitivity analyses—such as stratification by age group or publication year—may have provided further insights. The inclusion of older studies, while justified by our broad inclusion criteria, may limit the relevance of some findings, given the rapid evolution of digital technologies. Despite these limitations, some of the strengths of this review are a thorough examination of the quality of the studies and the use of a preregistered protocol.

Recommendations

Given the findings from our SR, which highlighted the variable effectiveness of mobile- and web-based interventions in improving health behaviors among children and adolescents, the following recommendations are proposed to enhance the design and evaluation of future interventions:

• Explore the differences in the features of different tools and the role of the features in the effectiveness of the tools. It is crucial to extend the follow-up periods in future studies and compare different durations of the use of the same tool to assess the sustainability of behavioral changes. Research should aim to measure long-term outcomes beyond the immediate postintervention phase to better understand the lasting impact of these interventions.
• Explore the role of parental involvement and family dynamics in the success of interventions, given their significant influence on children’s health behaviors.
• Investigate the scalability and real-world applicability of successful interventions, including assessments of cost-effectiveness and barriers to implementation in diverse settings.
• Enhance the methodological quality of research by ensuring adequate sample sizes and blinding where feasible to strengthen the evidence base supporting these interventions.
• Enhance the theoretical underpinnings of interventions by applying behavior change theories more systematically. Future research should clearly outline how behavior change theories are integrated into intervention design and evaluation.

Conclusions

Our SR explored the impact of mobile- and web-based interventions on dietary habits, physical activity, and obesity-related health outcomes in children and adolescents. Despite the diverse array of interventions examined, the results illustrated a complex efficacy landscape. We found that while some interventions, particularly games, showed promise in improving short-term dietary behaviors, notably increasing fruit and vegetable intake, their impact on long-term health outcomes and changes in behavior, such as reducing sedentary time, was inconclusive.

The effectiveness of these interventions varied, with a notable number of studies showing improved nutrition knowledge and some reporting positive shifts in dietary behaviors. However, in the few long-term follow-ups, the changes were not always sustained over time, highlighting the challenges of maintaining health behavior changes initiated through digital interventions.

Given the mixed outcomes, it is important to further investigate the specific features of the tools that are effective. Future studies should focus on developing interventions that not only engage children and adolescents effectively but also include components that support sustained health behaviors.