Introduction
Childhood obesity has become one of the most pressing public health challenges of the 21st century, driven by poor dietary habits, including insufficient fruit and vegetable (F&V) consumption and high intake of sugar-sweetened beverages (SSBs). Eating habits, early-life experiences and sedentary lifestyle increase the risk of obesity persisting into adulthood (Sahoo et al., 2015; Oude Luttikhuis, 2009; Langford et al., 2015; Cassidy & Shaver, 1999).
Globally, rates have risen sharply, with nearly 30% of European children aged 7-9 now overweight or obese (WHO Europe, 2018). This trend is fuelled by poor nutrition, physical inactivity, and environmental factors, including marketing and socioeconomic conditions (Duffey et al., 2012; Mackenbach et al., 2008). Given that non-communicable diseases account for most global deaths, early, multi-level interventions are essential to promote lasting healthy habits (WHO, 2021).
Globally, childhood obesity has risen sharply, affecting both developed and developing nations (WHO, 2021; Kelishadi & Soleiman, 2014). For instance, in Europe, childhood obesity rates have tripled with nearly 30% of 7-9 year olds classified as overweight or obese (WHO Europe, 2018). This trend is shaped by poor nutrition, sedentary lifestyles, and environmental factors (Nollen et al., 2014; Sahoo et al., 2015), with socioeconomic status, cultural norms and marketing also playing critical roles (Mackenbach et al., 2008; EU Action Plan on Childhood Obesity 2014-2020).. Given that obesity is directly linked to chronic diseases including cardiovascular conditions, diabetes, and respiratory disorders (Bandini et al., 2005; WHO, 2021) and these account for 71% of global deaths, multi-faceted early interventions are vital to promote lifelong healthy habits (WHO, 2021).
The consequences of childhood obesity affects both physical and psychosocial wellbeing, increasing the risk of chronic diseases later in life. The urgency to promote healthier eating behaviours among children through early interventions and structured environments like schools, has never been greater. In this context, communication technologies and digital media offer promising tools to enhance the delivery and effectiveness of health interventions. From educational games and mobile applications to web-based learning platforms, digital tools can personalise information, provide immediate feedback, and increase motivation through interactivity and enabling scalable and cost-effective options.
This study explores on the potential of communication strategies and digital technologies in delivering school-based health interventions aimed at promoting healthy nutrition behaviours among children aged 8 to 12. Schools represent a strategic setting for fostering health behaviour change in children, as they offer a structured environment where interventions can be implemented consistently (Whittemore, Jeon & Grey, 2013; Fertman & Allenworth, 2010; Moore et al., 2009). This study investigates how different media formats (printed vs. digital games) influence outcomes related to knowledge, behaviour, self-efficacy, and nutrition self-care. By examining the interplay between media and behavioural outcomes, this research contributes to the growing field of interactive health communication and supports the development of evidence-based interventions for public health.
Literature Review
The role of communication in school-based health interventions
Health communication is a growing field at the intersection of health and communication sciences. The combination of these fields aims at better fostering, reinforcing and guiding behaviours, policies and practices that advance the health and well-being of people (Society for Health Communication, 2017). Health communication plays a key role in health promotion and interventions (Thomas, 2006; Hornik et al., 2008; Rimal & Lapinski, 2009) being paramount in shaping public health outcomes, particularly when targeting children’s dietary behaviours in school settings.
Schools offer an ideal environment for implementing structured, evidence-based interventions that improve health knowledge, influence behaviour, and support long-term well-being (Fertman & Allenworth, 2010; WHO, 2009). By engaging students, families, and communities, school-based programmes can foster sustained changes in eating habits and promote healthy lifestyles from an early age (WHO, 2009; Langford et al., 2014; Roseman et al., 2011). Studies highlight the effectiveness of school-based programs in increasing F&V consumption and reducing unhealthy dietary habits (Langford et al., 2014). Multi-component approaches that combine education, parental involvement, policy changes, and community support have shown the most success (Kubik et al., 2003; Whittemore, Jeon & Grey, 2013). According to WHO (2009), long-term success in interventions depends on consistent reinforcement and follow-up. Incorporating interactive methods, technology, and peer support further enhances engagement and impact (Moore et al., 2009; Yang et al., 2015).
Even though school-based nutrition interventions have expanded in recent decades (Whittemore, Jeon & Grey, 2013), traditional interventions often come with high costs and effort from educators. Digital interventions offer motivating and engaging cost-effective alternatives (Kroeze, Werkman & Brug, 2006; De Vries & Brug, 1999; O’Donnell, Greene & Blissmer, 2014; Yang et al., 2015).
Nowadays, children grow up in media-rich environments, engaging with new media on a daily basis, regularly using smartphones, tablets, and online platforms (Ofcom, 2019; EU Kids Online, 2018; Anderson & Jiang, 2018). From a socio-cultural angle, media are embedded into users’ daily routines and settings (Silverstone & Hirsch, 1992) which represents an important consideration in school-based interventions, where the classroom context may mediate how such tools are received and integrated. As a result, digital media offers a promising channel for delivering health messages and interventions that are engaging, scalable, and cost-effective (Kroeze, Werkman & Brug, 2006; De Vries & Brug, 1999; O’Donnell, Greene & Blissmer, 2014).
Digital media, including mobile apps, games, SMS-based systems, and web platforms, have shown positive effects on both diet and physical activity, often outperforming traditional approaches (Moore et al., 2009; Cullen et al., 2016; Nollen et al., 2014; Silva et al., 2015; Melo et al., 2017). These media formats enable real-time feedback, tailored content, and self-directed learning, contributing to outcomes such as improved nutritional literacy, increased F&V consumption, and reduced intake of SSB (Struempler et al., 2014; Baranowski et al., 2012; Yang et al., 2017). On the other hand, despite their potential, digital interventions require careful implementation to mitigate potential risks associated with excessive screen time, privacy, and unhealthy foods marketing exposure (Taylor et al., 2005; Livingstone & Bovill, 1999). Different communication media activate distinct cognitive and behavioural mechanisms, functioning not only as channels for delivering content but also as tools that shape learning and engagement. Media theory helps to further clarify these differences. Uses and Gratifications Theory (Katz, Blumler & Gurevitch, 1973) suggests children engage more with media that fulfil needs, such as enjoyment, curiosity, or mastery, making interactive formats such as digital games motivating. Complementarily, following a media affordance perspective (Hutchby, 2001; Treem & Leonardi, 2012) each format enables or constrains different possibilities. For example, posters afford visibility, reflective learning repeated consultation, while games afford interactivity, feedback, and trial-and-error learning. These affordances make posters more suitable for reflection and planning, relevant in early stages of behaviour change (BC), while digital games may better support engagement and action through immersion and fun. Understanding these media-specific dynamics helps tailor communication strategies to behavioural goals and strengthens the theoretical and practical relevance of intervention design.
Effective health communication is key to promoteBC, as coit shapes health-related decisions and actions. BC theories are essential to design, plan and implement successful interventions, contributing with an evidence-based insights into the psychological and structural factors influencing human behaviour. Research highlights that theory-based approaches enhance intervention planning and outcomes (Cane, O’Connor & Michie, 2012; Painter et al., 2008; Noar & Zimmerman, 2005).
Theoretical approaches to health behaviour change in school-based interventions
Health BC theories are essential for understanding the psychological and structural determinants that influence human action. These frameworks offer evidence-based foundations for designing, implementing, and evaluating interventions aimed at improving public health outcomes. Their intersection with health communication research provides valuable insights into how messages and media can be strategically used to support BC. Numerous studies confirm that theory-based interventions are more effective and consistent (Cane, O’Connor & Michie, 2012; Painter et al., 2008; Noar & Zimmerman, 2005).
The range of available BC theories is extensive—Michie et al. (2014) identified 83, while Davis et al. (2015) found 82—highlighting the growing relevance of this field. Despite some less convergent factors in these theories, a widely accepted principle in BC is that change is a dynamic process that occurs in stages. Individuals are placed along a behavioural path that reflects their likelihood of change based on specific constructs (such as motivation, intention, knowledge). Interventions aim to move individuals along the path, while research seeks to identify and isolate key predictive variables and how to impact them.
The Transtheoretical Model (TTM) and the Health Action Process Approach (HAPA) are two prominent stage-based models, arguing that change is not linear, and relapses may occur (DiClemente & Prochaska,1982; Velicer, Prochaska & Redding, 2006). These phases typically begin with a state of unawareness, followed by a phase of awareness and motivation, leading to the formation of intention. Intention then progresses to action, often involving initial attempts that may include relapses, before reaching a more stable behavioural pattern. Some models include a final maintenance stage, where the new behaviour is sustained without significant risk of reverting to the previous one (DiClemente & Prochaska, 1982; Velicer, Prochaska & Redding, 2006). The TTM, developed by Prochaska and DiClemente (1982), describes health BC as a spiral progression through five stages — pre-contemplation, contemplation, preparation, action, and maintenance — each requiring tailored intervention strategies (Prochaska, DiClemente & Norcross, 1992; Prochaska & Velicer, 1997). For example, in the pre-contemplation stage, individuals may be unaware of the risks associated with their behaviour, whereas in the preparation stage, they begin planning concrete steps toward change. HAPA, in turn, builds on limitations identified in earlier models, particularly the gap between intention and behaviour and divides BC into a motivation and a volition (action) phase. In the motivation phase, individuals assess risk, outcome expectancies, and their self-efficacy, culminating in the formation of an intention. The volitional phase focuses on translating intention into behaviour through action planning, coping strategies, and maintenance of self-efficacy in the face of challenges (Schwarzer, 1992; Schwarzer, 2008). This model addresses post-intentional processes often overlooked in other frameworks.
Both models have been applied in school-based interventions to promote healthy eating by aligning strategies with individuals’ readiness to change. The TTM was successfully applied to increase F&V intake and reduce SSB consumption among children (Struempler et al., 2014; Baranowski et al., 2012). These studies highlight the value of tailoring messages to specific behavioural stages, such as contemplation or preparation (Prochaska, DiClemente & Norcross, 1992; Prochaska & Velicer, 1997). HAPA has similarly been applied in youth-focused interventions to bridge the gap between intention and action. Constructs such as outcome expectancies, planning, and maintenance self-efficacy have proven useful in supporting long-term dietary change (Maes et al., 2011; Silva et al., 2015), namely in digital interventions that offer personalised feedback and self-regulatory tools (Yang et al., 2017; Melo et al., 2017).
Integrated theoretical framework
Considering the extensive list of BC models, a central and relevant current key challenge lies in integrating these perspectives to generate added value from their combined strengths, building from existing models, more than creating new ones. This can either be done by identifying similarities and complementary constructs, or conversely, by addressing divergences. The design and interpretation of this study was guided by an integrated perspective developed in earlier work by the authors (Author 1 & Author 2, Year) . This perspective draws on the Transtheoretical Model (TTM), the Health Action Process Approach (HAPA), and other core behaviour change theories, and aligns them with stages of change to support the planning and implementation of health interventions. It specifically highlights how intention formation (as addressed by HAPA) complements the motivational and volitional phases of TTM, facilitating a more comprehensive understanding of the behaviour change process. According to the authors (Author 1 & Author 2, Year), the integrated perspective is grounded in the principle of complementarity - using the strengths of one theory to fulfil limitations of others. Behaviour change theories are inherently interrelated, not only because they belong to the broader domain of behaviour change, but because they often build upon one another, incorporating and expanding on previous concepts. This indicates that potential for integration.
The integrated framework combines constructs from some well-established models, focusing mainly on the TTM and HAPA, applying the principle of complementarity - merging theories when constructs overlapped and focusing on dissimilarity to overcome limitations. The TTM served as the structural backbone, allowing constructs from other models to be mapped onto its stages, providing a comprehensive understanding of the BC process across precontemplation, contemplation, preparation, action, and maintenance. The integrated framework goes beyond simple aggregation of constructs, resulting from a comparative analysis that identified gaps in individual theories and strategically incorporated complementary elements to address limitations. It was developed to enhance explanatory power and practical utility for intervention design also incorporating a communication layer, and arguing how assertiveness, clear language, and positivity influence progression across stages (Author 1 & Author 2, Year).
Objectives
This study examines the role of communication and media technologies in school-based nutrition interventions. Its main objectives are to: 1) explore effective ways of communicating nutrition content to children; 2) identify BC indicators influenced by communication strategies; 3) compare the effectiveness of different media formats in promoting healthy eating habits; and 4) contribute to understanding the value of media technologies in school health promotion. The intervention aimed to increase F&V intake and reduce SSB consumption among children aged 8 to 12, in the school context.
Research questions
RQ1. Is a media-based intervention more effective than traditional teaching methods in improving knowledge, behavioural intentions, food consumption, and health-related behaviours among children?
The core hypothesis argues that all media-based intervention conditions (poster and game) will lead to measurable improvements in key behavioural outcomes, as assessed from baseline to endline, compared to a control group receiving no media input.
Additionally, the study compares the impact of different media formats:
RQ2. Do different communication media (static posters vs. interactive digital games) differentially influence behavioural change indicators?
It is hypothesised that technologies with visual and interactive features will be more effective in advancing BC, particularly in the post-intentional stages, by enhancing engagement, self-efficacy, and action-related outcomes. Therefore, the main hypothesis argues that the digital games group will report significantly greater improvements in post-intentional indicators (e.g. action, self-efficacy and actual food consumption).
Method
This study follows a cross-sectional quasi-experimental research design to to assess a school-based nutrition intervention aimed at promoting healthier eating habits among children. It explores whether media-enhanced interventions improve nutrition-related outcomes, such as knowledge, behavioural intentions, food choices, self-efficacy, risk awareness, and self-care, compared to traditional methods. It also explores whether digital games are more effective than static media in fostering these outcomes.
The intervention included three groups - a control group and two treatment groups. Group 1) was a control group used a traditional teaching-learning exposition, focused on oral explanation, brainstorming and the usage of the white board available in the classroom. Group 2) was a treatment group receiving the same intervention content supplemented with printed communication materials, such as flyers, leaflets and posters, that the students could take home or leave in the classroom, hang in the walls or keep in their notebooks. Group 3) was the second treatment group that engaged with digital learning games (serious games) alongside the core instruction. All groups received the same content, with message design adapted to maximise the communicative affordances of each media format. Random assignment to treatment and control group was completed at class level (cluster randomisation), meaning participants were not randomly assigned to each group individually, but within their class, due to convenience, organizational purposes and ethical reasons.
Participants
Students from 4th, 5th, and 6th grades, aged 8 to 12 years (n = 67, mean = 10, SD = 0.873) attending a private school located in [city], [country]. Some 25 students from the 6th grade were assigned to the control group, 20 students from the 4th grade were assigned to the print communication materials group, and lastly, 22 students from the 5th grade were assigned to the serious games’ treatment group. In total, 57% were male and 43% were female. Due to organisational constraints, group assignment followed a convenience sampling approach by class grade. Although this may introduce baseline demographic differences, the main analytic approach focused on within-group comparisons from baseline to endline.
The intervention
The intervention was integrated in students’ classes, in the different grades, Informed consent was obtained from parents. Sessions were co-facilitated by a trained researcher and the students’ regular classroom teacher, who was well acquainted with the children’s learning needs, behavioural patterns, and any specific challenges or disabilities. This ensured a supportive and familiar learning environment, tailored to the students’ developmental level. Eight sessions of 45 minutes were held with each group, during school time (total 360 hours per group). The intervention addressed the following main topics: 1) w the basics of healthy nutrition, including macro- and micronutrients, 2) food groups (food circle, food pyramid, eatwell plate), 3) balanced meals, 4) food labels, 5) choosing food in the supermarket, 6) what are real foods and what are artificial foods, 7) Malnutrition consequences. Emphasis was consistently placed on the importance of F&V consumption and the risks of sugar, particularly SSB. All students received a printed booklet containing the lesson content.
The control group received traditional educator-led lessons including oral explanation, whiteboard work, discussion, and slideshow presentations. Treatment group 1 received the same content, supplemented with the creation and use of printed communication materials (flyers, posters), which remained visible in the classroom or could be taken home. Treatment group 2 also received the core lessons, but incorporated digital serious games aligned with the curriculum topics. Game play occurred in class for approximately 15-20 minutes per session and included freely available online games focused on food groups, food plates, food label literacy, healthy meal preparation, supermarket choices, and distinguishing real vs. artificial foods (see figures for examples). Games were organised per topic, including the following: food groups, food circle/ food plate, prepare a healthy meal, food label literacy, choosing foods in the supermarket, real and artificial foods. For examples of the games played in class, see the following figures.
Digital access equity was assured during the intervention, as the school provided the necessary computers, and all students had equal opportunity to engage with the games. Moreover, as the school was private and students shared a similar socio-economic background, substantial differences in digital access at home were not anticipated.
This intervention followed national ethical guidelines for research with minors and the ethical principles outlined in the Declaration of Helsinki. Besides written informed consent from parents or legal guardians, verbal assent was obtained from the participating children, the school and the teachers involved. Particular attention was given to ensuring that all procedures were age-appropriate and implemented during regular school activities to minimise any disruption or discomfort. Moreover, the researcher was a trained developmental psychologist with many years of experience in research with children and vulnerable populations.
Intervention plan, outcome variables and measures
The intervention was primarily based on the TTM and HAPA models, which guided the selection of health indicators, intervention stages, and communication strategies. It assumes that BC occurs in stages, with different strategies fostering progression to the next phase. This intervention specifically targeted the contemplation and preparation stages, as well as the initial steps of the action stage.
The outcome variables assessed before and after the intervention included nutrition knowledge, food label literacy, actual behaviour (frequency of F&V) and SSB consumption), barriers to change, self-efficacy, and nutrition self-care.
•Nutrition knowledge was measured using an updated version of the General Nutrition Knowledge Questionnaire (Kliemann et al., 2016) and the Nutrition Knowledge Questionnaire (Turconi et al., 2003), both employing multiple-choice questions. Only sections 1, 2, and 3 of the General Nutrition Knowledge Questionnaire were used, aligning with the intervention’s focus.
Food label literacy was assessed using the latest version of the Food Label Literacy for Applied Nutrition Knowledge (FLLANK) questionnaire (Reynolds et al., 2012), adapted with real Portuguese food labels from supermarket products (five labels were used).
•Actual behaviour was evaluated through a nutrition frequency survey (Turconi et al., 2003) and a two-day 24-hour recall (one weekday and one weekend day). The frequency survey included six yes/no questions, such as “Do you eat fruits and vegetables daily?” with a follow-up on portion sizes. The 24-hour recall asked participants about fruit and vegetable consumption on the previous day and on the last Saturday.
•Barriers to change were measured using the Turconi et al. (2003) survey with nine Likert-scale items, including questions like “Do you know which foods should be limited to reduce sugar intake?” and “Do you know how to eat healthily?”.
•Self-efficacy was assessed through Turconi et al.’s (2003) self-efficacy scale, consisting of eight Likert-scale items, including “Do you think you can modify your diet if needed?” and “Do you think you can apply nutrition advice to improve your health?”.
•Nutrition self-care was measured using the Moore Index Nutrition Self-Care (MIN-SC) (Moore, 1995; Moore, Pawloski & Baghi, 2005). This 50-item scale assessed dietary habits, meal planning, adjustments, and nutrition skills on a frequency scale from “never” to “always/daily,” with items such as “I read about nutrition.”
All constructs measured in this study were previously validated and used with the target audience and its operationalisation was carefully aligned with the theoretical frameworks that guide the intervention. For instance, barriers to change and nutrition self-care were aligned with pre-action stages based on their theoretical links to motivation, perceived difficulty, and behavioural readiness, as conceptualised in the TTM and HAPA. These measures used where constructs were previously validated among adolescents and children in school settings (e.g. Bashatah & Alahmary, 2020; MIN-C, Santiago, Chile). To ensure age-appropriateness, items were linguistically simplified and supported with visual aids such as smiley faces and colour-coded Likert scales to aid comprehension. All materials were pilot tested with children in the target age group. While self-report among children presents challenges, prior studies have demonstrated its feasibility when tools are carefully adapted (Maes et al., 2011; Silva et al., 2015). These procedures aimed to ensure that constructs such as self-perceived capability and perceived barriers were developmentally appropriate and meaningful for the target group. Moreover, a researcher and a teacher were always present during the assessment phases, supporting children in reading and understanding the measures.
The following table outlines the stages, along with the indicators measured and promoted at each phase and the measures used (Table 1).
Data collection
Data collection took place before and after the intervention (baseline and endline assessments). All measures were compiled into a single document, resembling a school activity sheet (see Annex II). Students were informed that there were no right or wrong answers and that it was not a school assessment.
Analysis compared responses for each outcome variable between baseline and endline using non-parametric repeated measures tests, specifically the Wilcoxon Signed Rank test. Group differences at baseline and endline were assessed using Chi-square and Kruskal-Wallis tests. Effect sizes were calculated for all comparisons (r and η2, respectively), and data were analysed with SPSS 28.0.1.1 (14), with statistical significance reported at a 95% confidence level (p-values).
Given the small sample size (n = 67) and the quasi-experimental cluster-level design, individual randomisation was not feasible due to school-level constraints. The resulting unequal distribution of school years across groups (e.g., 4th vs. 6th grade) may introduce age-related variation. While curriculum content and intervention materials were standardised across classes, this variation presents a limitation. The use of non-parametric testing was selected to address assumptions of normality and sample size limitations; however, future research with larger, stratified samples would enhance statistical power and enable more precise effect size estimation.
Results and findings
The intervention included 67 students (ages 8-12, M=10, SD=0.873) from 4th to 6th grade at a private school in [city], [country] (table 2).
Differences between pre- and post-tests were analysed for each outcome measure. Non-parametric tests were applied due to sample size and lack of normal distribution (p < .05), with a 95% confidence level. Effect sizes (r) are reported for each Wilcoxon test and interpreted according to Cohen’s convention: small = 0.2, medium = 0.5, large = 0.8.
Nutrition knowledge: Experts, Food groups, General knowledge
Descriptive analysis showed an increase in scores from pre- to post-test across all experimental groups for the three sections. of the nutrition knowledge questionnaire (Experts, Food groups, General knowledge). A Wilcoxon Signed Rank test confirmed these increases were statistically significant in all groups for all sections of the survey (Table 3. Part A - Experts: control: Z = 3.62, p < .001, r = .72; treatment 1: Z = 3.72, p < .001, r = .83; treatment 2: Z = 3.43, p < .001, r = .73; Part B - Food groups: control: Z = 2.82, p = .005, r = .56; treatment 1: Z = 2.18, p = .030, r = .49; treatment 2: Z = 3.75, p < .001, r = .80; Part C - General knowledge: control: Z = 2.30, p = .022, r = .46; treatment 1: Z = 3.04, p = .002, r = .68; treatment 2: Z = 2.98, p = .003, r = .64). Effect sizes (r) indicate large effects for most comparisons. In the Experts section, all groups showed large effect sizes highlighting strong practical impact. In the Food groups section, digital games showed a large effect, while control (r = .56) and posters (r = .49) reflected moderate to large effects. For General knowledge, all groups demonstrated moderate to large effects (r = .46 to .68). These results suggest that the intervention had meaningful impacts on nutrition knowledge across different formats, with digital games generally yielding the strongest effect sizes.
Food label literacy
A descriptive analysis indicated an increase in the scores achieved from pre-test to post-test in all experimental groups for both parts of the nutrition knowledge questionnaire. A Wilcoxon Signed Rank test confirmed these increases were statistically significant in all groups for all parts of the survey (Table 4 - Part A - Experts: control: Z = 3.62, p < .001, r = .72; treatment 1: Z = 3.72, p < .001, r = .83; treatment 2: Z = 3.43, p < .001, r = .73; Part B - Food groups: control: Z = 2.82, p = .005, r = .56; treatment 1: Z = 2.18, p = .030, r = .49; treatment 2: Z = 3.75, p < .001, r = .80;Part C - General knowledge: control: Z = 2.30, p = .022, r = .46; treatment 1: Z = 3.04, p = .002, r = .68; treatment 2: Z = 2.98, p = .003, r = .64).
Effect sizes demonstrate strong practical effects. The control group exhibited a large effect (r = .59), while the posters group showed a very large effect (r = .81), and the digital games group demonstrated a large effect (r = .75). These findings suggest that both media-based interventions were particularly effective in enhancing students’ ability to interpret food labels, with posters yielding slightly stronger improvements than games. Given the exploratory nature of the study and the large observed effect sizes, these results represent meaningful trends in strengthening children’s food label literacy through both static and interactive media formats.
A descriptive analysis indicated an increase in the scores achieved from pre-test to post-test in all experimental groups, which was relatively higher for the treatment group 1 and treatment group 2. The following table (table x) shows the data. A Wilcoxon Signed Rank test confirmed these improvements were statistically significant in all groups: control (Z = 2.95, p = .003, r = .59), treatment 1 - posters (Z = 3.64, p < .001, r = .81), and treatment 2 - digital games (Z = 3.53, p < .001, r = .75). These results reflect large effect sizes across all groups, particularly in those using media-enhanced interventions. These trends suggest that both poster-based and game-based formats had meaningful and practical impacts on students’ ability to read and interpret food labels, with posters yielding the largest observed effect. (Table 5).
Nutrition self-efficacy
A descriptive analysis indicated that students in all experimental groups tended to perform better in the post-test than in the pre-test, showing increased levels of self-efficacy, as it is possible to observe in the following table (table 20). A Wilcoxon Signed Rank test indicated that post-test scores were statistically significantly higher than pre-test scores in the treatment 2 group - digital games (Z = 1.77, p = .043, r = .38), indicating that children in this group increased their self-efficacy after the interventions in a meaningful way. These trends underscore the potential of media-enhanced interventions to support motivational constructs relevant to behaviour change. Results for the control (Z = 1.77, p = .076, r = .35) and treatment 1 (Z = 1.86, p = .063, r = .42) groups did not reach statistical significance, both showed small-to-moderate effect sizes, suggesting potential trends toward improvement. (Table 6).
Barriers to change
A descriptive analysis indicated that students in both treatment group 1 and treatment group 2 had better scores, than students in the control group. The following table shows the data achieved in this analysis, both descriptive values and test statistic values (standardised). A Wilcoxon Signed Rank test revealed that only the treatment group 1 - posters and flyers - showed a statistically significant improvement from pre- to post-test (Z = 2.03, p = .045, r = .45), indicating a moderate effect, thus suggesting a trend in reducing perceived barriers to healthier eating. Although the treatment 2 - digital games group showed some improvement but not statistically significant (Z = 1.12, p = .265, r = .24), and the control group remained unchanged (Z = 0.10, p = .920, r = .02). These findings support the potential of static communication materials to influence self-regulatory constructs like perceived barriers, while suggesting that interactive digital tools may require further tailoring to impact this specific domain (Table 7).
Nutrition self-care
A descriptive analysis indicated an increase in the scores achieved from pre-test to post-test in all experimental groups for nutrition self-care practices. A Wilcoxon Signed Rank test confirmed these increases were statistically significant in the treatment group 1 - posters ( Z = 2.86, p = .005, r = .005) and in the treatment group 2 - Digital games (Z = 2.00, p = .045, r = .43), indicating large and moderate effect sizes respectively and suggesting that students in these groups have increased their nutrition self-care practices after the intervention in a practical and meaningful way. On the other side, control group (Z = 0.14, p = .886, r = .03) showed no significant change, reinforcing the value of media-enhanced interventions in fostering healthier self-care behaviours. (Table 8,).
Actual behaviour - Frequency of fruits, vegetables and of sugar-sweetened drinks consumption
A descriptive analysis indicated that, in general, students in all experimental groups performed better in the post-test than in the pre-test, indicating an increase in F&V and decrease of SSB consumption. A Wilcoxon Signed Rank test confirmed statistically significant changes only for treatment group 2 - digital games: fruit consumption (Z = 2.15, p = .032, r = .46), vegetable consumption (Z = 1.97, p = .049, r = .42), and SSB consumption (Z = -2.53, p = .012, r = .54). These results reflect moderate to large effect sizes, suggesting a meaningful behavioural shift in the digital games group. No significant differences were observed in the control or poster groups, where effect sizes were small.
Actual behaviour - 24-hour recalls
The 24-hour recall was conducted for both ‘yesterday’, which represented a weekday, and Saturday (a weekend day). The following table (Table 10) presents the descriptive and inferential results achieved in this analysis. In general, students slightly performed better from pre-test to post-test, which sometimes is not observed in the median score, but in the range values, showing higher minimum and maximum values. A Wilcoxon Signed Rank test indicated that post-test scores were statistically significantly higher than pre-test scores for:
- Fruit consumption - treatment 2 group (digital games): yesterday (Z = 1.98, p = .048, r = .42) and Saturday (Z = 2.08, p = .038, r = .44); and treatment 1 group (posters): Saturday (Z = 2.17, p = .030, r = .49).
- Vegetable consumption - treatment 2 group: yesterday (Z = 2.07, p = .039, r = .44) and Saturday (Z = 2.95, p = .003, r = .63); and treatment 1 group: Saturday (Z = 2.67, p = .049, r = .60).
These effect sizes were moderate-to-large, supporting the practical significance of the findings. Improvements were particularly evident during weekends, where school meal plans no longer structured intake, highlighting the influence of intervention materials on children’s autonomous food choices.
Discussion
This study evaluated the effectiveness of a school-based nutrition intervention using communication technologies to increase F&V intake and reduce SSB consumption among children aged 8 to 12. According to WHO (2020; WHO/FAO, 2003), children should consume over 400g or five portions of F&V daily, with sugar intake limited to 5-10% of total energy. A diet rich in F&V and low in sugar helps prevent malnutrition and diseases like cardiovascular disease, type 2 diabetes, and obesity (EC-HPDPG, 2022; DGE, 2016; DGAC, 2015). Poor diets are linked to health risks and premature deaths (WHO, 2020).
At baseline, children consumed fewer F&V than recommended. Moreover, on weekends children ate even less F&V and more SSB than on weekdays. These findings are likely influenced by school nutrition plans, which provide structured meals and healthier environments. Higher SSB intake on weekends may reflect relaxed dietary habits at home. F&V consumption increased in all groups, but the increase was only statistically significant in the digital games group for weekday intake and in both the digital games and posters groups for weekends. No significant changes were found for SSB. Statistically significant improvements in knowledge and food label literacy were observed across all groups, including the control, indicating the intervention’s effectiveness (Reynolds et al., 2012; Kliemann et al., 2016). Self-care also improved significantly in the posters and digital games groups, suggesting that communication strategies and technologies enhance self-care more than traditional teaching methods. The posters group showed better results for barriers to nutrition change, while the digital games group outperformed others in self-efficacy and actual behaviour, as measured by the 24-hour recall and consumption frequency questionnaire. While findings indicate that all groups showed improvements, the digital games group exhibited the most significant behavioural changes, reinforcing previous research on gamification in health education (Suleiman-Martos et al., 2021).
While digital games demonstrated stronger effects on actual BC, posters and flyers were more effective in enhancing self-regulatory constructs, such as nutrition self-care and perceived barriers to change. This difference may reflect the distinct cognitive and motivational qualities of each media. Posters provide accessible, static content that supports reflection, planning, and internalisation of health messages over time. In contrast, digital games offer immersive and interactive environments that promote behavioural practice, reinforce action through repetition and feedback, and engage users affectively, strengthening the link between intention and action. These complementary mechanisms suggest that combining static media to support self-regulation with interactive tools to drive behavioural enactment may enhance the overall impact of school-based health interventions. These different effects can also be better understood through the lens of media theory. The greater behavioural impact of digital games may reflect not only content but also media-specific affordances — such as interactivity, feedback, and immersion — that support self-regulated learning and behaviour rehearsal. Children may also have been more intrinsically motivated to engage with this format, as posited by Uses and Gratifications Theory, which emphasises the selection of media based on perceived benefits and preferences. Additionally, the visible presence of posters in the classroom may have supported reflection and message internalisation through processes of media domestication, whereby communication materials become integrated into the learners’ physical and cognitive environments. Together, these perspectives highlight how media are not neutral channels, but active components of the BC process and how the use of different media may result in different outcomes.
In general, these results indicate a positive potential of nutrition interventions in schools, particularly interventions using digital technologies. Even when only applying traditional teaching-learning strategies, children were able to learn more about nutrition and to better understand information in this field, which represents the first step to effectively adopt healthier nutrition practices in the future (awareness and contemplation phases). On the other hand, results also show evidence that using media technologies and communication strategies within a nutrition intervention promotes a stronger effect in the process of BC, showing that both treatment groups performed better than the control group regarding self-efficacy, self-care and barriers to nutrition change. This indicates that communication and media technologies are powerful tools to be included in an intervention when working with children promoting a positive impact. Moreover, the digital learning games group performed better regarding actual behaviour towards food consumption, showing significantly higher results at endline (after the intervention) in the food frequency questionnaire, for F&V and for SSB, and in the 24-hour recall for F&V (yesterday/ school day). Although the constructs were analysed independently, their selection and positioning within the intervention were grounded in an integrated theoretical framework. This framework views knowledge, self-efficacy, perceived barriers, self-care, and actual behaviour as interrelated components of the BC process, each contributing to progression across stages of change. The observed improvements across these constructs align with the expected sequence and interplay defined by the framework. For instance, increased knowledge and food label literacy are expected to contribute to enhanced self-efficacy, which in turn facilitates self-care practices and helps overcome perceived barriers, ultimately leading to actual behaviour change.
In interpreting the findings, effect sizes were used to evaluate the practical relevance of observed changes. While not all comparisons reached traditional levels of statistical significance, several outcomes (such as improvements in self-efficacy, barriers to change, and dietary behaviours), showed moderate to large effect sizes. These reflect meaningful trends, particularly in the treatment groups using communication technologies. Given the exploratory nature of the study and the small sample size, these trends offer valuable insight into the potential impact of media-based interventions and warrant further investigation in larger, more powered studies. The use of digital games revealed a consistent pattern of stronger behavioural outcomes, suggesting practical significance in real-world applications despite statistical limitations.
Previous studies have shown a positive impact of serious gameplay in health outcomes, and also in nutrition knowledge and behaviour (Ferrari, et al., 2022; Baranowsk et al.,2016; Arnab, 2015; Pakarinen et al., 2017). For instance, a systematic literature review of 25 learning games showed positive effects on learning outcomes (Baranowski et al., 2008). Another systematic literature review of 11 videogames for diabetes showed most had a positive significant impact on knowledge and disease management (DeShazo et al., 2010). A meta-analysis of 64 games promoting healthy lifestyles revealed games had stronger and significant effects behaviour determinants, actual behaviour and health outcomes (DeSmet et al., 2014). In general, Digital games seemed to enhance engagement by making learning enjoyable and encouraging repeated gameplay. They promoted planning, decision-making, and trial-and-error learning, helping to understand risks and overcome barriers without real-world consequences (Baranowski et al., 2016). By reinforcing desired behaviours, games boost self-efficacy and confidence, fostering positive behaviour change.
Limitations and recommendations
Despite its promising results, this study has several limitations. The small sample size (n = 67) and quasi-experimental cluster-level design prevented individual randomisation and resulted in unequal grade-level distribution across groups, potentially introducing developmental differences. While pre- and post-tests were mainly analysed within groups to confirm that observed differences were intervention-related, thus minimising the impact of age differences in the interpretation of results, age-related cognitive variation remains a consideration. The possibility of novelty effects linked to the digital game condition should also be acknowledged, as the engaging and unfamiliar format may have temporarily heightened attention or motivation. Additionally, the digital game was not custom designed in the participants’ native language, and the study lacked long-term follow-up, limiting conclusions on sustained impact. Data were also collected during the COVID-19 pandemic, which restricted greater involvement from the school community and parents. Finally, the research was conducted within a private school context, potentially limiting generalisability to other cultural or socioeconomic settings. These limitations underscore the need for future studies using randomised designs, longitudinal tracking, qualitative components, and more diverse samples.
Future research should explore longitudinal effects, involve larger samples, and integrate qualitative insights to assess engagement levels. In addition, although the current study examined each behaviour change construct independently, future studies should investigate the theoretical and causal pathways that connect these constructs—such as how increased nutrition knowledge may enhance self-efficacy, reduce perceived barriers, and ultimately influence behaviour. Larger sample sizes would allow for the application of multivariate methods to better capture the interrelations and mechanisms underpinning BC in response to communication-based interventions. In addition, future interventions should consider integrating parental involvement and community-level strategies to reinforce BC beyond the classroom. In parallel, practical implementation aspects must be addressed, including ensuring equitable access to digital tools, providing adequate teacher training, and aligning digital content with school curricula. These considerations are essential to enhance scalability, sustainability, and real-world impact of school-based digital health interventions.
Conclusions
This study underscores the potential of media communication technologies, especially digital games, in delivering cost-effective and engaging school-based health interventions. While further research is still needed to assess long-term impact of media technologies in health interventions and across diverse populations, findings indicate digital tools offer a scalable and effective alternative to traditional methods. The effectiveness of digital media in school-based nutrition interventions was demonstrated through its ability to enhance nutrition knowledge, food literacy, and dietary habits. Although all interventions increased awareness, digital games proved more impactful in facilitating actual behaviour. Future research should investigate strategies to sustain these effects long-term and explore the broader application of digital tools in health promotion.
Effective health communication remains a critical challenge, requiring ongoing research, innovation, and investment. Bridging theory and practice, communication plays a vital role in driving sustainable behaviour change and improving public health.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Print
