The World Health Organization (WHO) defines mental health as “a state of mental well-being that enables people to cope with the stresses of life, realize their abilities, learn and work well, and contribute to their community. It has intrinsic and instrumental value and is a basic human right” [1]. This definition highlights that mental health is more than just the absence of a disorder; it is a major global health concern, with an estimated 20% of disability-adjusted life years (DALYs) worldwide in 2023 [2]. Adolescence is widely recognized as a sensitive period for the development of mental disorders [3-5], whose incidence has been increasing in recent decades [5-8]. The WHO specifically emphasizes the poor mental health among young people, who often struggle to access mental health services and avoid seeking professional help due to fear of stigmatization [9]. Consequently, these disorders often remain undetected and untreated until later in life [2,10]. Depression is highly prevalent among young people worldwide [2,3,5,7,8,11], and suicide is one of the leading causes of death among European adolescents [10-13].

Prevention recommendations for improving adolescent mental health include connecting with other people, self-help strategies after self-harm, accessibility of care, and a supportive school environment [6-8,14,15]. In the wake of the COVID-19 pandemic, which imposed contact restrictions, online interventions for preventive measures have been developed [16]. Initial research indicates that these interventions have the potential to significantly prevent increasing depression scores; however, more evidence is needed regarding their efficacy in addressing anxiety and stress [14,17-24].

Serious games are interventions where gaming elements are an integral and primary method for achieving goals related to mental health [25]. Gamification, in turn, is a technology used in digital interventions that incorporates game elements in nongame contexts (eg, reward or point systems) [26]. Both approaches have recently emerged as promising methods to enhance digital mental health technologies and increase adherence [14,18,23,27,28]. Existing serious games for adolescents typically include adventure worlds with tasks of different complexity levels and social simulations, often addressing anxiety and/or depressive symptoms. However, several important gaps remain. First, many existing interventions are disorder-specific, whereas there is a growing need for transdiagnostic approaches that address the complex and often comorbid nature of adolescent mental health difficulties [29,30]. Second, prior studies have frequently relied on relatively homogeneous samples, limiting generalizability and underscoring the need to include adolescents from more diverse school settings and socioenvironmental contexts [31-33]. Third, prior work has predominantly relied on self-report questionnaires to assess outcomes [20], indicating a need for more comprehensive evaluation approaches that incorporate additional sources of evidence, such as qualitative interview data [34], and follow-up with adult stakeholders directly connected to study participants [31]. In addition, most research on digital mental health interventions for young people has been conducted in the United States and Australia, with comparatively few studies implemented and evaluated in European settings [35]. As a result, there is limited evidence regarding the feasibility, acceptability, and effectiveness of such interventions across diverse European cultural, educational, and health care contexts. This is particularly important given that differences in school systems, service provision, and sociocultural factors may influence both engagement with and outcomes of digital interventions, necessitating contextually tailored approaches. Furthermore, there is increasing recognition—particularly within European research and policy frameworks—of the importance of involving adolescents [30,36,37] and key stakeholders (eg, parents and teachers) [38] throughout the design and implementation process to ensure relevance, usability, and sustainability. Taken together, these limitations highlight the need to develop and rigorously evaluate novel, transdiagnostic, and contextually adapted digital interventions for adolescents. The SMILE (Supporting Mental Health in Young People: Integrated Methodology for Clinical Decisions and Evidence-Based Interventions) project addresses these gaps by developing a serious game designed to support young people in building meta-skills and cognitive competencies that foster resilience to everyday stressors and enhance overall well-being.

The European project SMILE aims to promote resilience in young people by providing a gamified platform that incorporates elements of digital cognitive behavioral interventions. This platform is designed to achieve change in mental health and well-being by enhancing key psychological factors, such as cognitive flexibility, self-efficacy, critical thinking, self-regulation, and self-confidence. Within the app, algorithms will provide a gamified environment focused on mental health, including virtual self-assessments and self-monitoring, all integrated into a mobile app. The meta-skills and cognitive competencies targeted in SMILE are core components of cognitive behavioral therapy (CBT) and can be understood as transdiagnostic skills to promote adaptive coping across a range of emotional and behavioral challenges by identifying and changing maladaptive thought patterns. Although digital mental health tools have been increasingly used to improve access to mental health support, many such tools primarily focus on symptom monitoring or psychoeducation and suffer from low engagement and adherence to the interventions [39]. A contributing factor to issues of low engagement and adherence may be a lack of digital interventions available that systematically operationalize CBT principles into interactive, developmentally appropriate activities to support skill acquisition and real-world implementation. SMILE aims to address this gap by embedding the digital intervention with structured, gamified, CBT-based techniques, supported by self-monitoring and tailored feedback, to facilitate the development of these meta-skills and cognitive competencies into real-world contexts.

Additionally, SMILE will measure specific gaming behaviors, known as “digital biomarkers,” described in this protocol as in-game measures in Table 1. These measures have been previously used to assess, for example, social anxiety by collecting proximity measures and movement patterns in interaction with an avatar [40] or engagement in social scenarios [41], motor and balance skills as biomarkers for physical well-being in exergames [42], and emotional regulation skills by cheat code usage [27] or to identify nonplayer characters’ (NPC) emotions [14].

An important element of the CBT engagement and digital biomarkers within the game is understanding how reflection on these aspects enhances learning and therefore, psychological resilience. Feedback in digital mental health tools is grounded in experiential learning, cognitive behavioral models, and self-regulation frameworks, supporting users as they practice and refine coping strategies. Within serious games, game-based feedback helps users test actions, observe outcomes, and transfer these skills from digital rehearsal to real-world situations. Delayed feedback, outside of the game, can deepen reflection and understanding of thoughts and behaviors.

The SMILE approach places special emphasis on cocreation, defined as “the participation of users/consumers in the development process (…) significantly improving the quality, usability, and social acceptance of new solutions” [43]. Our previous studies [44] within this project included focus groups and workshops organized with adolescents and young adults in 7 participating European countries (aged 10‐24 years), as well as the Living Labs testing phase, where a first version of the game was tested by a small number of representatives from each age group (see a comparable approach with Living Labs in the study by Santonen et al [45]). The current stage of the project comprises 2 phases: a proof-of-concept study using a cluster randomized multisite multiarm adaptive trial design and key stakeholder interviews.

The overarching aim of the proof-of-concept study is to investigate the acceptability, feasibility, and preliminary efficacy of the gamified platform designed to enhance the well-being of adolescents and young adults in a pragmatic cluster randomized multisite multiarm adaptive trial. This study has 2 primary objectives. First is to investigate feasibility and acceptability in terms of reach (ie, participation, engagement, and adherence), preliminary efficacy operationalized as greater improvement of psychological distress, well-being, and resilience posttest, in the experimental conditions compared with the control period as the primary outcome, and engagement with the gamified intervention platform. This will provide the basis for assessing the public health impact of the use and engagement with the gamified platform. To provide context for the stated hypotheses, the SMILE intervention is delivered via 2 interconnected smartphone apps: the SMILE Game App (Nurogames and DigitalTwin Technology), which provides 5 modules of interactive CBT-based gameplay, and the SMILE Companion App (RDIUP), which facilitates self-monitoring, experience sampling methods (ESMs), and the collection of study metrics. Participating clusters are randomized to 1 of 2 intervention variants: a feedback condition or a no-feedback condition. While both groups engage with identical game content and monitoring assessments, participants in the feedback condition additionally receive automated, personalized data visualizations within the SMILE Companion App. These visualizations integrate in-game metrics, daily mood, and contextual data, aiming to enhance experiential learning, self-reflection, and the transfer of acquired skills to real-world contexts. A comprehensive description of these digital tools is provided in the “Methods” section. Specifically, we expect the following:

Our second aim is to identify contexts, psychological processes, and mechanisms of change, and how these are associated with outcomes of the intervention in participants. To this end, we will test the following hypotheses:

Finally, we will conduct exploratory analysis to map digital biomarkers from in-game metrics and diaries to subjective anxiety and mood scores collated from in-game, ESM, and reflective exercises from the SMILE Companion App.

The aim of this part is to understand how participants leverage the SMILE intervention to support their mental health and well-being and to evaluate the process of implementing the intervention within their day-to-day contexts using a realist evaluation framework [46] to identify in vivo configurations of contexts, processes, and mechanisms of change, and how these are associated with outcomes of implementation and intervention. To achieve this, we will conduct interviews with young people who participated in the proof-of-concept study, their legal guardians, school professionals, and clinicians. The objectives are to (1) gain deeper insight into how, when, and in what contexts adolescents used SMILE tools, and how useful and engaging they perceived them to be; and (2) identify barriers and facilitators relevant to the future use and implementation of SMILE tools in educational and clinical settings.

This study is conducted as a series of single-site trials that are governed locally and share a common protocol but obtain individual ethical approvals from their respective institutional review boards or ethics committees prior to commencing the study. Sites will be linked through a common protocol and measures and a cross-site data sharing agreement which will allow for combined analysis of findings. Specifically, the trial was approved by the Institute of Psychology at SWPS University in Warsaw, Poland (opinion number 23/2025); the Research Ethics Committee at the Faculty of Arts, University of Maribor (038-26-192/2024/27/FF/UM-4) in Maribor, Slovenia; University of Edinburgh School of Health in Social Science Research Ethics Committee (24-25CLPS015) in Edinburgh, United Kingdom; Heriot-Watt University Social Sciences Ethics Committee (2024-4831-13233) in Edinburgh, United Kingdom; Cyprus National Bioethics Committee (102283) in Nicosia, Cyprus; The Ethics Committee for Research with Medicaments in the Health Areas of Valladolid (PI-25‐358 O) in Valladolid, Spain; Ethics Committee of the Medical Faculty Heidelberg University (S-623/2024) in Heidelberg, Germany; and Bioethics Committee of the University of Bologna (0174183) in Bologna, Italy.

For participants who are minors, both the participant’s informed assent and the legal guardian’s informed consent are required. Only minors for whom both forms of agreement are obtained are eligible to participate in the study. The age threshold requiring legal guardian consent varies across participating countries, and the study procedures were implemented in full compliance with the respective national regulations. Specifically, in Poland, Spain, Germany, Cyprus, and Italy, parental or legal guardian consent was obtained for participants younger than the age of 18 years, whereas in Slovenia, such consent was required for participants younger than the age of 15 years, and in the United Kingdom for those younger than the age of 16 years. Participants in the proof-of-concept study do not receive financial reimbursement for their participation. In the stakeholder interview component of the study, participants in Germany will receive small book vouchers valued at €25 (approximately US $28), while no compensation is planned at other sites. Participation in the study is entirely voluntary. All participants are informed, in age-appropriate language, that they may withdraw from the study at any time without providing a reason and without any consequences.

The SMILE game aims to build resources and competencies without requiring players to disclose or explore personal negative experiences. However, gameplay may increase awareness of challenges in participants’ lives. While the game suggests coping strategies for some challenges, it does not address all possible difficulties. Participants (and legal guardians of minors) are informed in advance about available support services in their respective countries. An adverse event monitoring procedure has also been implemented. All adverse events occurring during the trial will be recorded, managed by the investigator as appropriate, and followed up. Events occurring after the completion of the study will also be documented if they are related to the intervention or study procedures. All adverse events and adverse device effects will be listed, statistically analyzed, and assessed by the sponsor for potential causal relationships with the investigation. The parties responsible for conducting the study meet regularly to monitor progress, review any adverse events, and oversee overall trial conduct. Additionally, a Data Monitoring and Ethics Committee has been established, comprising experts in Youth Mental Health, Clinical Trials, and Digital Interventions. The committee is independent from the sponsor and funder.

The proof-of-concept study was prospectively registered with the UK Clinical Study Registry (ISRCTN32523126) on April 29, 2025 by the chief investigator, before the enrollment of the first participant. A full statistical analysis plan will be registered before the end of recruitment and before data are available to the investigators. Any changes to the protocol will be transparently reported when disseminating study results.

In the proof-of-concept study, personally identifiable information (PII), such as names, email addresses, and birth dates, is only stored for the purpose of adverse events handling and is never stored on any central servers. During the setup of each pilot sequence, the users are imported and a pseudonym is created. Each of the study locations individually performs this import and receives a key for mapping users to their pseudonyms. This key is stored locally by a designated investigator of the local study and is kept physically separated from any other data. This local key is only accessed for the purpose of handling adverse events. The SMILE apps, that is, SMILE Game App and SMILE Companion App (see “SMILE Game App” and “SMILE Companion App” sections below), do not store data; they only collect data according to the specifications and send them to the SMILE API for storage and processing. These data are deidentified and pseudonymized and do not contain any PII. Data in-flight are never transmitted in plain text and all connections use Transport Layer Security (TLS) for encryption. The data are stored on a server that uses the Fast Healthcare Interoperability Resources standard. Once the data are stored, access to the data is guarded by the Streaming Attribute Policy Language (SAPL) [47] engine that is part of the SMILE project, allowing for fine-grained access control. The rules for viewing the data are made based on discussions within the consortium to ensure that all parties have access only to the necessary data. Data visualization tools and analytics are available for consortium partners (based on the established rules) using the SMILE DSS (Decision Support System). All SMILE tools use the open-source solution Keycloak for federation, strong authentication, and user management.

In the stakeholder interview component, interviews will be recorded and subsequently transcribed. All transcripts will be anonymized prior to analysis, and recordings will be securely stored and permanently deleted after transcription and verification. Upon study completion, deidentified participant data, along with codebooks and all materials necessary for replication, will be made publicly available in open repositories.

The intervention is delivered through 2 components: the SMILE Game App and the SMILE Companion App. Participants download and use both apps, which are required for participation. Both apps use the same login credentials and are linked through a single account. Both the Game and Companion apps are designed for use across age groups and in a manner that supports simultaneous accessibility for both adolescents and young adults; it was not feasible within the study constraints to provide age-specific content as originally planned.

The game follows a player who moves to the cyberpunk city of Hopetown with their sister, Seraphina. In this world, automatic thoughts appear as small, robot-like creatures. When a cybernetic dragon attacks the city and captures Seraphina, the player embarks on a quest to rescue her. Guided by key characters, the player learns CBT-based techniques to understand and manage negative automatic thoughts. Across 5 story-driven modules, players engage in interactive challenges designed to promote cognitive restructuring, resilience, and healthier thinking patterns.

Module 1: “Hopetown.” The player investigates the city’s attack and discovers that the villain manipulates citizens’ negative automatic thoughts. Challenges introduce the CBT “hot cross bun” model [48], teaching players to distinguish between thoughts, feelings, bodily sensations, and behaviors, identify automatic thoughts, and recognize cognitive distortions.

Module 2: “Starfall District.” In a corrupted district, the player helps NPCs trapped in negative thinking cycles. This module deepens understanding of the links between thoughts, feelings, bodily sensations, and behaviors and introduces cognitive restructuring and paced breathing as tools for breaking negative cycles and strengthening emotional regulation and resilience.

Module 3: “The Technicians’ Quarters.” The player uncovers the technician’s core beliefs to unlock key information to find their sister. This module focuses on identifying and challenging core beliefs. During a fight-or-flight scenario, players must actively apply previously learned CBT-based tools to overcome negative automatic thoughts.

Module 4: “Library.” While searching for information to defeat the villain, the player evaluates evidence and tests the validity of competing narratives. This module strengthens critical thinking and evaluation skills, reinforcing cognitive restructuring and self-regulation techniques.

Module 5: “The Data Vault.” In the final confrontation, the player applies all acquired skills to defeat the villain and rescue Seraphina. A closing social scenario requires assertive communication, targeting fear of negative evaluation, reducing avoidance, and strengthening social confidence through behavioral practice.

Clusters of participants are randomized (see details in the “Randomization: Proof-of-Concept Study” section below) to either a feedback or a no-feedback variant of SMILE. The feedback component is expected to function as an intrinsic motivator for engagement and learning and to facilitate greater transfer of skills acquired in the game to real-world experiences. The feedback system is based on Nicholson’s [49] “Recipe for Meaningful Game Engagement” and delivered via the SMILE Companion App. Feedback is generated across time points using a triad of data sources: in-game metrics, ESMs, and self-report measures. Upon completion of each module, players receive feedback and are invited to explore specific time points of interest. The system provides brief, automated summaries (eg, “your mood appeared to be slightly more positive than earlier in the day”), linking contextual and gameplay data with daily mood and well-being ratings. The SMILE Companion App (see below for more details) presents this feedback in graphical form, integrating in-game, ESM, and reflective exercise data. Visualizations include simple daily averages, trends over time, and basic associations between activity, context, and well-being. Players can select specific indicators to explore how they relate to in-game activities, real-world experiences, and other measured variables. Participants assigned to the no-feedback condition complete the same game modules, ESM prompts, and reflective exercises, but do not have access to personalized feedback or data visualizations.

The SMILE Companion App supports self-monitoring, feedback, and data collection during the proof-of-concept trial. All participants are required to use the app, as baseline and first-week outcome assessments are mandatory. The app delivers self-report measures, ESM prompts, and diary entries via push notifications. Participants can individually adjust the timing of these prompts to fit their daily routines. In addition, users can customize selected features (eg, light or dark mode) and receive rewards for completing study activities. The app also provides an overview of study progress, including the current module, scheduled play periods, and pending questionnaires.

Participants assigned to the feedback condition receive personalized feedback within the app. The feedback acts as an intrinsic motivator for engagement and a link between the skills learned in the game and real-world scenarios and experiences. This feedback is delivered to only half of the sample to evaluate its effectiveness. Feedback is based on Kolb’s experiential learning model [50], Gibbs’ Reflective Cycle [51], and Nicolson’s recipe [49] for linking experiential learning from the game, feelings from the ESM measures, analysis, presented in graphical form from the in-game measures and decision-making in the cognitive restructuring exercises. Reflections on these elements are then used for real-life goal-planning strategies. The clusters of participants who do not receive the feedback will still participate in the game and ESM exercises, but will not be able to visualize and reflect on their data.

Proof-of-concept study uses a cluster randomized multisite multiarm adaptive trial design, meaning that randomization occurs at the cluster level and involves assigning each cluster to either feedback or no-feedback variant. Randomization is conducted in blocks (using block randomization) at the consortium level using REDCap software (Vanderbilt University) by a statistician who is not involved in the study. Confirmed clusters for the upcoming sequence were randomized and this process was repeated for new clusters before each sequence. Clusters are allocated pragmatically between sites and adjusted in size and distribution if indicated by the adaptive interim analyses. There is no randomization at the individual level. Clusters will be blinded to their allocation (ie, whether they receive the intervention with or without feedback). No unblinding is planned during the trial.

There are 7 types of data that we will collect in this phase. First, there are self-reported primary and secondary measures that will be collected at the following time points (Table 2): baseline (T0; 1 to 2 weeks before the intervention), right before the intervention (T1), during the intervention (T2), and at posttest (T3; after completion of the intervention). Additionally, measures of depression and anxiety (ie, primary outcomes) will be distributed on a weekly basis, alternating between a depression scale 1 week and an anxiety scale the next. Secondary measures (ie, well-being and resilience) will be collected every other week. Second, participants will be invited to respond to a short daily survey using ESM, 5 times a day (Table 3). Third, there will be in-game measures related to the current activity in the game (Table 1). Fourth, we will collect vocal and facial data via weekly diary recordings. Fifth, we will assess the feasibility and acceptability of the game through measures other than participants’ responses. Finally, we will collect user data that are restricted to the SMILE apps: the Game App and the Companion App. All measurements described below will be collected by the SMILE end user apps.

Primary outcomes are depression and anxiety. To measure anxiety, we use the Penn-State Worry Questionnaire for Children (PSWQ-C [60]) for participants aged 10‐17 years and the Generalized Anxiety Disorder-7 item (GAD-7 [61]) for participants aged 18‐24 years. To assess depression, we use the Patient Health Questionnaire for Adolescents (PHQ-A; [62]) for participants aged 10‐17 years and the Patient Health Questionnaire-9 item (PHQ-9; [63]) for participants aged 18‐24 years. Completion of primary outcome measures at T0 and T1 is mandatory within the SMILE apps.

Main secondary outcomes include self-reported well-being and resilience, assessed with measures appropriate to the age group: the Warwick–Edinburgh Mental Well-being Scale (WEMWBS [64]) for ages 10‐24 years, the Child and Youth Resilience Measure (CYRM-12 [65]) for ages 10‐17 years, and the Brief Resilience Scale (BRS [66]) for ages 18‐24 years. Moreover, we measure online privacy concerns [67], technology acceptance [56], and user experience for both apps using the short User Experience Questionnaire [68]. Completion of secondary outcome questionnaires is optional for participants.

In addition to questionnaire data, selected secondary outcomes will also be derived from daily ESM responses (Table 3).

In addition to the primary and secondary outcomes, we will assess feasibility and acceptability outcomes, user data, in-game metrics, and diary recordings.

Feasibility and acceptability outcomes will include (1) cluster recruitment yield, defined as the proportion of clusters included relative to the total number of clusters invited, with a predefined success criterion of at least 80%; (2) uptake, defined as the proportion of participants who completed the minimum intervention dose (defined as completing baseline measurement and at least one game module) relative to all included participants, with a predefined success criterion of at least 66%; (3) engagement with the SMILE serious game based on in-game measures (Table 1); (4) adherence, operationalized as the average proportion of completed game modules relative to the total of 5 modules, with a predefined success criterion of at least 50%; (5) dropout, defined as the proportion of participants who disengaged from the study at any point between baseline and posttest relative to all participants, with a predefined criterion of less than 45%; (6) adverse events, assessed and summarized.

We will also collect user data restricted to the study apps (ie, the Game App and Companion App). User data will include app usage frequency; drop-off points (where users typically stop using the app); session duration; session start and end times; whether sessions are completed with or without interruptions; completion rates; frequency of engagement with the game; and the average time spent per session.

During SMILE gameplay, the Game App will collect various in-game metrics. These include measures of behaviors such as proximity to NPCs, performance in specific tasks, or in-game affect ratings. Table 1 provides a detailed description of these metrics.

As part of the intervention, in the SMILE Companion App, participants will be invited to complete brief weekly diary video recordings (approximately 3‐4 prompts) and longer recordings at baseline and posttest (approximately 8 prompts). Baseline and posttest prompts are general (eg, self-perception, mood and energy, coping, self-esteem, and potential anxiety or depression symptoms), while weekly prompts are tied to the current game module and ask participants to reflect on their experience. Recordings are encrypted and uploaded; audio and video streams are separated on the server and are not viewed or listened to by humans. Data remain encrypted until automated processing, and speech transcriptions are not stored. AI-based processing will extract exploratory digital markers previously linked in the literature [68-70] to anxiety and depression, including facial movement features (eg, action units, gaze, and head movement metrics) and acoustic or text-derived features (eg, energy and other signal parameters; counts of words and sentences and symptom-related terms), to examine associations with validated self-report outcomes and inform development of an explainable screening approach in young people.

Semistructured interviews with adolescents who have completed the SMILE intervention, as well as with their parents or legal guardians, school professionals, and clinicians, will cover (1) the main concerns of children and young people; (2) perceived resources and coping supports; (3) feedback on the game from parents and school professionals; (4) perceived barriers to implementing SMILE tools at home and in school; and (5) perceived facilitators to implementing SMILE tools at home and in school.

We will conduct a process evaluation to examine the implementation of the SMILE intervention and to understand what works, for whom, under what circumstances, and why. Quantitative data will describe outcome patterns and assess implementation and intervention fidelity, while qualitative data will help explain the underlying processes shaping these patterns. Using a mixed-methods realist evaluation approach[23], we will examine how contextual factors, implementation mechanisms, and outcomes interact within the adolescent population, as well as how broader socioeconomic and environmental factors influence engagement and decision-making. This approach will also enable us to explore how adolescents adapt SMILE to meet their individual needs.