Co-Designing, Developing, and Testing a Mental Health Platform for Young People Using a Participatory Design Methodology in Colombia: Mixed Methods Study

The study addresses the growing burden of mental health problems among young people globally and in Colombia, where service access is limited and treatment gaps are substantial. Evidence shows long delays between symptom onset and first contact with services, and digital health tools have emerged as potential supports for information seeking and early help, though challenges with access, usability, and retention persist. Participatory design approaches are proposed to improve adoption and engagement by centering user needs. In Colombia, with high prevalence of MH problems and low service use, there are no comprehensive digital tools dedicated to youth in Bogotá or elsewhere. The research question focuses on how to co-design, develop, and preliminarily test a youth-focused MH platform using participatory design to ensure contextual fit, usability, and feasibility in real-world settings.

Prior work highlights major barriers to mental health care access for youth, long delays to treatment initiation, and disparities in digital access between and within countries. Digital mental health tools often face significant attrition and limited longer-term evaluations. Human-centered and participatory design methods are recommended to boost usability and engagement and to culturally and contextually adapt solutions. In Colombia, previous participatory adaptations of digital mental health tools have been reported, but comprehensive platforms for youth are lacking. The literature also underscores ethical and privacy considerations for young users and the importance of integrating digital tools with existing services to improve early detection and intervention.

Design: Mixed methods within a structured research and development cycle aligned with WHO frameworks. Phases included co-design workshops, iterative development, usability testing (alpha and beta), and a 15-month naturalistic observational trial (July 2022-October 2023). Recruitment and participants: Convenience and snowball sampling via peer groups and advisory representatives. Stakeholders included young people, people with lived experience, supportive others (carers, family, friends, educators), and health professionals. Ethics: Approved by Pontificia Universidad Javeriana and Hospital Universitario San Ignacio ethics committee (FM-CIE-0103-21). Informed consent (and assent for minors) collected via REDCap; data deidentified and stored securely. Participants received vouchers. Co-design: Nineteen online workshops (Google Meet), ~6–8 participants per 2-hour session, using surveys, word clouds, Cartesian planes, sketching pads, and scribble sheets. Qualitative analysis in NVivo using thematic analysis (deductive and inductive coding); interrater reliability achieved (Cohen’s kappa=0.490, moderate). Knowledge translation sessions with a multidisciplinary team (researchers, clinicians, legal, engineers, youth) guided mock-up refinement and feature prioritization. Development: Responsive web prototypes (mobile/desktop) created considering feasibility, complexity, prioritization, and compliance. Key implemented features: social media outreach, registration with anonymous option, MH screener (K6) and automated feedback, targeted psychoeducation via integration with Mental Punto de Apoyo repository, tailored recommendations, track-as-you-go monitoring across 13 MH and well-being areas, progress graphs, well-being plan, SOS/emergency information, telecounseling integration with Mentes Colectivas, and basic gamification (avatars). Some proposed features (forums/community, AI chatbot, broad interoperability, public health surveillance) were deferred for the MVP due to resource and safety constraints. Usability testing: Two stages. Alpha: 90-minute one-on-one think-aloud sessions; task difficulty via Single Ease Question (SEQ). Beta: similar protocol plus System Usability Scale (SUS). Session notes thematically summarized and frequencies of comments calculated; attention to negative/divergent feedback. Naturalistic trial: Platform deployed July 2022–October 2023. Quantitative analytics captured usage, feature engagement, and sociodemographic data. Statistical analysis: descriptive stats (medians, IQRs, SDs), chi-square tests for associations (95% CI), using SPSS v23 and Excel. No age restriction to accommodate supportive users; roles recorded (youth, carers, educators, health professionals).

- Co-design participation: 146 total (110 in co-design/development activities; 36 in usability testing). Majority female (70.5%); median age 18 (IQR 16–29); included vulnerable groups (survivors of armed conflict 4.1%, ethnic minorities 3.4%, disabilities 0.7%). - Co-design outputs: Prioritized features included reliable and tailored information, optional anonymous use and multiple profiles, virtual care, screening and continuous monitoring (track-as-you-go), gamification, social media integration, well-being nudges/plan, emergencies support, community participation, chatbot, interoperability, and research/surveillance. UI preferences emphasized appealing, non-childish visuals and inclusive language; privacy clarity and consent complexities for users <14 years noted. - Implemented MVP features: social media advertisements; K6 screener with automated feedback; registration with anonymous option; targeted psychoeducation via integration with Mental Punto de Apoyo; tailored recommendations; track-as-you-go across 13 domains (stress, loneliness, well-being, emotions, physical activity, sleep, subjective health status, self-esteem, relationships, eating, alcohol/substances, bullying, sexual/reproductive health); progress graphs; well-being plan; SOS/emergency info; telecounseling integration; avatars. - Usability testing: Alpha (n=16) median task time 43 s (IQR 23.5–66.5); SEQ median 1.00 (IQR 1–1). Beta (n=20) median task time 28 s (IQR 16–48); SEQ 1.00 (IQR 1–2). Beta SUS median 85.0 (IQR 80–92.5). Most surveys rated easy to use; physical activity survey hardest. SOS considered useful; avatars visually appealing but requests for more customization. - Naturalistic observational trial (Jul 2022–Oct 2023): 435 users total; 314 (72.2%) registered; 121 (27.8%) anonymous. Registered users median age 34 (IQR 23–46); females 73.9%. - Roles among registered: youth 21.0%, carers 26.8%, educators 22.0%, health professionals 20.7%. - Emotional distress (K6): registered 63.7% (mild 21.7%, moderate 22.3%, severe 19.7%); anonymous 61.2% (mild 24.0%, moderate 27.3%, severe 9.9%). - Distress vs role: registered χ²=37.9, P<.001, Cramer’s V=0.20 (small); youth showed higher distress; health professionals more often low distress. Anonymous χ²=17.0, P=.048, V=0.22 (weak). - Engagement duration (registered): 81.2% used for 1 day; 8.9% up to 15 days; 2.5% ~1 month; 7.3% >1 month up to 1 year. Only 18.8% continued beyond first day. Distress level associated with platform use duration (χ²12=24.5, P=.02; V=0.16, weak). - Feature use (registered): K6 screener 100% (mandatory); track-as-you-go 37.6%; telecounseling request 32.5%; well-being plan 10.2%; SOS 4.8%; avatar 4.5%. - Tracking surveys use among trackers (n=118): emotions 27.1%; sleep 21.2%; stress 17.8%; physical activity 8.5%; others <5% each. Continued monitoring beyond initial use was low (e.g., emotions 47%, sleep 44%). Baseline insights: neutral emotion most reported (41%); sleep problems in 96% of respondents; high perceived stress in 71%; median moderate activity 10 minutes/day; median sitting time 210 minutes. - Telecounseling: 102 requests; 26.5% (27/102) completed one session. Reasons: anxiety 67%, sadness 11%, everyday problems 11%. Modalities chosen: chat later 48%, immediate call 33%, immediate chat 19%, call me later 15%, video call 4%. One case (4%) required emergency referral for self-harm/suicide plans.

The study demonstrates that a participatory co-design approach can produce a youth-oriented mental health platform with high usability and contextually relevant features for Colombia. By embedding user preferences—such as optional anonymity, tailored psychoeducation, screening with automated feedback, and telecounseling—the platform addresses early help-seeking and information needs. High SUS scores and favorable task performance indicate strong usability. However, consistent with broader eHealth literature, sustained engagement remains a major challenge; most registered users disengaged after the first day despite positive usability, suggesting that first-session experience, motivation, and integration with existing supports are critical determinants of retention. The significant association between higher distress and longer use implies that the platform is especially valuable for those already experiencing substantial symptoms, supporting its role in early identification and linkage to care. Privacy and consent complexities for minors highlight the importance of design choices (eg, anonymous mode) and policy alignment. Overall, the platform advances digital mental health efforts in a middle-income setting, signaling the need for continued user-centered iteration, stronger service integration, and strategies to boost long-term engagement.

This study co-designed, developed, and tested a minimum viable youth mental health platform for Colombia, achieving high usability and delivering features aligned with user needs, including screening, tailored psychoeducation, continuous monitoring, well-being planning, emergency support, and telecounseling integration. Real-world deployment showed substantial reach and detection of emotional distress but highlighted significant engagement drop-off after initial use. Future work should enhance initial onboarding and sustained engagement, expand customization and gamification, strengthen integration with local services and care pathways, and address youth privacy and consent challenges. Partnerships with industry could improve design fidelity, sustainability, scalability, and data capabilities (e.g., sensors/wearables) while maintaining rigorous, evidence-based approaches.

- Engagement attrition: Despite high usability, only 18.8% continued beyond the first day, limiting longitudinal data and impact. - Feature scope constraints: Community forums, AI chatbot, broad interoperability with local services, and public health surveillance were deferred due to resource, safety, and administrative challenges. - Lack of industry partnership: Limits high-fidelity design, advanced data collection (e.g., sensors/wearables), sustainability, and scalability. - Generalizability: Co-design conducted in Bogotá; adaptations are needed for diverse Colombian regions and minority groups. - Consent and privacy complexities: Legal requirements for parental consent may deter youth use; balancing autonomy with caregiver involvement is challenging. - Sample composition: Lower participation of men and gender-diverse individuals may limit insights into their specific needs and engagement patterns.