Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
Many young people in the Middle East face anxiety problems, and social stigma often stops them from getting help. Doctors' usual talks sometimes miss the right diagnosis, and most treatments only target one part of anxiety (thoughts, feelings, or body reactions). This study tested a free mobile app designed to help teens aged 13-19 manage anxiety better on their own or with support.
The app uses artificial intelligence (AI) to chat with users, ask questions based on standard medical guidelines, and suggest a personalized plan. It combines four proven therapy styles:
Users can also write in a private journal, review their past entries to track patterns, join video group sessions with other teens (after a short readiness check to ensure a good fit), and try virtual reality exercises to face fears safely. The app works in many languages, including Egyptian Arabic dialects, keeps chats private and encrypted, and only shares info with a psychiatrist for diagnosis help if needed. It also uses AI to read facial expressions and voice tone for real-time feedback on emotions.
We enrolled 587 teens (some used the app, others did not as a comparison group) and measured how well the app diagnosed anxiety and reduced symptoms.
Anxiety disorders show rising prevalence in the Middle East (estimated at 6.33% in recent data), compounded by cultural stigma that limits help-seeking. Conventional diagnostic interviews are prone to misclassification, and unimodal therapies often fail to address the full cognitive, emotional, and physiological dimensions of anxiety.
This randomized controlled trial evaluated a mobile application delivering an integrated, AI-assisted intervention for adolescents aged 13-19 years with anxiety symptoms or disorders. The app employs an unstructured diagnostic model powered by the Gemini API, synthesizing DSM-5 and SCID-RV criteria via natural language processing to inform personalized treatment pathways.
The core intervention integrates four evidence-based psychotherapeutic modalities within a single platform:
Additional components include:
The trial used a parallel-group design with an experimental arm (app access; n=217) versus control (n=370; likely treatment-as-usual or waitlist based on standard practices).
Not provided
Not provided
Not provided
Not provided
| Label | Type | Description | Intervention Names |
|---|---|---|---|
| Mobile App Intervention | Experimental | Participants in this arm received access to a multilingual mobile application designed for adolescents aged 13-19 with anxiety disorders. The app featured:
|
|
| Control Group | No Intervention | Participants in this arm did not receive access to the mobile application or any study-specific intervention during the trial period. They continued with standard care or no additional anxiety management support (treatment as usual or equivalent passive control; n=370). This group served as the comparison to evaluate the efficacy of the mobile app intervention on anxiety symptoms (measured by SCAS) and diagnostic performance. |
| Name | Type | Description | Arm Group Labels | Other Names |
|---|---|---|---|---|
| Ísland app | Behavioral | A mobile application delivering an integrated behavioral intervention for anxiety disorders in adolescents aged 13-19. It uses an AI chatbot for diagnostic support (DSM-5/SCID-RV criteria) and personalized delivery of four evidence-based therapies: Cognitive Behavioral Therapy (CBT), Acceptance and Commitment Therapy (ACT), Mindfulness-Based Cognitive Therapy (MBCT), and Mindfulness-Based Stress Reduction (MBSR). Additional features include secure journaling, session history review, AI real-time emotional feedback via facial expression and voice-tone analysis, video-based group therapy with readiness screening (AGTRQ), Virtual Reality Exposure Therapy (VRET) modules, multilingual support (including Egyptian Arabic dialects), and encrypted data privacy with limited psychiatrist access for diagnostic purposes only. |
| Measure | Description | Time Frame |
|---|---|---|
| Change in Anxiety Symptoms (SCAS Total Score) | The primary outcome is the change in anxiety symptom severity, measured by the total score on the Spence Children's Anxiety Scale (SCAS; child self-report version, 44 items, 0-3 Likert scale per item, total range 0-114, higher scores indicate greater anxiety severity). The SCAS assesses symptoms aligned with DSM anxiety disorders (e.g., separation anxiety, social phobia, generalized anxiety, panic, obsessive-compulsive, physical injury fears). A clinically significant reduction is defined as exceeding the 4-6 point Minimal Clinically Important Difference (MCID) threshold established for pediatric anxiety measures. This outcome evaluates the efficacy of the mobile app intervention in reducing anxiety symptoms compared to the control group. | Baseline to post-intervention (8 weeks) |
| Diagnostic Performance of the AI Chatbot (Sensitivity) | Sensitivity (true positive rate) of the app's AI chatbot diagnostic model (based on DSM-5 and SCID-RV criteria via Gemini API) in correctly identifying anxiety disorders compared to structured clinical assessment or expert confirmation. Reported as percentage. This evaluates the accuracy of the unstructured AI diagnostic approach in adolescents. | After 8 weeks of Chatbot use |
| Diagnostic Performance of the AI Chatbot (Specificity) | Specificity (true negative rate) of the app's AI chatbot in correctly ruling out anxiety disorders. Reported as percentage. | After 8 weeks of Chatbot use |
| Measure | Description | Time Frame |
|---|---|---|
| Diagnostic Performance Metrics (PPV, NPV, Accuracy, MCC, F1 Score) | Additional diagnostic validity indicators including Positive Predictive Value (PPV), Negative Predictive Value (NPV), overall accuracy, Matthews Correlation Coefficient (MCC), and F1 score of the AI diagnostic model. These provide a comprehensive evaluation of the chatbot's reliability for anxiety disorder classification in adolescents. Reported as percentages or coefficients. |
Not provided
Inclusion Criteria:
Exclusion Criteria:
Not provided
Not provided
Not provided
Not provided
Not provided
Not provided
| Facility | Status | City | State | ZIP | Country | Contacts |
|---|---|---|---|---|---|---|
| STEM Egypty | Cairo | Egypt |
Not provided
| ID | Term |
|---|---|
| D001008 | Anxiety Disorders |
| ID | Term |
|---|---|
| D001523 | Mental Disorders |
Not provided
Not provided
Experimental (app users, n=217) vs. control (n=370), with no crossover, switching, or sequential elements.
Not provided
Not provided
Not provided
Not provided
|
| After 8 weeks |