This study aims to develop and evaluate a virtual companion agent for providing psychological support to adolescents who have recently experienced stressful life events. This is a single-arm, pre-post pilot trial. We plan to recruit 35 adolescents aged 14-25 to receive a 4-week intervention via the virtual agent. The intervention primarily consists of multi-turn conversations focused on helping adolescents cope with stressful events, identify emotions, enhance social support, and adopt positive coping strategies. The primary objectives are to evaluate the effectiveness of the virtual agent in improving psychological symptoms, resilience, emotional intelligence, and perceived stress, and to comprehensively assess its feasibility, acceptability, and safety.
Over the past decade, mental health issues among adolescents have become increasingly prevalent. Approximately 14% of adolescents globally suffer from conditions like depression and anxiety. Adolescence is a critical developmental period, and mental health problems during this stage can have long-term negative impacts on relationships, academic performance, and future careers. However, due to factors such as cost, accessibility, and social stigma, many adolescents struggle to access timely and effective professional psychological services. In recent years, AI-based conversational agents have emerged as a potential solution. For some sensitive adolescents who are reluctant to confide in others or resistant to professional counseling, an anonymous, non-judgmental virtual conversational partner may serve as a lower-threshold avenue for them to open up and seek emotional support. Against this backdrop, this study aims to develop and evaluate a virtual companion agent specifically designed for adolescents. A minimum of 35 participants are required to complete all study procedures. Accounting for an estimated attrition rate of 20%, a total of 50 eligible adolescent participants will be enrolled. Participants will engage in a series of structured conversational interactions with the agent over 4 weeks and undergo multiple psychological assessments before, during, and after the intervention.
Study Type
INTERVENTIONAL
Allocation
NA
Purpose
TREATMENT
Masking
NONE
Enrollment
50
A conversational virtual agent designed to provide psychological support to adolescents through multi-turn dialogues. Intervention components include: psychoeducation, emotion recognition training, coping strategy development, social support exploration, and personalized behavioral task ("to-do list") setting. The intervention lasts for 4 weeks, involving multiple structured session nodes and follow-up reminders during inter-session intervals.
Level of Psychological Symptoms
Level of Psychological Symptoms: Assessed by change in scores on the Brief Symptom Inventory-53 (BSI-53). This 53-item instrument, which encompasses 9 dimensions, is rated on a 5-point Likert scale. A cut-off point of 63 or higher is used to define a positive case.
Time frame: Baseline [week 0], Week 1 of intervention, Week 2 of intervention, Week 3 of intervention, Week 4 of intervention
Psychological Resilience
Psychological Resilience: Assessed by change in scores on the Connor-Davidson Resilience Scale (CD-RISC). The instrument consists of 25 items distributed among three dimensions: Hardiness (13 items), Strength (8 items), and Optimism (4 items). Responses are recorded on a 5-point Likert scale, yielding a total score between 0 and 100, where a higher score denotes a higher level of psychological resilience.
Time frame: Baseline [week 0], Week 1 of intervention, Week 2 of intervention, Week 3 of intervention, Week 4 of intervention
Social Support
1\. Social Support: Assessed by scores on the Perceived Social Support Scale (PSSS) .The Perceived Social Support Scale (PSSS) consists of 12 items, which are divided into three dimensions: family, friend, and other support. Using a 7-point Likert scale, the total score (range: 12-84) is calculated, with a higher score reflecting a higher level of perceived social support.
Time frame: Baseline [week 0], Week 1 of intervention, Week 2 of intervention, Week 3 of intervention, Week 4 of intervention
Emotional Intelligence
Description: Emotional Intelligence: Assessed by change in scores on the Wong and Law Emotional Intelligence Scale (WLEIS). The Emotional Intelligence Scale contains 16 items across four subscales: Self-Emotion Appraisal and Expression, Others' Emotion Appraisal, Regulation of Self-Emotion, and Use of Emotion. 7-point scoring system is used, and the total score for each subscale is used to determine the competency level in that specific dimension.
Time frame: Baseline [week 0], Week 1 of intervention, Week 2 of intervention, Week 3 of intervention, Week 4 of intervention
Coping Strategies
Coping Strategies: Assessed by scores on the Coping Strategy Indicator - Short Form in Chinese (CSIS-SFC).The scale is composed of 21 items divided into four subscales: Task-Oriented Coping, Emotion-Oriented Coping, Distraction Coping, and Social Diversion Coping. It is a 5-point Likert scale.
Time frame: Baseline [week 0], Week 1 of intervention, Week 2 of intervention, Week 3 of intervention, Week 4 of intervention
Participant Satisfaction
Acceptability of the virtual peer agent will be assessed using a Client Satisfaction Questionnaire (CSQ).
Time frame: Within 1 week after intervention completion
Session Completion Rate
The duration of each participant's conversation per session, the total weekly interaction time, and the overall interaction time throughout the intervention period will be recorded.
Time frame: Week 1 of intervention, Week 2 of intervention, Week 3 of intervention, Week 4 of intervention
Dropout Rate
Feasibility will be assessed by the study dropout rate, calculated as the percentage of participants who discontinue the study before the post-intervention assessment
Time frame: From enrollment through the post-intervention assessment (at Week 4)
User Experience
The overall user experience during interaction with the virtual agent will be assessed using the User Experience Questionnaire (UEQ).
Time frame: Within 1 week after intervention completion
Working Alliance
The quality of the therapeutic relationship between the participant and the virtual agent will be assessed using the Working Alliance Questionnaire(WAQ).The instrument consists of 12 items comprising three subscales: Task Goal, Emotional Bond, and Engagement. Each subscale contains 4 items and is rated on a 5-point scale.
Time frame: Within 1 week after intervention completion
Perceived Warmth and Professionalism
Participant perceptions of the virtual agent's interpersonal qualities will be assessed using scales measuring Perceived Warmth and Professionalism.
Time frame: Within 1 week after intervention completion
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.