A Scalable Trans Diagnostic Intervention Targeting Adolescent Agency Supported by Conversational AI (AGENCIA)

A Randomized Controlled Trial Protocol of a Scalable Trans Diagnostic Intervention Targeting Adolescent Agency Supported by Conversational AI

The aim of this clinical trial is to evaluate whether AGENCIA, a brief psychological program supported by digital technology and artificial intelligence, can help reduce emotional and behavioral difficulties in adolescents aged 12 to 18. These difficulties may include irritability, impulsive behaviors, conflicts at home or at school, or difficulties in managing intense emotions. The study also aims to determine whether the effects are similar across adolescents with different symptom profiles or neurodevelopmental characteristics.

Participants will be randomly assigned to one of three groups:

AGENCIA Digital: a self-guided online version completed at home. AGENCIA in-person with a digital assistant: a clinician-delivered version supported by an interactive digital assistant to guide the exercises.

Digital psychoeducation (control): a self-guided online program providing general information about adolescent well-being.

The main research questions are:

Does AGENCIA reduce overall emotional and behavioral difficulties? Does the program improve functioning, family accommodation, and personal agency (a young person's sense of being able to act and make changes)? Are the effects similar across adolescents with different profiles or neurodevelopmental characteristics?

Participants will:

• Complete three structured sessions depending on their assigned group.
• Complete brief online questionnaires at baseline (T0), immediately after the sessions (T1), and at 1-month (T2) and 6-month (T3) follow-ups.
• Receive brief phone calls during follow-ups to support questionnaire completion.

A total of 465 adolescents will take part in the study. Participation is voluntary and does not replace usual clinical care. The study does not involve medication or invasive procedures, and all digital tools operate within secure institutional systems.

This study is situated at the intersection of mental health, technological innovation, and artificial intelligence, incorporating advanced digital tools and conversational systems as supportive elements for psychological assessment and intervention in clinical settings. Within this framework, the project explores how technology can facilitate the delivery of brief, scalable, and more personalized interventions in child and adolescent mental health.

This study is a three-arm randomized clinical trial designed to evaluate the effectiveness of AGENCIA, a brief, mechanism-focused psychological intervention for adolescents aged 12 to 18 receiving care in routine mental health services. Adolescents in this age range often experience emotional and behavioral difficulties that affect daily functioning at home, school, and in social contexts. These difficulties may include irritability, impulsivity, conflicts with caregivers, or challenges in managing strong emotions, and they may occur with or without a formal neurodevelopmental diagnosis.

AGENCIA is a structured psychological program delivered in three sessions. It aims to strengthen personal agency, improve emotional regulation and self-compassion, support action toward meaningful goals, and help families reduce accommodation behaviors that may limit adolescents' autonomy. The study evaluates two delivery formats: a self-guided digital version and a clinician-delivered in-person version supported by an interactive conversational assistant. A third group receives digital psychoeducation, which provides general information about adolescent well-being without active therapeutic components.

The trial will recruit 465 participants, who will be randomly assigned in a 1:1:1 ratio to one of the study groups. The design is trans diagnostic and dimensional, including adolescents with emotional or behavioral difficulties that interfere with daily functioning, regardless of diagnostic status. Recruitment takes place across child and adolescent mental health services and primary care settings in Spain.

Assessments are conducted at baseline (T0), immediately post-intervention (T1), at 1-month follow-up (T2), and at 6-month follow-up (T3). All assessments are digital and include measures of emotional and behavioral difficulties, functioning, quality of life, family accommodation, and personal agency.

The primary outcome is the change in the Strengths and Difficulties Questionnaire (SDQ-Total Difficulties Score) from baseline to the 1-month follow-up. Secondary outcomes include symptom subscales, emotional dysregulation, functioning, interference in daily life, quality of life, family accommodation, and personal agency. Exploratory analyses will examine whether early changes in personal agency or family accommodation predict later improvement, and whether treatment effects differ across clinical profiles or neurodevelopmental traits.

The study is considered low-risk. It does not involve medication, invasive procedures, or unsupervised generative AI tools. Digital components run on secure institutional servers. Safety monitoring includes routine screening for acute clinical risk at multiple time points, with established referral procedures within the mental health system.

Participation is voluntary, and adolescents continue to receive their usual clinical care during the study. Data are de-identified, securely stored, and may be shared upon reasonable request and ethical approval.

Inclusion Criteria:

• Adolescents aged 12 to 18 years at enrollment.
• Presence of emotional or behavioral difficulties causing functional interference (e.g., irritability, impulsivity, emotional dysregulation, conflicts at home or school, avoidance).
• Difficulties compatible with neurodevelopmental profiles, regardless of formal diagnosis.
• Ability to use digital materials through a personal device and basic reading skills in Spanish.
• Availability to attend assessments and participate in the intervention and follow-up schedule.
• Informed consent from caregivers and assent from the adolescent.

Exclusion Criteria:

• Acute clinical risk at pre-screening or screening (e.g., imminent self-harm risk, severe agitation, aggression, disorganized behavior, or unsafe behaviors requiring immediate clinical care).
• Score of "Severe" on any of the three HoNOSCA screening items (self-harm, substance-related problems, or bullying/social problems).
• Uncompensated sensory, motor, or language barriers preventing valid completion of assessments or participation (e.g., severe visual or hearing impairment, significant receptive/expressive language difficulties, motor limitations preventing device use).
• Intellectual disability defined by screening tools: ABAS-II GAC ≤ 70 or Kaufman Brief Intelligence Test Composite Intelligence Quotient ≤ 70.
• Any condition judged by the clinical team to require immediate alternative care or that would prevent safe participation.