Developing an Artificial Intelligence (AI) Chatbot for Adolescents and Young Adults With Eating Disorders

Developing an Artificial Intelligence (AI)-Powered Exposure Therapy Chatbot for Adolescents and Young Adults With Eating Disorders

The goal of this study is to develop and refine an artificial intelligence (AI)-powered Exposure Therapy for Eating Disorders chatbot web-app that effectively engages adolescents and young adults and decreases anxiety about weight gain. Participants will use the web-app as a supplemental therapeutic tool during their first six weeks of treatment in a Partial Hospitalization Program. Participants will complete assessments every two weeks, including: (1) semi-structured interviews about their experience with the chatbot web-app; (2) implementation outcome measures (e.g., feasibility, usability, acceptability); and (3) anxiety about weight gain.

The overarching goal of this study is to develop and refine an AI-powered Exposure Therapy for Eating Disorders (Exp-ED) chatbot app that effectively engages adolescents and young adults (AYAs) with eating disorders (EDs) and decreases anxiety about weight gain. The investigators will enroll 20 AYAs (ages 12-26 years) entering the Partial Hospitalization Program (PHP) for EDs at the Penn State Milton S. Hershey Medical Center. The initial version of the Exp-ED chatbot app will be developed by an interdisciplinary team with expertise in exposure therapy for EDs (Dr. Essayli), AI (Dr. Nawab), and implementation science (Dr. Lenker). Participants will use the Exp-ED chatbot app as a supplemental therapeutic tool during their first six weeks of PHP treatment. Participants will complete assessments every two weeks, including: (1) semi-structured interviews about their experience with the Exp-ED chatbot app, including design issues and engagement barriers; (2) implementation outcome measures (e.g., feasibility, usability, acceptability); and (3) anxiety about weight gain. Success will be evaluated using pre-determined benchmarks derived from previous mental health chatbot research. This project has two aims:

Aim 1: Iteratively adapt an exposure therapy chatbot app for AYAs with EDs. The investigators will conduct monthly reviews of mixed-methods data and use the Framework for Reporting Adaptations and Modifications-Enhanced (FRAME) to systematically document and evaluate modifications to the chatbot app. Each revised version will be tested with new participants, creating an iterative development process that continues throughout the project period. All adaptation decisions and their empirical foundations will be documented to create a comprehensive record of the app's development via implementation science frameworks.

Aim 2. Evaluate the feasibility, usability, acceptability, and preliminary efficacy of the Exp-ED chatbot app. H2.1: The investigators will recruit ≥50% of eligible AYAs with EDs. H2.2: Over the six-week project period, participants will use the app ≥18 times. H2.3: Participants will engage with the chatbot for ≥100 minutes total. H2.4: Participants will rate the Exp-ED chatbot app as acceptable (score ≥68 on the System Usability Scale). H2.5: Participants will report a ≥50% reduction in anxiety about weight gain.

Inclusion Criteria:

• Age is between or equal to 12-26 years
• Admitted to the Penn State Health Eating Disorder Partial Hospital Program (PHP) or Intensive Outpatient Program (IOP)
• Have a personal smartphone
• Have a diagnosis of anorexia nervosa (AN), bulimia nervosa (BN), binge-eating disorder (BED), or other specified feeding or eating disorder (OSFED)
• Be English speaking

Exclusion Criteria:

• Diagnosis of avoidant/restrictive food intake disorder (ARFID)
• Have acute suicidality
• Diagnosis of an intellectual disability