NCT07218302

Eating Disorder Chatbot Optimization

Optimizing an Automated Chatbot to Achieve Efficient, Scalable Treatment for Eating Disorders

• Status: Not Yet Recruiting
• Phase: Phase 2
• Study type: Interventional
• Enrollment: 800 (estimated)
• Sponsor: Washington University School of Medicine · Academic / Other
• Sex: All
• Age: 18 Years
• Healthy volunteers: Not accepted

Summary

The purpose of this study is to optimize an automated digital chatbot intervention for adults with clinical or subclinical eating disorders.

Detailed description

Eating disorders (EDs) are common-affecting an estimated 10% of individuals in their lifetimes-and have disabling consequences being associated with high psychiatric comorbidity, impairment, and mortality. Despite this, less than 20% of individuals with an ED ever receive treatment. When individuals do seek treatment, many are not evidence-based and come with high costs and burden on the healthcare system. One solution to this is a chatbot intervention that is evidence-based, affordable, and accessible for people with EDs. This project will assess four candidate components of a cognitive behavioral therapy (CBT)-based chatbot to determine the optimal chatbot package. CBT is the recommended first-line treatment for EDs, yet there is no research to date on which components are most effective or needed for meaningful change in ED psychopathology. In this study the intervention will be divided into four components targeting: (1) over-evaluation of weight and shape; (2) dietary restraint; (3) emotion dysregulation; and (4) resisting urges to binge. The investigators will recruit 800 adults with a binge/purge-type ED who are not currently in treatment and randomly assign participants to one of 16 experimental conditions (each combination from 1 to all 4 components) or the waitlist-control group. Participants will be assessed on changes in ED psychopathology, ED behavior frequencies, comorbid symptoms (i.e., depression, anxiety), and clinical impairment at 1-, 2-, and 6-months following randomization. An optimized package of these candidate components that is most effective, while remaining efficient, will be identified using multiphase optimization strategy (MOST). The investigators' established relationships with industry and non-profit partners will ensure this optimized package has high potential for scalability and dissemination into the community.

Conditions

• Eating Disorders

Interventions

Type	Name	Description
BEHAVIORAL	Digital Chatbot Program	Wysa is a digital rules-based chatbot program platform offering cognitive-behavioral therapy-based guided self-help modules. Participants in an arm with access to Wysa will engage with the digital chatbot program daily for eight weeks.

Timeline

• Start date: 2026-03-01
• Primary completion: 2029-07-31
• Completion: 2029-07-31
• First posted: 2025-10-20
• Last updated: 2025-10-22

Locations

1 site across 1 country: United States