National Eating Disorders Association

Chatbot to Improve Body Image and Prevent Eating Disorders

Ellen Fitzsimmons-Craft, PhD

Preventing eating disorders is of the utmost importance. This is because not only are eating disorders incredibly impairing, but there is also a very wide treatment gap that exists once individuals develop these problems. In our work, including in collaboration with NEDA on our collaborative screening tool, we’ve found that less than one in five individuals with an eating disorder ever receives treatment. This gap exists for lots of reasons—stigma, shame, few trained providers, limited or no insurance coverage for treatment, lack of problem identification due to stereotypes about who experiences eating disorders, not feeling like one’s problem is serious enough to warrant treatment—the list goes on and on.

We know prevention of eating disorders is important, including by way of addressing known risk factors, like body image concerns, but how do we make it widely available? Programs that require human time and resources to implement them are difficult to scale, particularly in our current environment in the U.S. where there is limited investment in prevention. Digital tools might offer one solution, but even with those, our team has found that having the support of a human coach—to provide motivation to continue or ideas about how to apply skills learned to daily life—makes these programs much more effective.

Our team had the idea that one possible solution to this set of problems would be a chatbot. A chatbot is a computer program that simulates a conversation with a human. Think about Siri or Alexa but in your text messages. We had the idea that a chatbot could deliver the core content from our team’s effective (but admittedly old school) web-based eating disorders prevention program in an engaging format—so, in bite-sized bits of information, with infographics, using emojis, and the warm, casual language we have all come to expect in our texts—and mimic some aspects of human moderation. In other words, even though the chatbot was a robot, we thought she could provide some of that motivation, feedback, and support that humans had in our earlier work and maybe even deliver our effective program content in a way that would make people really want to engage. And actually, other work has shown that some individuals may even be more open and honest with a chatbot because it is not a human and because of the anonymity it can provide. Some people may not yet be ready to talk about these issues with a real person, and that is OK.

We were very fortunate to receive funding from NEDA’s Feeding Hope Fund to develop and test this chatbot in a study of 700 women who were already at high risk for the onset of an eating disorder, meaning that they were already experiencing very high levels of concerns about their weight and shape. We partnered with X2AI, a mental health chatbot company, to program the chatbot, named Tessa, to deliver what we referred to as the Body Positive program. The program consisted of a series of conversations that were designed to take about 10 minutes each and covered the core content from our previous digital program in this new delivery format. Content covered topics like challenging the thin body ideal, media and social media literacy, comparisons, healthy eating strategies, and how to handle critical comments about your body.

We randomized the women in our study to either receive the chatbot or to the control condition, where we told them and gave them access to the chatbot after the study. What we found was that those who got access to Tessa right away had greater reductions in their concerns about their weight and shape, and they even sustained these changes up to 6 months out from receiving the chatbot. We also found indications that the chatbot may even reduce risk of eating disorder onset. You can learn even more about our findings in an article we published in the International Journal of Eating Disorders (the article is available open access for the next 2 months) or in this write-up published in The Verge.

So, now what? Now that we developed this scalable, easy-to-use tool—how can people get access to it? This NEDAwareness Week, I am so thrilled to announce that through a partnership between NEDA and X2AI, Tessa is now available free of charge! You can get access by taking the NEDA Screening Tool and seeing if Tessa is a good fit for you. If you’re interested to learn more about Tessa before diving in or would like more info on what to expect when chatting with her, you can check out the very brief video below. Note that while we originally tested the program with women, the program could be helpful for individuals of all genders, and we hope to do more testing with an even more gender-diverse sample in the future.

I am a researcher, and while extremely frustrating, I’ve gotten very accustomed to slow progress—there is a known lag between what the research shows us and what is available to our patients and individuals in the real world (usually 17 years, on average!!). The NEDAwareness Week theme could not be more on point this year, as it is incredible to #SeetheChange so quickly in this instance. To go from our study being published one month to Tessa being available to the public the next is completely unheard of! We have the support of the Feeding Hope Fund and X2AI to thank. This NEDAwareness Week, I also invite you to #BeTheChange—engage with Tessa to improve your own body image or share about our screening tool or Tessa with your network—you never know who might benefit.

Ellen Fitzsimmons-Craft, PhD is an Assistant Professor of Psychiatry at Washington University School of Medicine, St. Louis, Missouri, USA and a licensed clinical psychologist. She has established programmatic lines of research centering on the use of technology for eating disorder prevention and treatment, eating disorder screening, sociocultural etiological and maintenance factors for eating disorders, and eating disorder recovery. Ultimately Dr. Fitzsimmons-Craft’s work aims to disseminate evidence-based interventions from research to practice as well as extend treatments in ways that will reach the large number of people in need of care for mental health problems but who are not receiving services. She is a Fellow in the Academy for Eating Disorders, is a current recipient of a National Institute of Mental Health K08 Career Development Award, and has authored more than 100 peer-reviewed publications that have been collectively cited over 3,000 times in the literature. Dr. Fitzsimmons-Craft is passionate about increasing access to scalable, evidence-based mental health services, collaborating with numerous non-profit organizations, including the National Eating Disorders Association (NEDA) and the National Association of Anorexia Nervosa and Associated Disorders (ANAD), and statewide groups in the U.S. in order to do so. Her work has been featured in high-profile media outlets including Wired, Forbes, Scientific American, Slate, and The Verge.