Generative AI chatbot may optimally influence patient behavior

Key takeaways:

• Presenters said a generative AI chatbot may be able to optimally influence patient behavior.
• The chatbot uses patient archetypes backed by the transtheoretical model of behavioral change.

ORLANDO — A generative AI chatbot may be able to optimally influence patient behavior, according to the winning abstract from the 2024 Abstract Competition at the AIMed24 Annual Meeting.

“Traditional methods to approach behavioral change in patients nowadays face significant drawbacks and significant challenges in the sense that they do not often account for the complexity and diversity of the psychology or the persona of the patient,” Samer Salman, an undergraduate student at the University of Texas at Austin, said.

In their presentation, Salman and Isha Tripuraneni, undergraduate student at the University of California, San Diego, discussed the capacity for a generative AI chatbot to personalize behavioral interventions through the development of patient archetypes backed by the transtheoretical model (TTM) of behavioral change.

Salman said patients would first complete a questionnaire to help determine which of the seven patient archetypes most accurately reflects the patient's “attitudes or beliefs toward a condition or health care or even a behavior or a habit that they have.”

Tripuraneni added that the patient archetype would be paired with a patient’s development through the five stages of the TTM to create a personalized behavioral nudge.

“The TTM informs the content of these nudges,” Tripuraneni said. “For example, if we have a patient in the preparation stage, they may use this process called self-liberation, which essentially provides this commitment to change through a specific action plan.”

Salman concluded that the ideal function of the technology would be to “deliver the most optimal behavioral nudge and more effectively influence patient behavior.”