Clinical Research Directory
Browse clinical research sites, groups, and studies.
LLM-Generated Coaching Prompts
Sponsor: Stanford University
Summary
This pilot study aims to evaluate whether personalized coaching prompts generated by a large language model (LLM) can effectively increase physical activity levels among participants. The study will involve 50 participants who will receive daily text messages, either personalized by the LLM or generic, over a 14-day period. Participants will share their HealthKit data for analysis. The findings will inform the development of future versions of the My Heart Counts application, enhancing user engagement and health outcomes.
Official title: Enhancing Physical Activity With LLM-Generated Coaching Prompts: A My Heart Counts Study
Key Details
Gender
All
Age Range
18 Years - Any
Study Type
INTERVENTIONAL
Enrollment
50
Start Date
2026-02-20
Completion Date
2026-12-01
Last Updated
2026-01-23
Healthy Volunteers
Yes
Conditions
Interventions
Pre-Generated Personalized LLM Coaching Impact
After the initial 7 days, participants will switch to the other type of prompting for another 7 days. This means that participants who initially received personalized LLM-generated prompts will then receive generic activity prompts, and vice versa. This crossover design allows for each participant to experience both types of interventions, providing a within-subject comparison of the effectiveness of personalized versus generic prompts in increasing physical activity levels.