To assess if AI-led interventions could help reach more people in need, Mathioudakis and colleagues carried out a study where 368 adults (71% female) with a combination of overweight or obesity and prediabetes were randomly assigned to an AI-led intervention, which included a responsive app and connected scale to measure weight loss, movement and meals, or an accredited human-led diabetes prevention program.
Study results show that use of an AI-led diabetes prevention program helped people with prediabetes improve their heath to a similar extent to a human-led program after 12 months or more of use. Approximately 32% of participants in both the AI- and human-led groups achieved the study endpoint of maintenance of glycated hemoglobin (HbA1c) of less than 6.5% plus weight loss of at least five percent, weight loss of at least four percent, and physical activity of at least 150 minutes per week, or reduction in HbA1c by at least 0.2 percentage points at 12 months. The primary combined outcome was achieved by 31.7% of those in the AI-led group and 31.9% of those in the human-led group at 12 months.
【MORE】