AI-Powered Diabetes Prevention Program Intervention Matches Human Coaching in Landmark Trial
A groundbreaking study has revealed that an AI-driven intervention based on the Diabetes Prevention Program (DPP) can match the effectiveness of traditional human-led programs in helping adults with prediabetes reduce their risk of progressing to diabetes. This is a significant finding, as it suggests that AI-powered solutions could potentially revolutionize diabetes prevention, making it more accessible and scalable.
The trial, published in JAMA, compared the AI-powered program with human coaching over a year. Results showed that 31.7% of participants referred to the AI-powered program and 31.9% referred to human coaching met CDC benchmarks for diabetes risk reduction at 12 months. This outcome is particularly promising, as it indicates that AI-based interventions can be just as effective as human coaching in reducing diabetes risk.
One of the key strengths of the AI program was its high rates of initiation and completion. According to the study, 93.4% of participants referred to the AI program started the program, and 63.9% completed it, compared to 82.7% and 50.3% for traditional programs. This suggests that AI-based interventions can overcome some of the barriers to entry and completion that traditional programs often face.
The study's design was pragmatic and noninferiority, conducted from October 2021 to December 2024, with 368 middle-aged adults with prediabetes and overweight or obesity. The participants were randomly assigned to receive referral to either a fully automated AI-powered DPP mobile app or one of four CDC-recognized human coach-led programs. The research team followed up at 6 and 12 months to assess real-world effectiveness.
The AI-powered program utilized personalized push notifications for weight management, physical activity, and nutrition, along with active and passive data collection. In contrast, the human-led programs involved CDC PreventT2 curriculum, group video conferences, trained lifestyle coaches, and 16 weekly core sessions followed by biweekly to monthly maintenance sessions.
The primary composite outcome required maintaining HbA1c below 6.5% throughout the study and achieving at least one of the following: 5% weight loss, 4% weight loss combined with 150 minutes of weekly moderate-to-vigorous physical activity, or an absolute HbA1c reduction of 0.2 percentage points. The AI program demonstrated clear noninferiority, with a risk difference of -0.2% (one-sided 95% CI, -8.2%) falling within the noninferiority margin of -15%.
The study's high retention rate of 85.1% and alignment with expected community-based DPP outcomes further support the validity of the findings. The results suggest that AI-based interventions can provide reliable personalized interventions, making diabetes prevention more accessible and scalable.
However, the study also acknowledges some limitations, including the unmasked design, use of surrogate outcomes, and recruitment of motivated volunteers from only two sites. The authors emphasize the need for further research to explore the effectiveness of AI-based interventions in broader, underserved patient populations.
Despite these limitations, the study's findings have significant implications for clinical practice. With only 3% of US adults with prediabetes currently participating in DPPs, the potential for AI-powered interventions to scale up diabetes prevention is immense. The authors suggest that primary care providers may consider AI-led DPPs for patients in need of lifestyle change programs, especially those with logistical constraints.
In conclusion, this landmark trial demonstrates that AI-powered diabetes prevention programs can match the effectiveness of human coaching. As AI technology continues to evolve, these findings suggest that fully automated behavioral interventions may offer a viable complement to traditional human-coached programs, making diabetes prevention more accessible and scalable for a wider population.
