The journal Clinical Nutrition ESPEN recently published the results of a trial comparing physician-guided dietary supplement recommendations for weight loss to artificial intelligence-guided supplement recommendations. This 6-month-long pilot study randomized 60 participants between the ages of 40 and 60 with a body mass index (BMI) of ≥ 25 kg/m2 to one of these two groups, with the primary endpoints of percent weight loss and the number of participants achieving a 5% or greater weight loss. The artificial intelligence (AI) recommendations were individualized by each participant’s genetic, metabolic, and behavioral data.
The mean loss in the AI group was 12.3%, significantly greater than the 7.2% loss observed in the physician-guided group. Additionally, over 83% in the AI group achieved a 5% or greater weight loss, significantly higher than the approximately 53% achieving this in the physician group. Significant differences favoring AI were also observed for fat mass and visceral-fat ratings, postprandial appetite suppression, hunger, satiety, etc. Adverse events were reported to be mild to moderate, with a higher incidence of GI symptoms in the AI group, but they did not affect adherence.
A study previously published by this same group had compared AI-guided to physician-guided supplement recommendations for LDL-cholesterol reduction. In that trial, LDL-C dropped by 25.3% in the AI group, significantly more than the 15.2% observed in the physician-guided group (and no difference in adverse effects). In that trial, the AI-group received more supplements on average (3.5 vs. 2.1), as well as a larger variety, including plant sterols, omega-3 fatty acids, red yeast rice, coenzyme Q10, niacin, and fiber, in contrast to the physician group which largely prescribed niacin and omega-3 fatty acids. Participants with specific genetic variants (in LDLR, APOB, and PCSK9, for example) experienced larger benefits in the AI-guided group.