Artificial intelligence and natural language processing of patient narratives to evaluate semaglutide for weight loss.

PURPOSE: This study used artificial intelligence (AI) and natural language processing (NLP) to analyze patient reviews of semaglutide, with the goal of better understanding its real-world effectiveness and safety for weight management. METHODS: A retrospective, cross-sectional analysis was conducted on 772 user-generated reviews of semaglutide posted on Drugs.com (July 2021-March 2025). Sentiment analysis was performed using a transformer-based BERT model on a five-point scale. Topic modeling with Latent Dirichlet Allocation (LDA) and Latent Semantic Analysis (LSA) was used to identify dominant themes. Cluster analysis was applied to segment users based on weight loss outcomes and side effect severity. Reviewers (n = 95) that explicitly reporting both weight loss and treatment duration were analyzed for real-world efficacy and safety. RESULTS: Users who took semaglutide more than 60 days reported a mean weight loss of 32.2 ± 3.1 lbs (14.6 kg). Frequently mentioned side effects included nausea (46.9 %), headache (18.4 %), vomiting (14.3 %), fatigue (9.2 %), and dizziness (4.8 %). The highest sentiment scores were observed in the ≤ 30-day group (mean: 3.38). Topic modeling identified themes such as appetite suppression, medication cost and access, and long-term experiences. Clusters analysis revealed distinct user profile, including super-responder group with substantial weight loss and another with more side effects. CONCLUSIONS: AI and NLP methods offer valuable tools for analyzing patient-reported outcomes, revealing semaglutide's real-world efficacy and safety profile for weight management. These findings contribute to ongoing efforts to integrate patient-reported data into post-marketing surveillance and treatment decision-making.