Trends in Glucagon-Like Peptide-1 Receptor Agonist Social Media Posts Using Artificial Intelligence

JACC Adv. 2024 Aug 28;3(9):101182. doi: 10.1016/j.jacadv.2024.101182. eCollection 2024 Sep.

Abstract

Background: Glucagon-like peptide-1 receptor agonists (GLP-1RAs) have surged in popularity in recent years, with discussions about their on-label and off-label use spilling into the public forum. No study has analyzed online discussions about GLP-1RAs.

Objectives: The purpose of this study was to analyze perceptions of GLP-1RAs on social media.

Methods: We analyzed GLP-1RA-related posts on Reddit between May 28, 2013, and June 1, 2023. All posts were identified that included generic or brand names of GLP-1RAs. Post volume on Reddit was compared to search interest on Google over time. An artificial intelligence (AI) pipeline consisting of a semi-supervised natural language processing model (Bidirectional Encoder Representations from Transformers [BERT]), a dimensionality reduction technique, and a clustering algorithm was used to cluster posts into related topics. Discussion sentiment was classified using a pretrained BERT model and assessed qualitatively.

Results: 14,390 GLP-1RA-related Reddit posts by 8,412 authors were identified. Ninety-four percent of posts were created after 2021, consistent with search interest trend on Google. We used the AI model to categorize posts into 30 topics which were hierarchically grouped by the model based on shared content. Posts were identified among communities for individuals with diabetes and obesity, as well as for diseases without a Food and Drug Administration-approved indication. Most posts had a negative sentiment using the pretrained model, acknowledging the pretrained model is at risk for misclassifying posts.

Conclusions: AI can generate insights on perceptions of GLP-1RAs on social media. Common themes included success stories of improving diabetes and obesity management, struggles with insurance coverage, and questions regarding diet, side effects, and medication administration.

Keywords: GLP-1RA; artificial intelligence; natural language processing; ozempic; social media.