AI analysis of Reddit reveals public interest in GLP-1 drugs for weight loss and mental health benefits

In a recent study published in the journal Communications Medicine, researchers in the United States of America (US) used large language models (LLMs) to analyze over 391,000 unique discussions on Reddit, a social media platform, related to glucagon-like peptide-1 (GLP-1) receptor agonists (GLP-1 RAs). The study revealed high interest in GLP-1 RAs, with discussions focusing on weight loss experiences, side effects, access issues, and positive psychological benefits, with mostly neutral-to-positive sentiment.

Study: Using large language models to assess public perceptions around glucagon-like peptide-1 receptor agonists on social media. Image Credit: Caroline Ruda/  Shutterstock

Background

More than 38% of the world’s population is overweight or obese, projected to reach 51% by 2035. Obesity is known to increase the risk of cardiometabolic diseases and all-cause mortality significantly. GLP-1 RAs are drugs that imitate the function of natural GLP-1, a hormone in the intestine that controls glucose metabolism and feelings of fullness. While this class of drugs was initially approved for type 2 diabetes, it has recently gained global attention for cardiovascular risk reduction and weight loss in patients with obesity, irrespective of the presence of diabetes. However, public views on GLP-1 RAs, crucial for treatment uptake and adherence, have not been thoroughly explored.

Social media platforms like Reddit offer anonymized public conversations on health topics, revealing real-world experiences often missed in clinical settings or trials. While manually analyzing large volumes of this data is resource-intensive, its analysis can be expedited using artificial intelligence systems such as LLMs. Therefore, researchers in the present study employed LLMs to analyze over 390,000 Reddit discussions on GLP-1 RAs, identifying topics such as weight loss, side effects, and concerns. They aimed to monitor side effects, gauge public sentiment, and guide future research and public health initiatives using the findings.

About the study

Reddit hosts user discussions in the form of posts and comments. It is organized into publicly accessible, topic-specific communities called “subreddits.” GLP-1 RA-related discussions were curated by indexing Reddit content based on the generic and brand names of GLP-1 RA drugs, including semaglutide. The dataset included 391,461 unique discussions (mostly since 2021) from 116,216 authors, with 71,982 posts and 319,479 comments.

A previously described “topic modeling” approach was employed for analysis, and various tools and algorithms were employed. Discussions are transformed into numerical representations and clustered to identify topics. Each topic was labeled and grouped based on similarities in discussion content. This approach aimed to extract key themes and insights from the extensive discussions on GLP-1 RAs available on Reddit. Additionally, this study employed a model named “RoBERTa” (short for Robustly Optimized Bidirectional Encoder Representations from Transformers Pre-training Approach) to classify sentiment. It used three probabilities (ranging from 0 to 1) to determine the nature of the sentiment within the text, classified as “negative, neutral, or positive sentiment.”

Results and discussion

About 97.1% of the discussions focused on GLP-1 RA medications prescribed for weight loss, such as semaglutide, tirzepatide, and liraglutide, with “Ozempic” being the most discussed (41.4%), despite not being approved by the US Food and Drug Administration (FDA) for weight loss. Only 2.9% of discussions were about GLP-1 RAs approved solely for diabetes. The volume of discussions surged significantly after 2022, following the FDA approval of “Wegovy.”

The model identified 168 discussion topics, indicating high public interest, with a focus on experiences with the drugs for weight loss. The topics included drug efficacy, comparison to other treatments, appetite impact, and side effects. Nausea was found to be the most frequent side effect, followed by vomiting, injection site issues, constipation, pancreatitis, and gastroparesis. Further, access issues, market shortages, insurance coverage, and the ethics of off-label use were found to be discussed. Positive effects on motivation and mental health and the value of avoiding bariatric surgery were also discussed. Topics were clustered into 33 groups, reflecting themes such as comparisons with other treatments, side effects, access concerns, and psychological benefits. Sentiment analysis revealed that 31.8% of discussions were negative, 50.1% were neutral, and 17.4% were positive. Notably, two topics were excluded due to illegal content related to acquiring illicit substances.

Scatter plot showing a 2D-projection of all discussion embeddings, where each point represents a discussion. The overlying color represents the associated group of that discussion based on the topic modeling. The x- and y-axes represent the two axes (Feature 1, Feature 2) onto which embeddings were dimensionally reduced using Uniform Manifold Approximation and Projection for visualization purposes.

The study is strengthened by its large-scale AI-based analysis of social media discussions to uncover public perceptions and experiences with drugs, offering insights beyond traditional clinical research. However, the study is limited by potential mislabeling due to spelling errors, inability to verify reported side effects, limited generalizability, and suboptimal general task benchmarks for LLM.

Conclusion

In conclusion, the study analyzed large-scale GLP-1 RA-related discussions on social media using LLMs. The findings reveal discussions centered on weight loss experiences, side effect comparisons, access issues, and positive psychological benefits. This indicates high public interest in GLP-1 RAs and highlights priorities for clinical and policy communities, including monitoring side effects, addressing access barriers, and acknowledging both the physical and psychological benefits of these drugs.