The use of Artificial Intelligence in analysing statin-related sentiments on social media – the perception is mostly neutral, or negative…

By Dr. Ahmed El-Medany

Despite their efficacy and safety profile, statins are underused, even in patients with the highest cardiovascular disease risk. Social media platforms such as Reddit may offer valuable insights into public perceptions and attitudes towards statins.

A recent qualitative study published in JAMA Network Open by Somani et al. analysed all statin-related discussions on Reddit between January 2009 and July 2022. The study identified statin- and cholesterol-focused communities and developed an artificial intelligence pipeline to cluster discussions into specific topics and thematic groups. The pipeline included a natural language processing model, a dimensionality reduction technique, and a clustering algorithm. Sentiment analysis was used to label discussions as positive, neutral, or negative.

The study identified 10,233 unique statin-related discussions from 5,188 authors. The number of discussions increased each year. The discussions were classified into six thematic groups: (1) ketogenic diets, diabetes, supplements, and statins; (2) statin adverse effects; (3) statin hesitancy; (4) clinical trial appraisals; (5) pharmaceutical industry bias and statins; and (6) red yeast rice and statins. The majority of the discussions had a neutral (66.6%) or negative (30.8%) sentiment.

This study highlights the potential of artificial intelligence to analyse large amounts of social media data and provide insights into public perceptions of medication pertinent to Cardiovascular medicine. The information obtained could help guide strategies for addressing barriers to evidence-based therapy use and adherence.

Read the full study here: https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2803988