Hacker News
UPenn Researchers Use AI to Analyze 400K Reddit Posts, Surfacing Unreported GLP-1 Side Effects
University of Pennsylvania researchers applied AI analysis to 400,000 Reddit posts about GLP-1 medications (Ozempic, Wegovy), surfacing side effects not captured in clinical trial data or FDA adverse event reports. The study demonstrates a practical pipeline for using LLMs as epidemiological tools — mining patient-reported outcomes from social media at scale. 24 points on HN. For builders: this is a repeatable pattern for any domain where user-generated content contains signal that structured databases miss.
Source
↳ Follow the thread