We have a research letter published in JAMA Network Open that we're excited to have out there. We used LLMs to study outcome switching in clinical trials, and found (perhaps) surprising amounts of it. Quick thread on 1) What is outcome switching? 2) why is it bad? 3) What did we do to detect it?
jamanetwork.com
Monitoring Changes in Clinical Trial Primary Outcomes Using Large Language Models
This cross-sectional study investigates the prevalence of meaningful changes to the outcomes in preregistered clinical trials using large language models.
