As AI becomes increasingly involved in healthcare decisions, transparency and reliability are critical to its success.
CUSP’s Dr. Sherry Yun Wang, winner of Women in AI’s 2025 Researcher of the Year Award, continues to drive progress in this space. Wang has published research in leading journals including JAMA, The Lancet, and Neuroepidemiology.
Wang’s most recent study found that NarxCare, a proprietary opioid risk algorithm used in over 20 states, could not be independently reproduced using its own publicly disclosed methods. npj Digital Medicine, part of the Nature portfolio, published the finding. The journal has an impact factor of 18.0 and ranks #1 among 194 journals in Health Care Sciences & Services. Wang’s team attempted to recreate the algorithm’s functioning using NarxCare’s own data alongside real-world data sets, but could not match its outputs.
The results raise real questions about accountability for a tool that actively shapes medication decisions for patients across the country. Wang’s research focuses on developing trustworthy, interpretable healthcare AI systems using large-scale real-world data. At CUSP, her work reflects a broader commitment to holding healthcare AI to the same standard of rigor as the medicine it supports; a standard that makes clear why algorithms used to guide patient care need to be verifiable.
This article was co-written by Silas Fernandes.
