Main content start
Amit Kaushal and Russ Altman: JAMA paper published on clinical deep learning dataset
Clinical datasets for deep learning come from three states: CA, NY, MA
Stanford Bioengineering Adjunct Professor Amit Kaushal and Professor Russ Altman have published a JAMA paper on clinical deep learning datasets. The paper is titled "Geographic Distribution of US Cohorts Used to Train Deep Learning Algorithms."
The main finding is that across many medical disciplines (radiology, path, ophthalmology, gastrointestinal, cardiology, and dermatology) most US datasets used to train clinical-grade deep learning algorithms for diagnostic imaging come from California, New York, and Massachusetts, with little or no representation from the other 47 states.
Read more:
- "The Geographic Bias in Medical AI Tools" (HAI, Sept 21, 2020)
- "Medical AI systems are disproportionately built with data from just three states, new research finds" (Stat, Sept 25, 2020)
-
"Algorithms used in medicine are trained on data from only a few states" (The Verge, Sept. 25, 2020)