Nightingale Open Science
The absence of high-quality, open medical data sets remains one of the major obstacles to realizing the promise of AI in health. Students and faculty can spend months or years arranging data use agreements, or worse, give up altogether. Public data sets are often drawn from narrow settings such as the intensive care unit or from limited US populations. Models developed from data in one health system are deployed in others, even if differences in population or practice mean that they function very differently from place to place.
CPH–In partnership with the University of Chicago’s Center for Applied Artificial Intelligence (CAAI)–is helping to change that by hosting Nightingale Open Science, a large, global data repository containing more than 400,000 de-identified medical images and other data. The repository, co-founded by CPH and School of Public Health faculty member Ziad Obermeyer was collected from diverse populations worldwide, and is labeled with actual groundtruthed medical outcomes. Data is available to CPH researchers and students free of charge, and includes:
- Waveforms
- Microscopy Images
- X-ray Images
- Multiple Diagnostics
The collaboration with CAAI to make Nightingale data available advances CPH’s twin mission of delivering world-class academic training and real-world impact in hospitals, clinics and community, offering a resource for teaching, research, model development and biomarker discovery. Along with such efforts as Open Platforms for Health, an initiative to enable secure privacy preserving streaming of data from wearables and multiple other sources to advance biomarker development and deployment, Nightingale advances CPH’s shared commitment to helping transform the health tech and data ecosystem.
For more information on CPH or engagement with the Nightingale data set at UC Berkeley, please contact CPH at cph.info@berkeley.edu.
