Machine Learning & Big Data: Is it possible to share de-identified data?

Machine Learning & Big Data: Is it possible to share de-identified data?

The 2019 JP Morgan Healthcare conference, along with the many others that congregate in San Francisco at the same time, draws a huge audience from across the healthcare spectrum, discussing any number of topics ranging from industry trends, intellectual property topics, to deal making discussions. An interesting discussion this year was around the use of previously collected clinical data by for-profit companies with artificial intelligence, machine learning platforms.

As AI and machine learning platforms become bigger players, these companies are interested in accessing de-identified patient pathology slides and clinical data to reanalyze through their models with the hope of identifying a predictor that will allow for better treatment of patients in the future.  Although the companies do not want exclusive rights to the dta, do not want the identity of the patient disclosed, will not harm or alter or even take posession of the data, issues remain with allowing access to such data. The American Cancer Society is one of the group struggling with this, as they have tremendous amounts of archived data from studies they have previously funded.

Doctors and health systems collect massive amounts of data on each patient. It starts with demographic information, which is often confirmed more than once, then vitals, the reason for the visit, and lab results. This is just be for a typical visit to your PCP, not even data that a specialist might need. While this data, when analyzed by AI platforms, could have the potential to impact the diagnosis and treatment of others, there are many things to consider. We collect this data for care, but never imagined the potential for repurposed use. So how do we create a process that is safe, ensures privacy and protection, and is equitable and legally sound?

By: Tony Giordano, PhD, Innovation Manager
Tony Giordano joined Cleveland Clinic Innovations as an Innovation Manager in 2018.  Tony has extensive experience in commercialization and drug development, having served as Vice President or President of seven different biotechnology companies he co-founded, including companies developing platform technologies, a cancer vaccine, and Alzheimer’s Disease and cardiovascular therapies.  Tony also served as an Associate Professor and Assistant Dean of Research and Business Development at LSU Health Sciences Center in Shreveport, where he led the licensing efforts at the campus and at Abbott Labs, where in addition to serving as a Senior Research Scientist, he was involved in technology assessment activities.  Tony has a PhD from Ohio State University and completed Fellowships at the NCI and NIA.

Get In Touch With Us