The privacy risk of patients sharing their deidentified data for medical research is low, an Oct. 6 study in PLOS Digital Health found.
Scientists use this type of data, which has been stripped of personal information, to develop artificial intelligence-based algorithms to diagnose diseases and predict their onset or progression.
Between September 2016 and September 2021, there were no instances of deidentified data being reidentified, according to the review of more than 10,000 U.S. media publications. During the same period, more than 100 million health records were stolen in data breaches from supposedly secure systems.
"Of course, it's good to be concerned about patient privacy and the risk of reidentification, but that risk, although it’s not zero, is minuscule compared to the issue of cybersecurity," said senior study author Dr. Leo Anthony Celi, a principal research scientist at MIT, in an Oct. 6 MIT news release. He is also an instructor at Harvard T.H. Chan School of Public Health and a physician at Beth Israel Deaconess Medical Center, both in Boston.
Dr. Celi said any risks of sharing the data are outweighed by the benefits to medical advancement and public health. He said studying deidentified data can help reduce health disparities by increasing the representation of traditionally marginalized minority groups. He added that better safeguarding the data sets is a better option than asking for patient consent, as that could exacerbate inequities by further excluding those communities.