08 Aug UW-Madison gets $1.6 million to develop medical privacy safeguards
Madison, Wis. – The National Science Foundation has awarded $1.6 million in federal research funding to UW-Madison’s Computer Sciences department for work on better data encryption and privacy safeguards in public health information networks.
The three-year grant is for research on data mining strategies that could still preserve privacy. The Division of Public Health in Wisconsin’s Department of Health and Family Services and the Public Health Information Network system built by DoIT will partner with UW-Madison’s Computer Sciences department to serve as the test bed for researching the strategies.
Federal regulations controlling the release of detailed data to researchers currently restrict access to many important datasets and databases. A consequence is the impediment of basic research in fields dependent on such datasets.
“De-identifying data by doing obvious things like removing Social Security number and name and address is insufficient,” said UW-Madison computer science professor David DeWitt, “because it’s possible to take the de-identified data … and combine it with publicly available data to re-identify individuals.”
A primary focus of the project is to develop automated techniques and tools that are capable of taking a very large database containing tens of millions of records and shielding the identities of the people within the databases, facilitating the release of datasets to researchers while still ensuring that the privacy of the people in those datasets is maintained.
Key methods of doing that, according to DeWitt, involve techniques of generalization — expressing data within ranges rather than specifics. One example would be giving information such as dates of birth and ages within certain ranges rather than giving specific dates.
Researchers could then more easily collect demographic data about people with diseases such as HIV, for example, without violating the privacy of the patients, he said.