Mary Mosquera reports on a privacy issue that I’ve discussed a number of times: that use of allegedly “de-identified” patient information risks re-identification and all that goes with it.
The Office of the National Coordinator has begun a study on how to overcome the privacy and security risks of using health information that otherwise has been stripped of its personal identifiers such as a patient’s name and address.
One danger is that the data might be able to be re-identified through the use of additional records publicly available on the Internet, according to Dr. David Blumenthal, the national health IT coordinator.
But health care experts say that the use of de-identified data is critical for tracking population health over time and for research purposes.
ONC wants to come to a “consensus on what risk we can tolerate for identification and then what level of removal, what kinds of removal of information, are required to get to that level of risk,” Blumenthal told a congressional hearing Sept. 30. But he did not provide further details.
[…]
Blumenthal said that ONC will analyze the science of de-identification and re-identification of data in its study. Determining an acceptable level of risk in order to be able to use the data will not be a one-time judgment call, he said.
“That will require that we continually look at the Internet and the information that is available. It’s a judgment that we will continually make based on how the technology advances. But it is something that we recognize is critical to ensuring public trust and enabling some of the most valuable uses of information to go forward,” Blumenthal said.
Read more on Government Health IT.
While there are certainly “research” studies that are of public health importance, as we saw with Texas, some “research” is much more commercial in nature. Those engaged in such commercial uses would likely argue that their commercial research serves important healthcare benefits such as indicating trends or identifying unmet needs, but where is the line between such applications and marketing?
Keeping an eye on the Internet and new information that becomes available, as suggested by Dr. Blumenthal, is important, but once the data are “out there,” there is no getting them back. Even if we think patients’ identities are safe for now, if we release their data and then find that they are no longer safe next year, then what?
I hope the ONC works closely with the researchers like Dr. Sweeney and Professor Paul Ohm who have much to offer in this discussion.