A "huge loophole" is being carved in the European Union's upcoming data protection regulation, according Ross Anderson, a professor of security engineering at the University of Cambridge in England.
The way the current draft of the law allows secondary uses of medical records is a privacy scandal waiting to happen, Anderson said during the Amsterdam Privacy Conference on Monday.
The loophole, he said, is in articles 81 and 83 of the draft data protection regulation, which covers the way personal health data may be processed in the 27 E.U. member states. The data can, for instance, be used for historical, statistical and scientific research purposes -- but the way the draft proposal is crafted poses a big privacy risk, Anderson said.
"The fundamental problem is that everyone from insurers to drug companies wants access to masses of personal data," said Anderson. And if Europeans are not careful they could reveal very personal information to both academic and commercial researchers, he said.
The data could, for instance, be used to set up patient registries for improving diagnoses, differentiating between similar types of diseases, and preparing studies for therapies. The same information could also be useful for pharmacists who might want to adjust their prescriptions, said Anderson.
There is a big tussle coming up about those secondary uses, he said.
While the draft regulation calls for the data made available to researchers to be made anonymous, it is very hard to anonymize data sufficiently, Anderson said.
"That's something that we geeks have known for thirty years," Anderson told the audience.
By querying a database often and asking it different questions it is possible to find out where someone lives, or to find the medical record of a colleague or friend.
"You can always find a set of queries that reveals the target," he said.
When it comes to making data anonymous, Anderson said, "The big practical problem is context." For example, he said he could find out more about a particular colleague by querying the database for 42-year-old women in the U.K. with a 9-year-old daughter, who both have psoriasis.
Anderson had hoped that the new data protection regulation would fix a loophole in existing U.K. legislation that was created by defining "personal data" too narrowly. However, the same is now happening with the new E.U. regulation, he said.
One of Anderson's colleagues, Douwe Korff, a professor of international law at London Metropolitan University, has already proposed amending articles 81 and 83 of the regulation in a way that would still allow member states to create exemptions for health care and research, but only for "exceptionally high public interests," with the highest standards of anonymization and with the prior agreement of the privacy regulator.
Anderson, though, wants to go further: he proposes notifying people in advance that their data will be used for research or other purposes, allowing them to opt out.
"Most people don't want wide sharing of information and most people don't want research without consent," he said.
If the articles are not amended, Anderson said, he could picture a situation in which a professor of psychiatry conducting research might lose a file containing 10 million patient records -- with them subsequently being posted to Pastebin, a site often used by hackers to leak stolen personal data. Such losses have already happened, said Anderson, referring to a case in which a laptop containing the health records for almost 8 million Britons was stolen.
And if such a database were posted to Pastebin, he said, then it's also easy to imagine someone comparing the data to what's known about the U.K.'s members of parliament. "And then things blow up," he said, "a scandal leading to public revolt."
While this scenario is of course fictional, it should be anticipated, Anderson said.
"I want to push back on what the E.U. is trying to do here," he said. "If civil society can get its voice heard there is every chance we can get these sections amended."