Respecting privacy with responsible data sharing
04.09.14
There is a lot of excitement about Big Data, even if the term seems to mean different things to different people. In the advertising and online media world, different data sources are used to “triangulate” what persons to receive what adverts or what links to display. In my industry, the researchers are very excited about connecting patient health records with other databases to better understand different diseases, or to track “real-world” performance of new medicines: how safe and effective are they, and are they worth the money. These applications are also about “big data”.
When sharing data, and pooling data, there must be safeguards so that individual patients cannot be identified. Computer scientists have already shown that it is possible to identify individuals in databases that are supposedly “safe”. Not long ago there was a minor scandal in the US when some clever computer wizards managed to identify the health records of a state governor, despite the “anonymisation” of publicly available health records. Here I am not talking about hacking – just about linking personal details from health records, with other public records, Facebook, and data that can be bought from Google etc.
As we move to share clinical trial data, we need to make sure individual patients cannot be identified. So far the debate has been focusing on what data is commercially confidential. More important must be the privacy of patients. Several of my member companies take part in a data-sharing consortium called TransCelerate. It aims to speed up the development of new medicines. It is only one of the many “open innovation” projects that are under way. The Innovative Medicines Initiative is the biggest of the kind.
My colleagues in TransCelerate have just published standards to make sure patient privacy is guaranteed when data is shared. These not only illustrate the complexity of the issue, but also how serious we are about protecting patients. We are committed to share much more data with the public and other researchers, but it has to be done responsibly.
