Claudia Pagliari
LinkedIn.com Post
Originally posted March 26, 2019
There are several fundamental truths that those of us working at the intersection of data science, ethics and medical research have recognised for some time. Firstly that 'anonymised’ and ‘pseudonymised' data can potentially be re-identified through the convergence of related variables, coupled with clever inference methods (although this is by no means easy). Secondly that genetic data is not just about individuals but also about families and generations, past and future. Thirdly, as we enter an increasingly digitized society where transactional, personal and behavioural data from public bodies, businesses, social media, mobile devices and IoT are potentially linkable, the capacity of data to tell meaningful stories about us is becoming constrained only by the questions we ask and the tools we are able to deploy to get the answers. Some would say that privacy is an outdated concept, and control and transparency are the new by-words. Others either disagree or are increasingly confused and disenfranchised.
Some of the quotes from the top brass of Iceland’s DeCODE Genetics, appearing in today’s BBC’s News, neatly illustrate why we need to remain vigilant to the ethical dilemmas presented by the use of data sciences for personalised medicine. For those of you who are not aware, this company has been at the centre of innovation in population genomics since its inception in the 1990s and overcame a state outcry over privacy and consent, which led to its temporary bankruptcy, before rising phoenix-like from the ashes. The fact that its work has been able to continue in an era of increasing privacy legislation and regulation shows just how far the promise of personalized medicine has skewed the policy narrative and the business agenda in recent years. What is great about Iceland, in terms of medical research, is that it is a relatively small country that has been subjected to historically low levels of immigration and has a unique family naming system and good national record keeping, which means that the pedigree of most of its citizens is easy to trace. This makes it an ideal Petri dish for genetic researchers. And here’s where the rub is. In short, by fully genotyping only 10,000 people from this small country, with its relatively stable gene pool, and integrating this with data on their family trees - and doubtless a whole heap of questionnaires and medical records - the company has, with the consent of a few, effectively seized the data of the "entire population".
The info is here.