We risk harming the huge potential of mass medical data in health research unless public concerns are properly addressed
Digital data are being amassed at an astonishing rate. This vast store presents great opportunities for science. Health data – for example, from biomarkers, medical imaging and genome analysis – offer a rich resource. Combined with data about lifestyle, diet and living and working environments, they can yield new insights on the causes of disease and the effectiveness of treatments, help to predict and prevent illness and allow therapies to be personalised.
Recent advances in information technology and data science – such as faster computing and improved methods of analysis – make this increasingly possible and are already being used in projects that rely on volunteers, such as the UK Biobank, which hopes to track the lifestyle and health of 500,000 over-45s.
So why has the UK government’s giant care.data scheme, which aims to aid research and improve the health service by centralising data on millions of people in England obtained from family doctors, met with such resistance?
When security and privacy doubts were voiced over care.data last year, the project was postponed. It is an active opt-out scheme; unless an individual objects, their details are included. But the reaction of GPs, the media and civil society demonstrated that the data uses intended in the scheme were not in line with people’s expectations.
|
Advertisement: Replay Ad
 |
|
Ads by ZINC
|
<a href=”http://ad.doubleclick.net/N6831/jump/NewScientist/ns_unclassified;key=opinion+dn26901+nologin+Opinion+health+medicine+big-data+privacy+technology+nhs+healthdata+UK-Biobank+research+comment-analysis;tile=7;sz=450×250;ord=1234567890?”><img src=”http://ad.doubleclick.net/N6831/ad/NewScientist/ns_unclassified;key=opinion+dn26901+nologin+Opinion+health+medicine+big-data+privacy+technology+nhs+healthdata+UK-Biobank+research+comment-analysis;tile=7;sz=450×250;ord=1234567890?” /></a>
Data worries
People have concerns about data misuse – such as worries over insurance pricing, marketing, political surveillance and identity theft – but we do not yet have a clear understanding of the impact of such misuse. Research commissioned as part of the Nuffield Council on Bioethics’s new report on the use of big data in health care and biomedical research, led by the Wellcome Trust’s Expert Advisory Group on Data Access, shows that the harms are likely to be broad in range and more prevalent than those that are acknowledged by regulatory regimes. We know that a particular public concern is the sharing of health data with commercial researchers: for example, pharmaceutical or insurance companies.
Re-identification from “anonymised” data is increasingly possible as technology improves and more and more databases become linked. People are being asked to consent to data uses that have implications which neither they nor the researchers may fully understand. And simply complying with data law is not enough: it is just a framework that may not fully capture the reasonable expectations of research participants about how their data will be used.
To win public confidence, we need to go further. We need to start by understanding people’s interests in how their data will be used. Appropriate security measures and participant consent need to be framed within governance systems that both involve and are accountable to those whose data is held.
Such systems must take seriously people’s reasonable expectations about data use: about who may have access to the data, for what purposes and for how long. This may seem onerous, but there are examples of good practice to draw from as well as sobering lessons from previous cases.
These recommendations are in the latest report from the Nuffield Council on Bioethics. It offers a set of principles and some practical guidance to help those contemplating data initiatives such as care.data. Other conclusions include tracking how health data are used and ensuring individuals are told of any privacy breaches. Tough penalties for deliberate misuse of data, including imprisonment, should be introduced, even if it appears that no harm was caused.
Only in this way will we build the public involvement and trust that is necessary to ensure that the benefits of personal health data are not squandered.
Martin Richards is chair of the Nuffield Council on Bioethics Working Party for the report The Collection, Linking and Use of Data in Biomedical Research and Health Care: Ethical Issues. He is also emeritus professor of family research at the University of Cambridge