The benefits to the public of sharing their health data must be articulated more compellingly if their trust is to be won, writes Harry Evans
Yesterday Dame Fiona Caldicott launched her Review of Data Security, Consent and Opt-outs which sets out a way forward for the collection of patients’ health data.
At Ipsos MORI, we have produced substantial pieces of research on this topic with the Royal Statistical Society, the Wellcome Trust and, most recently, a piece of work speaking to the public about government use of data science with the Government Digital Service.
Much of the research focus has been on giving the public a voice in the data debate, and discovering the underlying nuances in their views on this important topic. As a result, Dame Fiona Caldicott’s recommendations for a wider dialogue in this area are welcomed, as is the consultative nature of the review itself.
A great deal of focus in the review is on the details of what an opt-out should look like. We know from our research just how important permission is to people when it comes to very personal and sensitive health information. In our work on Commercial Access to Health Data with the Wellcome Trust, people spoke from the heart about how they felt they should be consulted when their data was being used. We found that this attitude is widespread, with more than half of the public wanting to be asked for their permission if their health data is to be shared with commercial organisations.
Intrusions on privacy
The review also provides an opt-out exception for anonymised data and this reflects a certain softening of public concerns when it comes to anonymised data. However, many still feel like anonymised data looks and feels like their data, even if they also acknowledge that anonymised data is safer than data with their name attached.
To some extent, as Dame Fiona Caldicott says, desire for consent is related to the low trust the public have in organisations using their data. Our research with the Royal Statistical Society from 2014 demonstrates that there is a ‘data trust deficit’ across sectors, with trust in organisations to handle data lower than trust in general. This crisis in confidence is a huge issue and to some extent is not the fault of any one sector.
Those with low awareness of what the NHS is using their health data for tend to be more likely to oppose commercial access to health data for the purposes of research
When discussing this with patients, they frequently refer to cold calling, spam and other intrusions on their privacy that they associate with any number of organisations giving their data away freely. Low trust is partly related to these bad experiences, but also a low awareness of what actually happens. Just a third of people admit to knowing a great deal or a fair amount about what the NHS does with their data.
And those with low awareness of what the NHS is using their health data for tend to be more likely to oppose commercial access to health data for the purposes of research. This is hardly surprising, but awareness of sharing leads consistently to a higher support for health data sharing.
Consent is often a way of taking back control, and people often express this desire as a result of not having information that answers key questions, such as how is this data to be protected? What are the penalties for misuse and how regularly are these enforced? Who could be using that data and why?
Answers to these questions are often missing from the public’s understanding of how data sharing works and this leads many people to rely on misconceptions and their biases about what happens. Where uncertainty is introduced into the mix is where concern starts to manifest itself and when having control of one’s data becomes increasingly important.
And this is key – providing clear and easily accessible information about data sharing is important. This is why a structured conversation with the public now needs to take place to communicate the benefits of health data sharing, but also to listen to their real concerns. Many people feel let down around their data being shared, and it will take more now than opt-outs to restore these people’s faith in the way the NHS handles their health data.
Risks and safeguards
Frequently missing from the public’s understanding of data sharing is the positive case for why policy makers are so interested in the wholesale collection of health data. Three in five members of the public support commercial organisations having access to health data when research would be at risk without that access. We know that support is much higher than this for the health sector using that data internally, such as for direct care.
The benefits, however, are not regularly articulated, partly due to many of the benefits not having been realised. It is vital, though, that this is done, in conjunction with education about the risks and safeguards. Clear and transparent examples of data usage are currently missing from the public’s examples of data sharing – mostly the public just talk about the negative experiences they have had and this is what is leading to scepticism.
Google DeepMind, following negative press about their data sharing with the Royal Free, have just published plans to work with anonymised patient data from Moorfields Eye Hospital in order to scan for signs of disease. While concerns will remain about commercial organisations receiving this kind of sensitive health data, the transparency displayed here is a step towards making a positive case for data sharing and may provide answers to concerns we have heard directly from the public.
Our research on health data suggests that proactive approaches that aim to build awareness of the uses of health data are key in reassuring patients and the public that there are positives to sharing. There are many real concerns about what health data can be used for, and this is why providing an informative opt-out model is an important first step, but conversation about further concerns, as well as potential benefits, are the vital next step.
Harry Evans is senior research executive at IPSOS Mori.