“We seem to spend all our time talking about the downside of making patient data available and little about the upside,” said a frustrated researcher at last week’s Sowerby eHealth Symposium organised by Imperial College’s Institute of Global Health Innovation. The problem seems to be that the downside—somebody’s health records being made public—is horrible, concrete, and easy to understand, whereas the upside remains vague, aspirational, and largely opaque to the public.
Everybody seemed to agree that care.data, last year’s attempt to make patient data available, was a disaster. Indeed, it reminded people of Connecting for Health, the attempt to wire the NHS, another top down disaster. Between them those two disasters may have set back by years attempts to move the NHS to a place where many other enterprises have long been, a place where everything happens online rather than on paper and where data are used constantly to improve the performance of the enterprise.
“Data,” said Lord Darzi, director of the Institute of Global Health Innovation, “are the most important enabler of the uptake of innovation,” allowing health services to integrate, improve safety, conduct research, and reap the full benefits of the “omic era.” (In case that phrase baffles you, Darzi meant genomics and metabolomics, which everybody hopes will lead to precision medicine where the uniqueness of each patient will be fully recognised.)
Equally upbeat was George Freeman, the world’s only minister for life sciences, who may soon, he told us, become minister for life and health sciences. The prime minister has given him two aims, to enable innovation that will mean that the NHS gets more health for every pound and to make the UK a world leader in health technology, which it can then sell and pay for the continuing development of the NHS. Patient data will be essential to achieve both aims. More fundamentally, wiring the NHS and making data available will, he said, move the NHS from a 20th century model where “things are done to you” to one where “you aspire to look after your health yourself.” He wants to encourage “active health citizenship.”
As a politician, he is, however, aware of the fallout from care.data and recognises the need to show benefits early. He’d like “an Uber moment for the NHS.” By this he presumably meant the public discovering in the NHS something equivalent to being able to get a taxi faster and cheaper than the usual service rather than taxi drivers (doctors) protesting about the competition.
East London has supposedly gone further than most areas with using data to improve services, but Charles Gutteridge, chief clinical information officer for Barts Health NHS Trust was unconvincing in his attempts to illustrate the improvements.
Paul Aylin, professor of epidemiology and public health at Imperial College, began his interest in using data for improvement by seeing if he and his team could use Hospital Episode Statistics to identify results for cardiothoracic surgery for children as an outlier. They could. Similarly when given mortality data from a thousand GPs they could identify Harold Shipman, the GP who murdered hundreds of his patients, as an outlier. Unfortunately they all also identified 11 other GPs: they presumably were not mass murderers but GPs who looked after nursing homes or had an interest in palliative care. These examples of the use of data were not convincing either as they were identifying problems after the event rather than before, but Aylin’s team did identify problems at Mid Staffordshire before they became widely known. Even here, however, simply listening to the terrible stories told by patients and their relatives should have been enough to identify the severe problems of the trust.
Leeds has managed to bring patient data together from hospitals and 104 of 109 general practices and make them available to clinicians for direct patient care, said Julie Oxley, head of IM&T for Leeds City Council. Clinicians, she said, greatly appreciated being able to have access to the data, and some patients have undoubtedly benefited—but again there were not simple stories that would convince the public, although only 44 of 673 000 patients have opted out of having their data available. (It wasn’t clear, however, that every patient had had a clear chance to opt out.)
So where are we with public trust in making data available? asked Nick Timmins, senior fellow at the King’s Fund in the final panel session.
In a bad place, answered Phil Booth, the coordinator of medConfidential, a body that campaigns for confidentiality and consent in health and social care. He campaigned against the cavalier attitude of care.data but emphasised that he recognised that there will be benefits from making patient data available–but it must be done in a way that is “consensual, safe, and transparent.” medConfidential made it possible for people to opt out of their data being available, and the projected figure is that around 900 000 to 1.6 million have done so. The data of those people are, however, being made available by the Health and Social Care Information Centre (HSCIC), and Booth thought that those people would be very angry if they knew so.
In broad terms patient data can be used for direct care, commissioning, research, and “commercial use.” The benefits to patients are most obvious with direct care, but even so, said Booth, about 1.5% of patients will opt out, fearful of their data being leaked, hacked, or made public in some way. Commissioning is not understood by the public (and many people involved in it), and the benefits from research are not yet convincing.
Many people seem to object to the commercial use of data. HSISC, “a creature of statute,” as Dave Roberts from HSISC repeatedly emphasised, does have individual patient data from hospitals and primary care and it can make anonymised data available to private companies who are working to improve health and social care. But data can also be made available to companies working to promote health, which can certainly mean pharmaceutical companies. Roberts emphasised that HSISC is being very cautious, and many researchers in the audience believe it to be far too cautious.
Booth was very blunt when he said “you can’t put out whole population data—it’s just too dangerous.” The problem is that by linking different datasets it can be relatively easy to identify individuals, particularly those with rare conditions. NHS England said in the debate over care.data that no patient had ever been identified from Hospital Episode Statistics, but a Freedom of Information Request showed that that was not true.
But, asked a researcher in the audience, isn’t the genie out of the bottle as more and more data become available and are shared? Booth didn’t accept that, nor did he accept Timmins’s analogy with flying: “we know some planes crash but we don’t ban flying.” The point was that airlines have done everything possible to prevent accidents but much more needs to be done to protect patient confidentiality. The only way to make potentially identifiable data available is to “safe” researchers in “safe” circumstances.
In an attempt to row back from the disaster of care.data the NHS is running pathfinder projects, where local teams will set up systems of sharing data and show the benefits. It’s a bottom up rather than a top down programme.
Most of those at the meeting—people from health and researchers–clearly agreed that there was too much emphasis on risks and not enough on benefits, but I left the meeting convinced that everybody involved—politicians, regulators, commissioners, clinicians, and researchers—needs to do a much better job of explaining benefits. And still more important is to show the benefits and let patients experience them.
Competing interest: RS is an adjunct professor at the Institute of Global Health Innovation and chair of Patients Know Best, a company that gives patients control of all their records from health and social care. The views expressed her are his own.
Richard Smith was the editor of The BMJ until 2004. He is now chair of the board of trustees of icddr,b [formerly International Centre for Diarrhoeal Disease Research, Bangladesh], and chair of the board of Patients Know Best. He is also a trustee of C3 Collaborating for Health.