Health and technology: how can data save lives?
The matter was discussed by a panel including Diane Abbott and Chi Onwurah at The Labour Party's 2016 Conference
This article was produced in association with BCS
“I’m a huge believer in the progressive power of technology,” said Chi Onwurah, shadow minister for the digital economy, opening the Prospect/BCS debate on the role of technology in healthcare at the Labour Party Conference in Liverpool.
Championing the power of shared data not only reduces the cost of public services, she said, but it can make those services more responsive, more personal and can empower citizens. “The old model of a remote, unresponsive state can be transformed into a personalised, empowering state putting people in the driving seat.”
Inevitably, perhaps, there was a condition to her enthusiasm. “People must feel they are in control,” Onwurah said invoking the case of Care.data, a scheme led by NHS England intended to bring together health and social care information from sources including GP practices, hospitals and care homes. Security issues meant the scheme was first postponed then quietly dropped.
“With Care.data, people didn’t believe they were in control—and they were right,” Onwurah, MP for Newcastle Central, said. Cases like this and evidence of privacy and security breaches undermine confidence. “[The danger is] we’ll end up in a regressive world where people are paranoid and feel persecuted by the state. We need to be sensitive to that.”
More on those concerns later but first, can data really save lives?
Mike Fisher is chief clinical information officer at Royal Liverpool and Broadgreen University Hospital NHS Trust. He is a cardiologist and deals with data—a combination of individual and trials data—in his day-to-day clinical practice. It allows him to do his job. But, he said: “Data doesn’t save lives. And certainly machines and computers don’t save lives. Data has to be processed into knowledge before it can save lives.”
Processing and extracting value often means triangulating different sources of data and identifying correlation. This means data sharing. Fisher pointed to innovative approaches to data processing including IBM’s Watson project and Google’s artificial intelligence programmes led by its London-based operation, DeepMind. Google founder Larry Page has said: “Right now we don’t data-mine healthcare data. If we did we’d probably save 100,000 lives next year.”
“We have a traditional model in medicine founded in the 18th and early 19th century when the things that afflicted us were almost always highly infectious diseases,” said Fisher. “The truth is the vast majority of illnesses that I see … are nothing to do with that model any more. They are chronic diseases and long term conditions.”
“People with that sort of condition do not do well in hospitals. We actually make people sick a lot of the time. If we are genuinely going to do something to address the health of the nation, we have got to stop herding people into these disease palaces.” Instead, said Fisher, we need to look after people in more “distributed care settings”—at home, a modern version of a cottage hospital, a GP surgery with specialist input and so on. “Data and technology enables that process in a way that nothing that has gone before ever has,” he said. “That should be the focus.”
Andy Kinnear describes himself as an “NHS life” having spent 26 years in digital roles within the health service. Currently director of digital transformation at NHS South and head of health at the BCS, Kinnear accepts Fisher’s distinction between the collection and application of information. “Data per se doesn’t save lives. It’s the way we process and utilise it, turn it into knowledge and take decisions on the back of it. Fundamentally clinicians day in, day out are making risk-based decisions and they are making those based on the information they have in front of them. The more information we can provide, the likelihood is that they will make better decisions.”
Why then has it taken the NHS so long to make best use data? Kinnear offered three reasons. The first: technical. “Until we get to a place where all of our data is digitised—collected electronically and in the kind of formats that allow us to share it across the system—then it’s very difficult to leverage the value of that data.”
Second, the way the NHS is organised—hospital trusts, clinical commissioning groups, federations of GPs, GP practices, social care departments, ambulance trusts, “a myriad of individual organisations operating under that NHS umbrella”—militates against sharing. Data is compiled in the interest of one organisation and rarely shared beyond its walls, said Kinnear.
The third reason is the failure to “empower” the public. Unlike in the commercial sector where individuals are taking control of banking, travel and social data, “We’ve not got to the place in health where it is possible for people to truly own and control their data in a way that will transcend all those organisational boundaries.”
Despite these barriers, Kinnear insisted there were many examples of successful data use, especially at local level. He cited a Bristol-based project called Connecting Care that he was behind. It joins up patient information so a doctor can look at one set of data that includes GP, hospital and mental health data. He cited the example of one out of hours doctor who says he makes different—better—decisions because of the newly available data. “Previously, he was largely flying blind,” said Kinnear.
Which returns the conversation to the controversial issue of data sharing. There are two fundamental objections to data sharing—access and identification.
On access, Diane Abbot raised the danger of “promiscuous data sharing” where a commercial company that offers health services as part of its business harvests that data for wider commercial ends. Fisher insisted that there were already regulations in place to deter such misuses of data and pointed out the distinction in place between primary and secondary use of data, “between sharing for the purpose of looking after you and sharing for another reason.” Fisher did acknowledge that sometimes “the borders between healthcare and secondary use can get a bit blurry. But for the purposes of direct healthcare, the vast majority of people have absolutely no problem about their healthcare data being shared.”
Fisher acknowledged, too, that although unlikely it is theoretically possible to identify previously unidentifiable individuals through the triangulation of data sets. This is something that concerns Onwurah. “The more data you correlate the easier triangulation and identification is, particularly with certain rare diseases. That’s why we need a data framework before we can begin to reap these huge rewards.”
“Being targeted with cough medicine ads because you went to the doctors last week would be annoying. Having your mortgage application refused because the combination of your medical records indicates [something problematic] would be traumatic.”
Onwurah said that what is needed is “a rigorous framework for handling patient data and a set of principles that cut across central and local government.” A review into information sharing in the NHS, originally set up in 1997, led to the drawing up of the Caldicott principles, named after Fiona Caldicott who chaired the review. The seven principles act as a code of practice and while Onwurah describes them as “good guidelines and targets” she believes “they are too vague, too open to interpretation and the vast majority of people have never heard of them.”
If the current principles are too vague what would a specific principle look like? Onwurah suggested the following: “’You own your own data. We won’t use it unless you tell us we can.’ That’s not vague. That’s very simple.”
The danger, of course, is that patients might choose, on balance, to refuse permission.
For those who believe data sharing is essential to enhance the well-being of the nation, allaying public fears and changing the tenor of the conversation is key. “There is already a digital dialogue with the public around the use of their data,” said Kinnear. “It can’t be beyond the wit of man to extend that into the health space.” Meanwhile, Fisher noted that, “If you ask people in hospitals, the vast majority just assume that data is shared between all people who are looking after them.”
So perhaps it not only depends how the question of data sharing is framed but when the question is asked. “I don’t think people have the information to make a judgement,” said Onwurah. “There is a reservoir of goodwill towards the NHS that we could build on… but too many bad headlines will turn people off.”
With the support of BCS, The Chartered Institute for IT, Prospect hosted a series of panel discussion at the 2016 Party Conference on health data. These discussions were chaired by Jon Bernstein, Associate Editor for Prospect. Speakers included: Chi Onwurah MP, Shadow Minister for Business, Energy and Industrial Strategy; Diane Abbott MP, Shadow Secretary of State for Health; Jo Churchill MP, Chair of the APPG on Personalised Medicine; Andy Kinnear, Chair of BCS Health; Dr Jeni Tennison, CEO, Open Data Institute (ODI); and Dr Mohammad Al-Ubaydli, Founder & CEO, Patients Know Best.
For speaker and partnership opportunities, please contact firstname.lastname@example.org.
We want to hear what you think about this article. Submit a letter to email@example.com