PassW0rd – 8 September 2021

Speaker A: This program is brought to you by Resonance FM. If you like what you hear, please support our work by making a donation at resonancefm.com/donate.

Speaker B: Hello and welcome to Password with me, Peter Warren. In today’s program, we look at the controversial world of health data. Why on September 1st the government pulled back from its announcement that it would seize control of all of the UK’s health data and use it as it saw fit. And why 5 days later Google announced it would be shutting down its AI-driven health data analysis operation streams. So what What is it that has made health data such an explosive issue?

Speaker C: For me, the issue is about how data is harvested, how data is managed, and how much data about me and my life is released into the— not into the world, but into the hands of others who I may not be able to trust.

Speaker D: Health data is important. It’s an important tool. It’s an important research tool. Not fine just to give it to the government with carte blanche to sell it to whoever, you know. I’m sorry, but that, that’s too far for me, and I think it should be too far for most people.

Speaker E: Find the research which marries the conclusion you come to. It means there’s an absence of research that’s gone on, and that isn’t an area worth exploring because all of a sudden through data analysis, you’re able to uncover a classification and a consequence that we didn’t, we didn’t think existed before.

Speaker F: I mean, I’d like to get to a situation where people could say, well, I’m donating my data for medical research. We got all these people are blood donors. Thank you very much indeed. Blood and platelet donors. That’s great. Take your time. But for being a data donor, you don’t have to give any time. All you have to do is just say, yes, I want to help medical research. Yes, I want to help the NHS plan so the right resources are in the right place at the right time.

Speaker G: If you compromise that trusted relationship at the very heart of not just the NHS but every healthcare system, yeah, and we’re talking about something going back literally thousands of years to the Hippocratic Oath, then, you know, you’re throwing the baby out with the bathwater.

Speaker B: When news of the government initiative was announced, it stated that anyone who wanted to opt their data out of what was being dubbed the NHS data grab had to do so before the start of September. That provoked a ferocious backlash. Over a million people a month logged onto NHS websites or obtained the relevant paperwork and filled out forms that slapped a “do not use” notice on their health records. So what was happening? In a world where we happily let massive US tech companies have as much access to our data as they want, It would appear that access to our health data is just getting too up close and personal with our bodies. It’s a point made by Duncan Hyning, a music writer who started a local campaign in Suffolk to raise awareness of the government’s plans and encourage people to opt out.

Speaker C: When I think about this issue, It seems fundamentally wrong from so many perspectives, including an ethical perspective. There are obviously issues about data security and privacy and the right to privacy, and we have to note that the NHS does not have a good record in terms of data leaks. There’s the issue of whose data is this. I for one do not accept that information regarding my health. That information is something that is shared by me with those medical practitioners who are responsible for my healthcare. It is not information that is to be shared without my consent, and my proper consent and informed consent. What I mean by informed consent is that this is not the kind of issue that should simply be pushed through without proper public debate, and that seems to me to be something that’s been severely lacking in this particular situation. Those are not by any stretch of the imagination the only issues that concern me. I’m concerned that essentially this is a further step towards privatisation of the health service. Those companies who are able to buy that data or access it clearly will have a huge competitive advantage. They will clearly be using that as a means of deciding which particular areas of health service delivery can be most effectively targeted for privatisation. Given the potential power of these companies, the danger then arises that as a society we lose the democratic ability to prioritise health to make decisions about where the priorities for service delivery actually lie. And the danger is that those priorities will be skewed by those companies in terms of commercial interests. And in recent times during the pandemic, the issue of trust has arisen. We’ve seen companies, basically companies being established overnight in order to obtain government contracts for PPE and other services during the pandemic. The awarding of those contracts has been a major cause for concern, and in these kind of circumstances, how can one trust the assurances that one is being given? But whether it’s by those involved in this particular NHS project, or civil servants, or government ministers. They do not have a good record in those terms. I’m also hugely concerned at which companies are likely to have access to that data. Will it be just pharmaceutical companies and companies engaged in other forms of medical research? Will it be companies who provide ancillary services to the government, for example Serco, which has an appalling record? If you talk to any NHS member of staff at any level, if you talk to them about Serco, they will tell you what an absolute disaster it has been. Will companies such as Serco have access to that? Will health insurers have access to that? Will Virgin Health have access to that data. And I think the public has every reason to be suspicious and concerned about how their data is actually going to be used and what the longer-term implications of the sharing of that data might involve.

Speaker B: Do you think you own your data?

Speaker C: Yes, I do. I think I own— in my view, I own that data.

Speaker B: Jaz wrote to Duncan Hyning from Suffolk on why he was encouraging people to sign up to an NHS data opt-out to prevent their health information falling into the wrong hands. Hyning, by his own admission, was just a concerned man on the street. But he was one of those millions who thought that the government announcement was a step too far. Quite how far was spelled out by Christina Zaba, a health journalist who has been a leading campaigner against health data leaking out of the NHS for decades. For Zaba, this is not only because of the intensely personal nature of the data, but also because of the honesty and accuracy of the data, which she says has been collected in an environment that is akin to a Roman Catholic confessional.

Speaker D: I mean, a lot of people don’t go and see a priest these days. I actually do. I’m a Roman Catholic, and I’m one of the few people in the country that does go to the confessional. And I remember being brought up by my mum to be honest in the confessional. And she always said, a priest will rather be murdered than divulge what people say to him in confession. Because she’d been through the war, and she knew priests who actually had been murdered or gone to concentration camps, been killed, rather than divulge what they knew from the confessional. So it is a kind of really, really deep sacred promise that you won’t divulge it. And you’re right, when we go and see our doctor, we have to be honest because we want to get better, we want to have our proper diagnosis, and to do that we have to tell the truth. There’s no point asking for a diagnosis of something that’s wrong with you if you’re going to falsify your symptoms, because then you’ll be treated for the wrong thing and bad things will happen. So it’s impossible. We have to be honest when we go and see the doctor.. And it’s not just about, well, I’ve got a cough, or, well, my backside hurts. It’s what have I been doing? Where have I been going? Have I been sleeping? With who? You know, it’s all these very, very personal things. Have I been biting my nails? Have I been waking up in the night tearing my hair out? Things that are really embarrassing and things that are really private. And things that you could be blackmailed for. I was a campaigner against identity cards 10 years ago, and it was the same problem. That if you had all your information in a digital biometrically identified format on a card, the easiest thing is to nick the card and blackmail the person. Well, I’ll let this out unless you do X. And, you know, it’s a similar thing here. It’s such secret information, it’s such private information a lot of the time. For example, it could affect people’s insurance policies, It could affect people’s reputation, it could affect their job, their standing. We all need a front door in our lives. Not everything is for everybody, and privacy is part of being human. It’s a human right, really, I think. That doesn’t mean to say you should keep secret things that shouldn’t be kept secret, but I think it’s our right to choose. And health data is one of those things. It’s private unless I choose to share it. And that’s this whole initiative which has been going on for a number of years, I might say the push, because of the value of it, the financial value of this database. There’s kind of an impetus, like a push, like a pressure almost, from the people who would like to use it, sell it, and exploit it, to allow people to just let it go, and that worries me very much. So there’s a big leap of faith there.

Speaker B: For Zaba, the confessional nature of the data, its truth, makes it so deeply personal that, like Heining, she says that this is a debate about whether the government has the right to take a part of you that Zaba says is as much a part of you as your fingernail. And whether we can trust the government with information that is so revealing.

Speaker D: Trust is something that I think is a commodity governments don’t necessarily care about. But I think ownership is something that everybody cares about and especially governments care about. And I think that, you know, we are in a sense our data. The more data that exists, electronic data that exists about us, and we’re talking about an electronic database here all the time, aren’t we, not paper records, The more electronic data that exists about us, the more we have like an avatar or a corresponding being in electronic space. And the question is, which is the real person? And does the human flesh-and-blood person own, and to what extent do they own, that electronic data version of themselves? And I think that in in the rush to get technology moving and working, that key ethical question has been almost completely ignored, and I think it’s vital. And it informs everything else, doesn’t it? Because you’ve got your electronic self. Now, here I am with— I happen to have 10 fingers, but only 9 fingernails, interestingly. So my 10 fingers and 9 fingernails, I can tell you I own them. Because you’d have to cut them off with a knife to get rid of them, take them away, right? And that’s a mutilation, and that’s a crime. But do you see what I mean? We don’t have ways of asserting ownership over our digital selves in the same way at all. So when my 10th fingernail was removed with a knife, it was done as part of surgery because it was a sick nail, and I gave loads of especially carefully crafted permission. In fact, I worked with my consultant for 4 years. There were lots of permissions given and given and so on. And by the time that the nail was actually cut off, I’d given lots of permission and I was happy about it. And even then, on the operating table, he said, are you sure you give permission? And I said yes. That’s the, that’s the degree to which our physical body is our ownership. That’s the degree to which we own our physical body. But with our digital body, with our digital self, there’s nothing like that. Yes, we click accept when we go on to things and they say, you know, have you read the terms and conditions? And we go, yes, although we haven’t, and we click accept. That relies on literacy, that relies on all sorts. Well, it’s a very, very weak thing, but it’s a legal sort of yes, I’ve accepted. Fine. But that’s not nearly enough. When we’re talking about a body of data that goes back to our birth, maybe our ancestry, that talks about our physical selves, that has all these other ramifications. It’s just not enough. We need to have a way, and we need to think about this way before we start giving away the National Health Service data willy-nilly to huge organizations who want to make designer drugs out of it, which may not necessarily benefit us at all, so much as charge us even more money for buying them back? I’m afraid that’s the cynical answer. And who gets the money? You know, who gets the money? Again, you’re saying, well, the government would say, well, we collected the data, we hold it. You can say that, but I would argue, well, without my data, you wouldn’t have it. So at least 50/50. It’s like a theft to me.

Speaker B: It’s this idea of data ownership which excites both Haning and Zabber, both of whom suspect that the government sees the data grab as a way of getting hold of one of the most valuable databases in the world. The data the NHS has is the longest continuous national population profile, and it’s growing all the time. Founded in 1948, it sees a million people every 24 hours a process that grows its 73 years of records and shows a unique pattern of the development and treatment of illnesses. It is this that many of those suspect is at the heart of the government’s interest. They suspect a plot to sell our data to technology, pharmaceutical, and insurance companies to help fund the NHS. Here’s Christina Zaber again.

Speaker D: So we don’t need to think, well, let’s sell our data as a nation and then we’ll be able to fund our NHS. It’s not the same as selling your old guitar on eBay and then you’ll have money to go down the pub. It’s not the same at all. We could manage to fund our NHS by raising taxes a little bit, by juggling things around, by doing all sorts of stuff that would require the political will to fund the NHS. Right now, I don’t see that happening. So I don’t think that selling our data to anybody would particularly solve that problem. I think that’s a political problem that we need to solve as a political problem. Having said that, there is something else hidden inside this, and that is the financial value of this data we’re giving away for nothing, aren’t we? I mean, my data is mine. Shouldn’t I be able to sell it if I want to and keep the money? Why is it the government’s data? It’s my health, do you see? So who owns the data is a very, very big ethical issue, really.

Speaker H: The government would say that it went to the trouble of collecting all of this data, it’s the one that is storing all of this data, so therefore it owns the data. In the same way that if you take notes as a journalist and you write them down, you say that you own those notes.

Speaker D: Well, but the NHS owns it. If anybody. That’s not the same as the government. The government comes and goes and is elected. The National Health Service isn’t part of the government, is it? Come on, transparency. That means you look into it. How does it work? You lift the bonnet, have a good look at the wiring. People are not stupid. One of the things that very much annoys me and depresses me really about the way these things are done is that those who, who are in charge of it assume Everybody’s thick. Now, sorry, you know, in my view, the public are not thick. They’re intelligent. You have to give them a chance. It’s up to those who have something to say to explain it properly. If it’s not being explained properly, then obviously people won’t understand it. That’s not their fault. It’s up to the explainers. You know as a journalist, I know as a journalist, that’s the first job: explain it properly. So people can understand it. And again, we do have good communicators. It’s not hard to explain things so everybody can understand. You have to keep it simple and truthful.

Speaker B: Journalist Christina Zabber explaining and highlighting why the government needs to strike a deal with the people that they can understand and buy into with their data. You’re listening to Password on Resonance FM with me, Peter Warren. After this, you can hear DJ Ritu with the World in London. Here on Password, we’re talking about the government’s cancellation of what the papers have been calling an NHS data grab. There’s been much written about this, and everything has become a little heated. So what’s it all about? What data does everybody want? What value does it hold? Here’s Amir Butt, the former CEO of TumorTrace, a company that combined AI and tissue scanning to diagnose cancer. I asked him whether opening up data records to artificial intelligence could cut waiting lists.

Speaker H: You’re very intimately involved. Is AI better health diagnostics in a human being?

Speaker E: It is better in some diagnostic fields for sure, and where it is not better, it will become better, because the more, more data you gather, the more accurate your predictions can become and your classifications of diagnosis or screening can become. So that’s a function of time. It’s not if, it’s when. When the areas that it’s not better at will become better. And one of the main reasons is that, you know, once the algorithms are trained, algorithms within the context of the way they’ve been written and the training they’ve undergone, they don’t make mistakes.

Speaker I: They will—

Speaker E: they don’t get tired, they don’t get overwhelmed, they don’t suffer from anxiety, or having had an argument with your partner in the morning when you come in and You know, it’s not subject to interpretation.

Speaker I: It is.

Speaker E: And just because of those, you know, if you take those factors into account, as long as they’ve been trained and the algorithms are right, they will keep giving you the classification.

Speaker B: And But went even further. Give the scientists more data and they could begin to spot patterns and trends that doctors are unaware of, identifying underlying conditions and discovering potential new new research areas.

Speaker I: If you have enough data where you’ve got measures and definite known outcomes from those various measures, you can use artificial intelligence to give you— to train the algorithms that given a set of other measures, it will give you a classification which is close to the truth. Or as close to the, you know, it’ll discover the underlying patterns in the data and give you the classifications. As long as you have the underlying data and the classifications, you can build AI to give you the predictions. But at the same time, you could look at it from the point of view of, you know, if somebody says, yes, you’ve got, you know, looking at your kidney readings, I predict that you’re going to have some sort of you know, degenerative ocular problems later down in life. I can, you know, I can— I know the biology, I know the, you know, the pathways that cause that. And of course, that’s important to understand. And it’s important to— because in building that knowledge, there are lots of, you know, there are lots of positive consequences as a result of understanding that pathway, which if you’re just operating in a black box AI environment, you never uncover pathways, then become very useful in other ways. I’ll give you an example of the oil industry. When you get the raw material out of the ground, there is the whole value chain of oil production. It goes anything from making kerosene to airline fuel to diesel to normal petrol you put in your car, gases that come out. Tar that you get manufactured, plastics that, you know, come out of it as well, where if you are operating on very narrowly, give me the data, I’ll give you the consequence, you are missing the value that’s contained in understanding that value chain and the benefits that you derive from it. So, but I think in my— and this is my opinion, I’ve not heard this anywhere else, this is very, very much my thinking, that there is a way of marrying the two together. And that is that all of this research and the pathways and the way doctors are trained— they’re trained to write, you know, do the research, write the papers that are then circulated and available to everybody, and people build on that as you go along. So there’s no reason why you can’t take a pathway, using that pathway, generate a whole heap of data and use that data to train the algorithms. So you can dovetail the two together. So you then know that this data that you’re training your algorithms has come from deep research. So that together, I think, is a, in my humble opinion, is a winning formula. It does mean But it doesn’t mean that you shouldn’t use, you know, sort of, just give me data and I’ll discover, you know, what’s in it. Once you do discover something which has not been a subject of research, if there’s no research papers which backs up those kinds of conclusions, that to me is a list of research then you can do. And I think the clever people are going to be the ones who are able to take the results of the AI and direct people to do some research to find out the why answers. In general, in the field of medicine, I mean, the applications of AI are just, you know, we’re just discovering them and they are, some people will be scared by them, but I’m exhilarated by them. Various things have an underlying pattern. Like, for example, put X number of chemicals together and they will form a particular pattern, shape of protein. And that’s an underlying pattern. If you have symptom A and symptom B and symptom C and symptom D, and the implication is that you have or have not got cancer, that’s an underlying pattern that you’re uncovering. So in the area of medicine, anything that actually has an underlying pattern will benefit greatly from artificial intelligence. And anything which is truly random is unlikely to be of benefit. And as I sit here and think about it, I can’t think of anything in biology which is which does not have a pattern that’s waiting to be discovered.

Speaker B: This is a tension between the data scientists and the people who the NHS hold records on. For those analyzing the information, they are data, and for the people whose records are being used, they are human beings. And they do not trust the government, particularly over health data, because of the recent fiasco over the track and trace app., which again was widely ignored by the public because of data concerns in part sparked by the government stating that it wished to keep the data for 25 years. Another factor is an ongoing discussion that has been fostered by right-wing think tanks like Reform, which have openly lobbied for the creation of an NHS data market. Phil Booth of Med Confidential, a group which has been a long-term opponent of the government, says that that’s because it cannot be trusted and that it wishes to turn our health data into a corporate free-for-all.

Speaker G: Yeah, so this is one thing which we’re often asked.

Speaker B: What’s—

Speaker G: why are they doing it? What’s the value? Well, they’ve said why, and they’ve been saying so for some years now. Back in 2014, There was language around the justification for the previous GP data grab, which was called Care.data. It said, to stimulate economic activity around the data of the NHS. That has now transformed, certainly post-Brexit, into an explicit life sciences industrial strategy. It’s been published since 2017. It was written by Professor Sir John Bell, who, one of the supporters of this GP data grab, And it explicitly says, I think the top recommendation, page 5, right-hand column, says we will create 2 to 3 new industries in the UK based on patients’ data, genomic data— that’s data to do with or extracted from your DNA— and from AI and big data. So it’s clearly the government’s intention that a large portion of our our country’s economic output be based on, you know, our medical records. What that means for patients, well, if you don’t have a choice, your confidentiality is compromised. You don’t know who’s getting a hold of this data. Is it going, you know, like we saw with the, the PPE procurement during COVID will there just be data VIP lanes to ministers’ mates and cronies? What we do know is that this is a fantastically valuable bunch of data, possibly.

Speaker H: Well, that’s, that’s the issue, isn’t it, Phil? This is one of the most valuable databases in the world. This is the NHS database. And because of that, because of the length of time over which the data has been collected, the number of people, there is not an equivalent data health database anywhere in the world. That’s why it’s so valuable, isn’t it?

Speaker F: Absolutely.

Speaker G: And this is the thing which I think people maybe overlook. We’re so used to having the NHS sort of there to sort of treat us and stuff. You know, no one else has had a National Health Service for 70, 73 years. No one else has delivered in quite the same way. We’re probably— the NHS in England, you know, the total amount of data, it’s not in one database, but the total amount of information that’s gathered is at least an order of magnitude, 10 times bigger, than any other similar aggregation of data anywhere in the world, even in big countries. So, you know, all of that means that it is fantastically tempting to all sorts of people, not just pharmaceutical companies, but increasingly data companies, you know, the Googles and Amazons of this world, to all sorts of people who want to sort of do startups, these AI or machine learning or chatbot type things. So everyone thinks that if they could just get a hold of the data, you know, they could start making a fantastic amount of money. And of course, they can claim they’re doing good because, hey, we’re helping the NHS. But when you look at who’s actually getting helped here, often it’s favoring the— the deals often favor the company. And not certainly the patients and not the NHS. We’re going through a revolutionary transformation of our society through the rise of data processing at global scale, yeah, and I think we have to fight in the same way as the government is trying to sort of divide us all up and grab our data and what have you. We actually have to face this full-on and say, okay, we’re going to fight for what we care about. Yeah, I think these are not peripheral values in an information society, if that’s what we’re in, and I think we are. Things that you do to my data affect me in my life, and sometimes they affect my family. Yeah, so I see these things as genuine attacks on aspects of my way of life, and I see it more. I’m a privileged guy, I’m a white middle-class male, I’m the very definition of the person who should be all right in this, but I see people who are being really badly hurt by systems that should be helping them.

Speaker B: Booth, like Heinin and Szaba, thinks that the issue is one of data ownership. Many of the supporters of the creation of a market for health data say that people in the UK get their healthcare for free, so they should be prepared to contribute their data in exchange. But that’s an argument that is fiercely rebutted by Booth.

Speaker G: Number one, the NHS isn’t free. It’s free at the point of delivery, which is very different. We all pay for the NHS through taxation. So the notion that we should have to pay a double tax, pay through our money taxes, if you like, and then pay through tax on providing our, or requiring our data to be used. That’s not a part of a social contract that was made. And if that’s what any government wishes to actually say is going to be the new social contract for treatment, you know, if you get treated by the NHS, you have to surrender your data, well, they should make that explicit, go to the country and see what people think.

Speaker B: So what does Booth think should be done? How does he reconcile the potential benefits of AI and big data with individual control of their own data and achieve the trust necessary to do that?

Speaker G: Well, from Medco Financial’s perspective, we want to ensure that every use of patients’ data, also service users’ data across social care, is consensual, safe, and transparent. We think those three things are absolutely vital, and they interrelate with each other. And without any of them, or any of them done properly, then patients are just simply not— their rights are not being respected, their confidentiality is not being respected, and their right to dissent or consent is not being respected. So we’ve been fighting this fight, if you like, since shortly after the 2012 Act, Health and Social Care Act, was brought into force because some of our doctor friends found some what looked like a rather nasty loophole for a GP data grab that later occurred in early 2014. But we look at the whole of the NHS and the systems around it and those that try to get a hold of data. We fought companies as big as Google DeepMind around their grab of 1.5 million patients’ data at the Royal Free Hospital. But we’re not just sort of fighting to stop stuff. Yeah, consensual, safe, and transparent means something. And so, you know, we have fought and won various rights to opt out, and we’re trying to improve those or that mechanism so that people can make their choices much more clearly and that their choices are fully respected across the whole system. The safe bit has sort of come to the fore in recent weeks following the sort of attempted GP data grab. The former Secretary of State said that everything should go into a trusted research environment, a piece of language which may be quite obscure to members of the general public, but certainly in research and in health is understood to mean not just a highly secure database, but one with all sorts of protections and rules and things around it. The sort of thing that, frankly, the, the Office for National Statistics has been using for all of our raw census data for over a decade now. They’re the really, really sensitive, very, very large population datasets which are incredibly valuable. You don’t want to be sending out copies. And the transparency bit, well, it’s often overlooked. People find out what happens to their data through the headlines, you know, when something goes wrong. And we believe that that sort of understanding, that knowledge, that evidence should be just more regularly available to you, like your mobile phone statement or your bank statement. You’d be able to sort of see how your data has been used, what it’s been used for. You have to check who’s accessed your record. You know, people’s summary care record, which I think probably lots of people may have become aware of during COVID when they’ve been treated outside of their normal GP practice maybe. Well, you know, if that summary care record is being accessed some part of the country you’ve never been to, you know there’s a problem. The system won’t know that. You know, the doctors don’t check this, but you, the patient, having access to your own records, obviously seeing what’s in them, but also seeing who has checked them, who has used them, I think could over time build a much greater, better understanding of what’s going on with your data and hopefully better trust.

Speaker B: Are you, are you, are you telling me that we can’t trust the government?

Speaker G: Well, Apparently not. The government’s made all sorts of moves here. The most recent one, which probably, well, while we’re speaking now, was the one in mid-May where they tried to get every single person in England who’s registered with a GP practice to try to get their entire medical history, take a copy of that, and basically make it available to be sold to third parties. I mean, used obviously by legitimate research and legitimate NHS bodies, but they’ve just persisted this loophole that means we cannot trust who gets a hold of our data once it’s been collected up and you’re scooped up into this sort of central system.

Speaker B: Long used to being a political football, is it just rolling with the punches and trying to get on as best it can? When Duncan Hyning started up his campaign on the social media platform Nextdoor Neighbour, which aims to be a genteel, well-mannered version of Facebook, one of those counselling contributing data was Anthony H, a former NHS IT employee. Anthony pointed out that there are some very simple advantages to sharing data, the main one being achieving the efficiencies advocated in the series of reports into productivity and waste carried out by Lord Carter of Coles, who identified savings of over £6 billion that could be made in the NHS. A point conceded by Birmingham University’s Professor of Life Science Innovation, Mike Lewis, who also chairs the UK’s National Institute of Health Research and is at the heart of the NHS’s health data plans.

Speaker F: I think it’s a really well-intentioned thing that they were trying to do, the NHS, NHSX, and NHS D. I think they made a couple of little missteps, but the overall aim is to get the data which is in the systems, EMIS, TPP, primary care systems, and use it for doing things like planning and doing things for research, both of which are really important to do. You can’t plan a health service if you don’t understand what the data is. And you can’t do research like we do at the National Institute of Health Research if you don’t have access to people to conduct the research on. So the intentions were absolutely right. I think where the missteps were, there were 3 missteps really. Communications were really poor. People weren’t told what’s happening, what the benefits were. I think they saw this as a technical issue. NHS Digital, they’re in the middle of a change of leadership, or they were at the time, interim CEO, new CEO being appointed. People really thought this was just a technical thing where you, you plugged in the data which is on the cloud servers, so EMIS and TPP, you plug them in and you connected them to some other cloud servers and it was a technical switch. So they made the misunderstanding, they thought it was purely technical and it wasn’t, it was a lot more important than that. I think the third misstep really is there’s been a lot of trust lost with the general public. With track and trace, vaccine passports, the Pingdemic. All of these are digital things where there’s been some missteps, and that’s overlaid onto this issue. So the public don’t have a great deal of trust, and we need to re-earn the trust if we want to take the data and use it for things like planning research. We have to get the trust back.

Speaker H: And when you say banning and research, give me a few examples of what you can do with that data.

Speaker F: Okay, where is the best place to build the next maternity unit? I don’t know, if you’ve got data in terms of pregnancies, data in terms of information about age and demographics, you probably have a good idea where you should build the next maternity unit. Where do you need supplies of flu vaccine to be sent? We need to plan for that, and you can’t send out a SurveyMonkey study to GPs and ask them to fill it in. You have to have access to data. Where should we build our next MRI unit? Where is cancer taking off? How can we plan where to train people? Do we need more medical students and more nurses trained in Birmingham, more nurses trained in Newcastle? Where should we get the next lab technicians? Where should we centralise the diagnostic labs so that we can get them the closest to patients and we can get the quickest results? That’s just 3 or 4 examples of how you need this for planning.

Speaker H: So it’s very important then that this debate should, should have happened, isn’t it? It’s very important because we’re in this new digital age, we’re in this new age of artificial intelligence and big data, so people are going to want to do this more and more.

Speaker F: Of course they are. I mean, just think about how much data HMRC holds on us all. We’ve got all our financial information stored away, everything about us, what we earn, where we, where we spend it, all our expenses, everything. They’re all stored in there on P11Ds and your, your tax returns. Massive amounts of data that the government already holds about us on a financial level. But then if you think about primary care data, it’s all held in little blobs all over the place in a disconnected format. And we need to understand what that data is and how we could how we can use it. Because how, how are we going to build a digital delivered healthcare service if we don’t have access to the basic fundamental building blocks, which are the pieces of data?

Speaker B: Like Amir Butt, Professor Lewis points out that there is value in that data, but as he also points out, that value is also in being able to achieve efficiencies in the NHS that we will all share in—better planning, better response, better care—not just medical breakthroughs. But to achieve that, the NHS and the politicians need goodwill and trust. Professor Lewis acknowledges, because as he admits mistakes have been made, that they need to be learned from.

Speaker F: We had whole GP practices as well, GPs in the practice deciding on the patient’s behalf without talking to the patients to unconsent them as well. So it was a reverse side. There were some hardline GPs out there who said, no, we’re not going to let you have access, we’re going to deconsent all these people. That’s because, and if you step back and look at the real reasons behind this, it’s because they didn’t understand why the data was being used, how long it was going to be used for, to take your point, when it was going to stop being used, and what were the benefits to them. What were the individual benefits? Now, if you’re going to sell or convince people to change their course of action and do something different, you have to give them reasons and rationale. You have to tell them that there’s going to be a gatekeeper of the data. It’s not going to be a free-for-all with people able to access the data, companies able to access the data without having a damn good reason. And there’s going to have to be governance and measures put in place to stop people. They talk They talk a lot about Trusted Research Environments, TREs. That’s an acronym you’re going to hear a lot of. But basically, that is a fortress front door where you have to come and say, knock on the door, say, I want access to this piece of data to do this piece of work to look at this. So maybe my colleagues at Birmingham would say, I need access to orthopaedic data on this age group people because I want to study this surgical intervention. And then you open the door, let them have the data, and close the door again, and it stays behind this trusted research environment. We’ve had instances in the past— I mean, you probably remember what happened to the Royal Free Hospital with the orthopedic data and DeepMind and Google. Google now own DeepMind, and they were using the data, and then suddenly people found out it was being used. It was a surprise. You can’t have those surprises with data. You have to tell people in advance how it’s going to be governed, how it’s going to be managed, who’s going to be governing it, and it better be independent, and then how long it’s going to be used for and what the benefit is at the end of the day.

Speaker H: That seems eminently reasonable. That seems to be the right thing to do. You know, if you say, I want to do some research, can I have your data please? Because you’ve got data that is of interest and is therefore useful. I want to use it to do this and I want it just for this amount of time.

Speaker F: So why—

Speaker H: I mean, the government doesn’t seem to be too willing to actually do that. And I’m talking about the government as opposed to separate from the NHS. It seems to want to be able to have carte blanche to use data because we’re not just talking about the NHS here, are we? We’re talking about a debate that’s going to start for the next 10, 15 years.

Speaker F: Yeah, and unfortunately for me, I sort of focus purely on the health service and the healthcare data, so I just have a view over that. But there’s all sorts of other data that the government could get access to, and it’s phenomenal if you think about how much data people, companies have right now about you in terms of, in terms of data. A use case I’d give to you is let’s say that you wanted to get some repeat medications from, from your GP. First of all, you, you go into eConsult and you book your session, and eConsult is a privately held company, so they’ve got data. They log the task into EMIS, once again a privately held company, so actually a public company, but it’s a private company, so to speak. They have your data. You request the meds, it goes on EPS, the order gets sent to Boots, private company, it gets shipped by UPS if they order the home shipping. So I mean, there’s a lot of data out there already already on healthcare, and there’s not too much governance around it. I think one of the benefits of doing the GP-PFI is actually getting the data and having a control, a governance, and management, and a means of accessing the data. So it is a tightly managed ship, and I think the TRE is the way to go, particularly for academics, particularly for the National Institute of Health Research, and also for any other, any other company that wants to do some beneficial research, they’d better come to us and tell us, we want to study this group of cancer patients, we want to try this new type of therapeutic, we have ethics approval, we’d like data approval as well before we go ahead with the study. And the benefit could be early studies have shown it can reduce mortality by X or increase longevity by Y. You better be able to tell us that this is what we’re expecting to see. And at the end of the day, the data comes back.

Speaker H: Some AI scientists are saying that the reason that they want unfettered access to data is that they can suddenly find some anomaly that they weren’t expecting to find, which will then indicate a new research area. Does that get cut off under the system that you’re talking about?

Speaker F: No, I think AI— you will find AI doesn’t look at everything everywhere every time. AI will be for a specific purpose. So let’s take some of the really good work that’s going on. People like Kira Medical Technologies, they’re looking at images of— they’re looking at mammograms and they’re trying to draw inferences there and trying to diagnose breast cancer from the mammograms alongside the pathologist. So they would want to see all mammogram data, all scans, this type of age group, and then run the machine learning over the top of that. And does some pattern recognition. So AI doesn’t mean everything everywhere. It’s not like Martini AI, anytime, anyplace, anywhere. It’s about specific datasets. And you better be able to tell us as well in this Trusted Research environment, I want the data to do this, and this is what I’m expecting to see out of it. I’ll use the data for this purpose. It will stay in my box over here until the work is done and then be returned as well. That’s the whole point of these TREs. Yeah, I think some early stumbles led to some fairly big falls, and I don’t think it was done with the wrong intentions. I think the people at NHS Digital, certainly the people I speak to, have had the right intentions about this, even if they thought it was technical. They were rightly intentioned, but if you create a vacuum or you allow a vacuum to exist of communication, people will fill it with their own stories. We’ve seen that with the COVID vaccine deniers. The people will fill it, and the ivermectin crew, They will fill this with whatever they want. There’s, there is none of the talk I’ve heard about sort of having a way of packaging data up so they can sell it to big American pharma or UK pharma for that matter. That’s not the intention. A lot of the intention is built around getting research the data it needs. UK-based research. We got 5,600 clinical studies running, 750,000 patients enrolled in those studies. We need the data. We need to be able to find patients to get onto studies. The right sort of patients. We need to make sure that the patients we do find are not the ones who are putting their hands up, but they’re all spread out so that we get rid of some of these health inequalities we have in this country. Some of the, some of the race inequalities, the poverty inequalities, the gender inequalities we can address by having access to this data.

Speaker H: Okay, so it’s not a question of the NHS data being put into a great big lorry, sent off to the Americans, and then the money that can be made from that will subsidise the NHS. That’s—

Speaker F: no, that’s never been— I’ve never heard discussions about that, and I sort of have fairly good discussions on a frequent basis with NHSD, NHSX, and the NHS England people as well. That’s not the intention. We want to get— we want to find patients, put on research, and get our planning right. You know, you can start to look at inferences, but you can start to look at inferences from fairly small sets of data. You don’t have to look at the whole population data, 58.5 million records in England, to start thinking about those things. What you do is you run pilots, you run small studies on small sets of data, you go, oh, this looks interesting, I found something interesting here. Therefore, then you come to the TRE and say, on this set here of this type of people, we found this. Is there a broader set of data we can check it on? But you don’t need to open the whole box to do AI on everything just to look for patterns.

Speaker I: You should be able to find that earlier.

Speaker B: In an interesting twist, Professor Lewis almost echoes the solution laid out by Phil Booth. This is, he says, the new way. There is a new team that has decided to earn our trust. The debate, it appears, will take place, though, as Professor Lewis points out, bruised its experiences with the Royal Free Hospital, where it had admitted that the hospital had given 1.6 million patient records to the tech giant without the patient’s consent, Google has closed down the project known as Streams.

Speaker F: Things can only get better. You know, it is interesting, they, um, once again they hit issues in terms of trust and telling people and communicating. And I think the whole issue in terms of at GPDFPR, it has been about all the people like Med Confidential. They got upset because of, because of trust. I think the BMA got upset as well because of trust. But I understand that the Royal College of General Practitioners, I understand that the BMA, they’ve now all been brought into the tent by NHSX, and they’ve been brought into the tent so they can check and make sure that everything at every step matches what the stakeholders want. So you’ve got people like the Open Safely Group, BMA, RCGP. I think you’ve got the National Statistician involved as well. You’ve got all these people who are an external reference set. So whereas before they were just blindly going ahead and thinking it was a technical change, now they brought in people to say, oh, hold on a minute, hold on, let’s narrow down this TRE, let’s narrow down who can get access. So they, they’ve listened And I think one of the biggest lessons they’ve learned is that this is going to take longer than they thought. I think there has to be a compact between your GP, between your, your specialist doctor as well, and the patient. And there has to be not only trust, you have to be treated like adults. So you can’t go in and start digging and burrowing into every single purchasing habit, every single eating habit, how many steps you walked on your Fitbits, unless you want to share that data with your GP., and that’s optional. But the basic stuff about your care, what was done to you, where, what medications you took, what were the side effects you had, what sort of dosage you were on them, how long you were on, all that sort of information, what procedure you had, that information is valuable for planning and certainly great for doing research as well. I mean, if I’m doing a study on diabetic patient, type 1 diabetic patients, I need to know where they are. And if you do studies today, Generally people put their hands up and say, ‘Oh, I want to be part of that study,’ and they fall into a certain demographic, and you have to work like crazy to make sure that you’re including South Asians, demographics from Blackpool and the Northeast, you’re making sure that you’ve got your right blend of gender as well. You really have to work hard, and I think with the GP data, if you can get access to that in the trusted research environment, it’s going to help research enormously.

Speaker H: Okay, so what you’re saying, we almost need a public information campaign, and we need to have this out there in the open where people say very, very clearly what they want, why they want it, and when they want it.

Speaker F: Yeah, and give people use cases, give them hard and fast use cases, and say, this is how it will help research, and here’s an example. This is how it will help planning. Give them an example. This is how it will help with access to vaccines or whatever. This is how it will help you get access to community care. This is how it will help your GP help your care. Give them 3 or 4 use cases, have a debate about it, and let people understand. We have to start treating— the one thing I hate about the NHS is they always call people patients, and we’re not. We’re citizens, and what they want is access to the citizens’ data, not the patient’s data. And if you’re going to be talked about as a citizen, then you have to be treated like an adult as well, and you have to have an engaged, informed, honest debate. And you can’t go— you can’t treat people with disrespect. You have to share the information with them so they can form their own opinions and they can see this is the right thing to do. And generally People will do the right thing. People will go and donate blood. They will sign themselves up to organ donation registry. They will go and help as charity volunteers at hospital. They will go and work for the Red Cross. They will do the right things. I mean, we are a very charitable, very open nation, and we should be. As long as you give people the information, then they will make the right decision, but you have to give it to them in the right format they can understand and they can comprehend.

Speaker B: And interestingly, Professor Lewis’s comments that a new dialogue needs to be learnt that stresses and respects people as people picks up on an argument that MIT professor Sherry Turkle, author of The Empathy Diaries, put forward in a previous episode of Password.

Speaker J: Our language has been filled with the language of engineering and science, but with the language of objects. To talk about people. I think our language for talking about people have been demolished, undermined. And so I think that computers are good for many things, and engineering culture is good for many things, but it’s not good for talking about human relationships and human understanding.

Speaker B: So there you have it. If data is going to be the 21st century lifeblood of the NHS, then we will have to be reassured enough to trust it with our data and give it like blood, so that once again we will be able to say without irony, “Trust me, I’m a doctor.” You’ve been listening to Password, presented and written by me, Peter Warren, and produced by Blue Buffery. To find out more about the issues raised in the program, go to www.peterwarren.com. Futureintelligence.co.uk or search for Peter Warren on Twitter.

Speaker A: Goodbye and thanks for listening. This program has been brought to you by Resonance FM. If you like what you heard, please support our work by making a donation at resonancefm.com/donate.