In conversation with Dr Mike Lynch and Matthew Gould, British civil servant and CEO of NHSX

2020 has been a year that not many people would have anticipated turning out the way it has, and it has put a huge focus on the NHS – both celebrating the staff on the frontline, but also showing its difficulties. And while in the corporate world we have seen an acceleration in the adoption of digital technologies, medicine is also one of the sectors in which emerging technologies are having the most profound effect.

In this podcast, Dr Lynch and Matthew Gould, CEO of NHSX, a joint unit between the Department for Health and Social Care and NHS England, set up to ensure that staff and patients have the technology they need, discuss the impact of the Covid-19 pandemic, the rise of personalised medicine and which technologies are going to have the biggest impact on our health.

Listen on YouTube, Apple Podcasts or Spotify.

Read:

How AI and automation is revolutionising healthcare and the NHS

On today's episode of Invoke Insights we hear from Matthew Gould, CEO of NHSX former British Ambassador to Israel, and previously Director General for Digital and Media policy for the UK government. In conversation with Dr. Mike Lynch, OBE, Matthew discusses the digital transformation of the NHS, the dawn of personalised medicine and the potential for AI and automation to transform healthcare.

Mike: So Matthew, thank you very much for taking the time to chat today, must be said you're probably one of the nearest things to a modern polymath I can imagine, and that you're the go-to man for recommendations on Persian meals, cybersecurity and technology and healthcare and a long list of other things that you have done in your career. So, in 20 minutes it's going to be quite difficult to get through all of those so we might have to do Persian meals another time. So one thing I do want to ask you about which is in the news today is the, some might say, final arrival of the internet harms legislation, and the idea that we are going to see some boundaries on what is acceptable in the social media and online world. I know that is an area that you have done a lot of work on behind the scenes, so are you pleased with how it turned out.

Matthew: Well, this isn't about my current job but my previous job when I was director general of digital policy at DCMS where I started an online harms and security team, and we came up with the plan and framework which is being announced today by the government. So, look, I think it's a serious, well thought-through attempt by the government to do something which hasn't been done anywhere else in this form in the world which is to tackle in a comprehensive and systematic way the risks that online harms pose. I haven't been in DCMS for the best part of two years so I'm a bit out of date in terms of what's happened since I left but it's identifiably the same plan that I was working on when I was there, and it's nice to see it come to fruition. And obviously a lot of thought has gone on in the meantime, mitigating some of the risks and issues associated with what we were planning. So, it leaves me optimistic.

Mike: I think it's really interesting, because obviously the reality is this online world is so central to so many people live now so making a start on producing a sophisticated and forward-looking framework is definitely a good thing. That is what you were doing then, obviously now you've been spending a lot more time thinking about technology and healthcare. We've come through this sort of incredible year which is really highlighted for people things that they probably weren't thinking about everyday. Things like data capacity, and how data is used and risk and things like that.

Do you think that the Covid experience is actually going to lead to an acceleration of technology in the health service?

Matthew: I think it has already accelerated both digitization and attitudes. So at the start of the year, just to give you an interesting data point, about 3% of GP surgeries were enabled for doing online consultations. We think that's now above 99%. At the start of the year, almost every consultation across the health service was done face-to-face. A very high proportion of them are now done remotely. At the start of the year, very few people, particularly clinicians, were working from home or felt able to work from home. We've now shown that it is totally doable. Now, some of this is obviously a function of the particular pressures of the pandemic but I don't think it'll go back to where it was before because we've shown that certain things are possible that are actually rather helpful, regardless of whether there is a pandemic or not. So it is much better for everyone if people don't need to keep going into their GP surgeries or hospitals or very routine conversations that could be done online or over the phone. So I think that will change. I think people both stores in health and care, and the public have realised that health care can be delivered differently. And so it's opened the door and it's opened attitudes. And then for me I think the most exciting thing of all is, it's not just about, and it hasn't this year just been about doing the same thing but online, but actually we started with digitising existing processes, but more and more, we are looking at genuinely transforming services and healthcare using digital technology. So to give you an example, if you have hypertension then unless you can afford your own blood pressure monitor, which quite a few people do but by no means everyone then if you want your blood pressure monitored you have to schlep into your GP or hospital on a regular basis to see how your blood pressure is doing. Now, we are distributing tens of thousands of blood pressure monitors, all with a digital underpinning so that people can monitor themselves, look after themselves at home, their clinicians can see how they're doing, the quality of data and the frequency of data will be much higher, it saves a huge amount of patient time, it also saves a huge amount of clinician time and it means that clinicians can focus on those patients who most need attention, rather than just the enormous flow of people who need to have blood pressure monitored on a regular basis and were previously going into GP surgeries.

Mike: The only thing I would say on that Matthew though is that for the past 20 years I have sat in on meetings about using more data in the health service and even more exotic things like diagnostic support tools and things like that, and even though to a technology person it all sounds like a wonderful thing to do it has taken longer than we thought it would. There seems to be good progress being made now, but why is it that, particularly in healthcare, what have been the barriers of getting that data used and the benefits of it

Matthew: So look, for me this is absolutely the central question and in tech terms, what we're trying to achieve with data is a long way from cutting-edge. It is what other sectors have done, sometimes years and years ago, so it is exactly the right question – why has it taken so long in healthcare? So I think number 1, the entire system has gotten itself into a tangle over information governance where it is right to put the highest value on preserving patient privacy, and not undermining public trust that they're very sensitive health data is being properly looked after. But this has, in practice, lead to really complicated guidance from lots of different places, and a fear of sharing data, even when it is absolutely the right thing and legal and appropriate to do so. So one of the things we need to do is sort that out. I think there are wider issues. It's not about the technology. It's about attitudes and systems, and incentives. And I think sorting it out will be about making sure that people have the confidence, the skills, making sure that the rules allow out and incentives align. So although on one level it is a tech question and a data question, solving it is much more about people and institutions and the systems we've got in place.

Mike: So on sort of a related issue, the general public are now understanding more about regulatory approval than they probably ever thought that they would, as we have all been watching with bated breath for the vaccines to be approved. Obviously there has been some quite amazing work in shortening the timescales there. One of the big debates in tech is obviously the effect of personalised medicine. And what we are likely to see is medicines which are personalised to one of the 10,000 variants of lung cancer and that sort of thing. But that doesn't really fit with the traditional regulatory framework which has generally been very much on a cohort basis. So, do you think that we perhaps even because of the realities of people following COVID and regulations, this is the right time to start thinking more sophisticated regulatory frameworks, which move away from cohorts because of personalised medicine, or do you think that actually we should stick with the old tried and tested methods. Where are you thinking on that sort of thing and is it something that the public can be brought along with now post-Covid?

Matthew: So, I think, again, it's a really important question, and just before the pandemic I convened all the chief executives of all the regulators and organisations involved in AI in healthcare, and we had 16 different chief executives around the table, and we agree that actually we need to create a set of regulations and the regulatory framework that gives the public and clinicians confidence, but also which allows innovation to flourish confidently in precisely the ways that you're talking about, because AI and data science, and the use of data allow at least the possibility of a personalised approach to medicine which has long been dreamt about but is very difficult to get to. So I think absolutely we need to look at regulation. Actually, I think, even within the existing regulatory frameworks if you could get data flowing safely but more effectively. If you were a doctor, an oncologist for example, looking at a patient who had a sort of particular situation and you could compare that patient against others with a similar risk profile and look at what therapies work and what therapies don't work in real time, and even with the existing regulatory framework, we should be able to offer a much more personalised approach to medicine. This has been brought home to me by my wife's experience with breast cancer, and was very well looked after by the NHS, but she went through chemotherapy, not because the oncologist knew that she needed that chemotherapy, but because there was a reasonable chance that if they gave that chemotherapy but because there was a reasonable chance that if they gave that chemotherapy to a dozen people in her situation, then at least one or two of them would have a massive benefit from it. Now, it was absolutely worth it, absolutely the right thing to do, but if you could ensure that chemotherapy went only to those who would benefit from it rather than for example, a wider set of people and you're then essentially playing the odds, the saving in terms of resource for the health service, in terms of the hardship of going through chemotherapy potentially unnecessarily, and the targeting of treatment, much more precisely on those who need it could across the health service have an enormous impact.

Mike: I think that's a great error. I actually work with a Swiss company which does sequencing of tumours, just so that you know with chemotherapy drug would actually affect that tumour. I think this is a technology we have seen filed in the NHS recently and that's a nice example. But what I wanted to do was give you a slightly unexpected twist on the data problem. So, I get lots of pitches in from startups and I had one a couple of months ago from a Californian company, and the basic idea was that for some amount of money I could subscribe to their service whereby every day it would measure all sorts of things about, and you know, on Wednesday tell me that suddenly my alanine transferase enzyme was looking a little low or whatever. And I was thinking about this and I thought well the problem is if large numbers of the relatively well-off, worried, well are getting told that their alanine transferase levels are rather low today even if it may have no clinical significance, isn't there a danger that all of these people are going to turn up at the doctor's surgery, and the NHS is going to get overwhelmed. Well, the thing that I started thinking about was obviously we have a health service which at the moment, uses clinical and scientific input to work out whether drugs are useful and work or whether tests are reliable but we're in danger with this consumer technology approach to healthcare that we're going to have lots of patients turning up, because their iPhone's telling them that their alanine transferase is low today. How is the NHS going to stop itself being swamped by the worried well if we end up with this personalised consumer healthcare which isn't really tied to the realities of clinical outcomes.

Matthew: Look, it's a brilliant question but I think my view would be that that is really a transitional problem. That the old model was essentially; you put yourself in the hands of your doctor or your clinical team, and you don't see your records, and you accept passively, the recommendations and the treatment prescribed for you, and it was it was very much an uneven relationship. And that model has held for a long, long time. I think it's no longer sustainable and there is an inexorable shift towards people having access to their own health records and data towards people producing their own data with whatever devices they're using. And as you say, people getting much more by way of signals of issues or potential issues around how they are, and their own health. I think it would be a mistake, though to respond to this by trying to calibrate how much information gets to people. Firstly because I don't think you can. I think that ship has long sailed. But more importantly, because I don't think the answer to problems around people knowing something but not enough, is to try and throttle information going to them. I think the reason I say it's transitional is because over time I think we will create the tools and the knowledge that people need so they know if they find out that their blood pressure has risen a bit overnight or whatever it is. They know when to be worried and when to consult a clinician, and when to say okay well, this is something I'm going to note, but it's within the normal range of variants and I'm not going to run up to A&E about it. I think the best thing we can do is accelerate those tools and that understanding, so that we give reassurance to the worried well, rather than try to bottle off information to them to stop them worrying.

Mike: Well as a fully paid up member of the worried well I shall look forward to the thing that tells me not to worry! Now we're very short on time but there is an area I'm passionate about hearing your views on: there's been a lot of debate about the interface of social and health care. And I've experienced some of the difficulties in that. Unfortunately my mother had Alzheimer's and she lived at home for a long time with a carer who would come in and then she did something which is quite common for those patients; she got a urinary tract infection. That has a strange effect in that there isn't much cognitive reserve so she actually started hallucinating. And then the only option was an ambulance, and she ended up going into Chelsea and Westminster where there were truly amazing staff. She ended up blocking a bed for about seven days and really, it was a question of, if the urinary tract infection had been caught early she could have avoided that situation, and ultimately those kind of episodes lead to her being in a home. So, obviously incredibly complex and difficult question, and one has to be very humbled by the people who are actually on the front line doing this but it does strike me that this is an area where perhaps a little use of technology could be accelerated. You know if we could have been told that today my Mother was acting slightly strangely, or indeed even by monitoring what was going on in the toilet that she had a UTI, then they could have handled her in a way that didn't involve ambulances and hospital beds. Is this a sweet spot, is this a place where technology really can make a difference, or are we really missing the clinical reality of these situations.

Matthew: I think it is, but it's not just about technology and I'm sorry about your mother's experience which sounds I mean really upsetting but the truth is, health and social care fit in quite sharp silos and information does not flow effectively between them, and it's definitely something we're trying to address and again. The technology is relatively the easy bit. It's the rules, and the culture and the systems that allow this to happen, but we are trying to short circuit it, so one of the things we're doing this winter and it's going on as we speak, is we're distributing 11,000 iPads to care homes around the country, which will allow for virtual consultations, virtual visits from the family and over time, allow for key bits of clinical information to be shared between health and care and finding ways, safely, to share data across the healthcare divide will be key both for the NHS and for social care because it's clearly unsustainable, and deeply suboptimal to be in a situation where people are discharged to care homes, without the care homes having access to essential information about their health, and people going from care homes into the NHS without the NHS being able to see essential information about their health before they were admitted, and it is solvable. It is a sweet spot. But the issue is, although we will use technology to solve it, the issue is not primarily technological.

Mike: We are just about to our time so I think two questions for you, quick answers, if you look at all of the things that are going on in technology across the whole spectrum, and not just in relation to health, which applications of technology do you feel most optimistic about which ones are you worried about?

Matthew: I am hugely optimistic, in health, around the use of AI to take out those tasks which can be automated leaving clinicians to do the things where they add absolutely the most value. The edge cases, the caring, the human elements, the judgments and so forth. So I think AI has the potential done well massively to improve healthcare. What am I most pessimistic about? Actually it's not a technology but it's a concern that although the technology opens enormous possibilities, to realise those possibilities we will have to drive change in an extraordinarily complex multi-dimensional system, which has so far shown itself to adopt change at a speed and with a consistency that isn't what we want. So, so my concern is how do we drive change effectively across such a complicated system.

Mike: Well Matthew, thank you very much I am mindful of letting you get back to that Gordian knot you're trying to untie. I am optimistic that we are going to see some great stuff in healthcare with new technology, however always humbled by the reality that you've been putting so eloquently that it's not really about the technology at the end of the day, it's very much about the people that are doing all those things. So with that we have run out of time - Persian cuisine will be our next session. Thank you for spending your time to share your thoughts with us.

Matthew: Really nice to speak to you Mike.