The Not Mini Adults Podcast - “Pioneers for Children’s Healthcare and Wellbeing”

Episode 4: DRIVE with Professor Neil Sebire

July 10, 2020 Season 1 Episode 4
The Not Mini Adults Podcast - “Pioneers for Children’s Healthcare and Wellbeing”
Episode 4: DRIVE with Professor Neil Sebire
Show Notes Transcript

In today's episode of The Not Mini Adults Podcast - "Pioneers for Children's Health and Wellbeing" we speak with Professor Neil Sebire.  Neil is Professor of Paediatric Pathology at Great Ormond Street Hospital (GOSH) in London, and Chief Research Information Officer and Director of the Digital Research, Informatics and Virtual Environment (DRIVE) Unit at Great Ormond Street Hospital.  The theme of our converstaion is 'Data' and there is no one better than Neil to discuss the importance that data has in bringing the future of healthcare to children. 

Neil is a clinical academic with research interests that include complications of pregnancy and the placenta, stillbirth and infant death, as well as paediatric oncology and nephrology. His current research and academic interests are however focused around the impact of digital transformation on health care, in particular how clinical informatics will improve child health.

DRIVE aims to become a world leading clinical informatics unit focused on data analysis, accelerating research and the deployment of cutting-edge technology.  DRIVE harnesses the powerful combination of rich health data with data science and digital innovation to develop scalable solutions to enhance health services not only for GOSH patients but across the wider NHS.

Thinking of Oscar Website and Contact details can be found here.

Follow us on Twitter here.

Theme Music - ‘Mountain’

copyright Lisa Fitzgibbon 2000
Written & performed by Lisa Fitzgibbon,
Violin Jane Griffiths

Podcast artwork thanks to The Podcast Design Experts.

David:

Hello and welcome to the fourth episode of the Not Mini Adults podcast Pioneers for Children's Health and Wellbeing. My name is David Cole and along with my wife Hannah, we're the co founders of children's charity thinking of Oscar. In today's episode, we're joined by Professor Neil Sebire. Neil is a professor of paediatric pathology at Great Ormond Street Hospital in London. He's also director of the Digital Research Informatics and Virtual Environment Drive Unit at Great Ormond Street Hospital. Neil's role at drive is to maximise how they can use data and to drive how to improve digital, for children in hospitals. The theme of today's conversation is very much centred around data, and how it is really being seen as a game changer in the future of health care for children. Neil, Hi, thank you for joining the Not Mini Adults podcast. We're delighted to have you on.

Neil Sebire:

Yeah, hi thanks for inviting me

David:

The first thing that we wanted to talk about was just you work for an amazing hospital. It's known, you know, throughout the world, but can you give us a little bit of an insight as to Great Ormond Street and I guess the role in which you play there as well, please?

Neil Sebire:

Yeah, sure. So just to give you a little bit of background around the hospital. So Great Ormond Street Hospital is a pure Children's Hospital. So we are a tertiary quaternary referral centre. So what that basically means is that we see patients with predominantly rare and complex paediatric disorders that have been referred in from, either all over the country or other parts of the world. So we don't really have a kind of a geographical population base that many other hospitals would have, because we see patients from across the whole country. So that gives us a slightly unique patient population, because it means that we see a disproportionate number of children with particular rare disorders. But it also means that our population is spread across the country. So in terms of some of the issues that we have around patients travelling to the hospital. Having to maybe stay over for the night in London because they live a long way away. That was actually one of the drivers for some of the things we're trying to do now around data and technology, to see if we can kind of improve that process. So my role there is, for the purposes of this podcast, is the Chief Research Information Officer. So I'm a clinical academic, so I still do clinical care. I'm a diagnostic pathologist. So I do some sessions a week still doing that. But in the CRIO role, the primary function that I have is around a.) maximising our use of the data that we're collecting to improve either patient care or patient outcome. And secondly, evaluating new technology. So in other words, trying to work out what we should be doing next, as opposed to the typical kind of CRIO role. Which is really around making sure that the hospital are using the tech and the things like the electronic patient record systems, etc, to the best of their ability. That's someone else's job. But my job is to say, okay, we're collecting this data, what do we now do with it to make maximum use of it? And what's the next thing that we should be bringing in?

David:

How did you get from, you know, your kind of clinical academic work, into data and data is going to be the theme of what we're going to talk about today? You know, that Hannah and I both worked in and work with data and with IT organisation. So, you know, it's, we kind of see it every day. It's been described in many different facets in terms of, you know, the new major resource and all this kind of stuff, but how did you go from, you know, your kind of clinical work to to being interested in data

Neil Sebire:

So I think in the same way as almost anyone else that you speak to, at the moment in healthcare, there was no proper path. So what you'll find is that almost everyone who is involved in this field has got into it, because partly by luck and partly because they were interested in it. So for me personally, I've always been kind of nerdy and into computers. Right from the early days. I've always been involved in data with my research. So a lot of the research that I was doing was around things like stillbirth and cot death, for example, which meant looking at data sets and some epidemiological work and modelling etc. So I'd always come in from that background and kind of understood the value of secondary use of data that we were collecting. Then for this particular sort of programme, it started really about five years ago when the hospital were looking at, essentially changing their whole IT infrastructure. So we had over 400 clinical systems in the hospital. We did a lot of research at Gosh, but it was a very fragmented, so each department was kind of doing their own thing. So we realised that there was a real need and a real opportunity to do something about that, because we were a very active research hospital, I was kind of brought in as someone that knew about tech and data, knew about research and knew about clinical and that there are not that many of us around in that sort of space at the moment. I'm very glad that it worked out. But it does raise the issue actually, more generally, that in order to really gain future benefit, it's super important that this area becomes more professionalised and becomes a root. So having clinical informatics as a specialty that some junior doctors or medical students going through now, can go into, is something that at the moment, you can't do in the UK, there is no official way of becoming a clinical informatician. Whereas in the United States, you can do for example, there are some things that have happened, like the faculty of clinical informatics, which across all the colleges has now been set up. So I think things are moving in the right direction. But the whole concept of both for medical staff, or clinical staff, but also non-clinical staff, so computer scientists, mathematicians, analysts. There are no jobs, there's no clear career structure to become a machine learning person in the NHS and that has to change. I think there's now a bit of momentum around that. So for me, it was very much just, it happened to be the skill set that I had.

David:

You touched on it there. But what you're now running, and I guess, what's come out of it is drive. So can you can you talk to us a little bit about drive what what you're trying to do from from that project?

Neil Sebire:

Yeah, so so very briefly, I think one of the things that we learned, as I say, when we first started looking at this a few years ago from other hospitals, we weren't around lots of children's hospitals worldwide, actually. We realised that there were essentially two very clear kind of buckets of activity if you like and one was around the hospital's own IT system and infrastructure. So unless a hospital can sort out his own electronic patient record system, and all of the other systems that it has to reliably capture, provide data and manage patients, you can't then do anything. But the second part was that you can't use those same systems to do other things with the data, you need a dedicated platform that is secure, that is auditable, where you can take the data where you can do the modelling, where you can test out algorithms, etc. You can't do that in the same places as the clinical systems. What we saw was that, what many places had done was to say, let's sort out the electronic patient record system. Then when that's sorted, we'll think about the secondary use of the data. What you see is that the electronic patient record system is never sorted. So actually, no one ever gets on to the second part or navigates onto it. But it gets pushed and it gets pushed. So one of the things that I have to say our board were forward thinking on was that they agreed to essentially procure and implement the secondary use data platform before we even got the EPR system implemented. So the intention was, we wanted to be a hospital that maximises use of the data. That could be data that comes from the weather, polution, patients themselves, wearables, one source of that data is the EPR system. That's very much how we kind of built the structure. So we we have this thing called our DRE, which is Digital Research Environments. This is a sort of secure platform within the hospital governance structure, where we can take data and provision it into analytic workspaces. So that was the core thing that we wanted to set up. Of course, then we realised that we needed a team to manage the data. So you need a data steward or data engineering team. But as we were scoping this out, I just almost touched on it there is that we realised that, what we currently think of as healthcare data is going to be a tiny proportion of what is actually healthcare data in the future. So the stuff that is generated In a box that says this is your haemoglobin level, is actually a tiny fragment of your total data. And this could include things like video recordings of gait, it could include things like real time data streams, or heart rate monitoring and ecgs. It could include genetic testing, proteomic analysis, and also other data that wouldn't currently be considered healthcare data, which is going to be home monitoring, and wearables and patient generated data. As we started looking at all of these things, you realise that you need structures, both to receive that data and also understand what you're then going to do with it. So it's all very well saying, oh, yes, great, we're going to take data from a sensor. But if you've now got a sensor doesn't really matter what it is, but a sensor that is monitoring something twice a second, that is a lot of data that's now coming in. It's not a case, you can't just say, well, we'll store it, you've got to have a plan for how you're going to use that data, what kind of algorithm you're going to develop, whose job will it be to look at the alerts that come out of that. We realised that it was really hard to do that testing because almost every hospital, only had a hospital. So you'd have a tech company or university that was somewhere else. So you could test something from a very proof of principle conceptual way. But the next step was a massive jump to actually getting it onto a ward. You've got all the things then of infection control, and the governance etc. So it's a huge hurdle. To only find out, for example, that the battery of this device was terrible, the Bluetooth range, you know, lasted a couple of hours, or whatever it happened to be. So we kind of realise that what we needed was a place that was within a hospital governance structure, where we could use real doctors and nurses, but bring them together with tech partners, computer scientists, students, etc. In a safe way where there weren't any patients, so that we could actually almost mock up what it would look like. From our perspective, we're not really interested in building a new device. There are, you know, lots of other people I'm sure you've talked to on this podcast that are much more expert than me in that area. But what we are interested in is saying, what does the data look like from this thing? And if we wanted to use it in a hospital for this purpose, how would we go about doing that? That was the intention of the drive unit, which is our digital research environment.

Hannah:

It's such an interesting story. Just to get to this point, one of the things that we want to ask you about this morning is about then what what are the kinds of things in layman's terms that you're able to do with that data? But the other question that struck me is about the types of data that you're collecting. You touched on lots of different types of structured data. But there's other data that I've been thinking about is I'm not even sure if it's, so we're talking about instinct and how do you get input from or do treat this as data where it's input from parents or children themselves. Where they tell you what they're thinking or what they are seeing? I only ever thought about it as instincts, but actually, that's also data, is that something that you consider as well.

Neil Sebire:

So it's really interesting area, and it's, this is going to be a huge thing. And it's kind of in its infancy at the moment. So there are a couple of elements to this. So the first thing is, can we take what is considered current unstructured data and do something with it to make it more objective and structured. And a real tangible example of that is a parent at the moment may describe what a child's seizure looks like. That's quite hard to sometimes get across what actually happened. So anecdotally, what some parents may do is they may film the seizure or the event on their phone, and then they'll kind of show it to the doctor. So one of the ways is if you can you start to use that so can we are there ways of capturing things like video and objectively doing things with it. The example that is further ahead than most at present is things like gait analysis, you'll all be aware that you can take a video of how someone is walking, and that's relatively easy to analyse. So from a technical perspective, it should be relatively easy to analyse any movement or posture. It's just that we haven't taught the systems to do that yet. So the first bit is around taking things that currently are not structured and making them as structured as possible. The next part is around taking what someone says, and saying, Can we get insights and structure from that? That's the huge area of kind of natural language processing and text analysis, etc, which is going to be massive in healthcare as it is in everything else and has all of the issues that I'm sure you're aware of. But that is, again, a big area for us. So we are looking at how is the best way of doing that. Currently, it's a bit like the wild west for this sort of, I'm going to bundle this all up as NLP. So that's natural language processing approach, because there are many companies that have tools that can do this. What we don't understand yet is if you take the same corpus and put it through the multiple companies, do you get the same things out? Do you get the same insights? Is the sentiment analysis the same across these? If so, which one is more appropriate in certain circumstances? So no one's done that work yet. Then the last part is much more general. This is really my my core interest in a ways around trying to standardise the data. You would think that there are clear standard ways of recording what happens in a hospital. That's not the case. I'm happy to give you some examples of that from some of the things that happened recently around COVID. But stepping a stage back to your question. Patient reported outcomes in general and quality of life measures are notoriously difficult to compare at the moment because of not having standard ways of either assessing them or reporting them. So if you have a group of children with a particular condition, and let's say that one of the things that's associated, that condition is nausea.Nausea is not recorded anywhere, because it's not a thing. So it will be possibly written in handwritten notes by the doctor or the nurse. But trying to evaluate how nauseous someone is, it's relatively hard. It's relatively hard because no one has got together and said, look, regardless of whether this is because of your chemotherapy, or because of your underlying gastro disease, or whatever, how about, we all come together and try and agree a framework around the severity of nausea that we will all use, regardless of that condition. Then you can start to do some much more clever things around comparing whether a certain drug gives you more nausea than another drug. But then again, we can't do any of that, because there are no standards around patient reported and quality of life measures. Personally, I think that will be an absolutely huge thing that is going to come out of, in fact, the COVID 19 emergency, because all of the focus at the moment in those studies has been around survival. Given that most patients survive, the next phase will be what happens to the survivors. And I think people are now starting to realise that we don't have good structures to evaluate anything that's slightly longer term and doesn't have a neat end point, like yes or no did you survive, but the quality of life impacts are likely to be much more profound and long term. That's a slightly long winded answer, I'm afraid. Sorry about that.

Hannah:

No, no, no, it's good. Yeah. I mean, you took us on to another area when you touched on COVID-19. We're recording this podcast with you today when we've gone past the first peak of the pandemic. We were curious to understand it's a different experience for you in a children's hospital, but what kind of impact that has had in terms of how you've been able to care for children and their parents being around them during care and all of those types of things that previously would have been business as usual.

Neil Sebire:

Yeah, so it has had an impact on us. Fortunately, COVID-19 affects children much less severely than adults. So our hospital was affected less in terms of the actual proportion of patients infected. There was of course, this paediatric inflammatory multi system syndrome, which a small number of children get, which obviously meant that we had a disproportionate number of those cases. But by far, the biggest impact for us has been the impact on our services for other patients. So, because what happened in London is that lots of the general hospitals in London that take adult patients and paediatrics were full with dealing with COVID patients. Therefore we were taking a lot of the paediatric patients that they would have had to free up beds, etc. But that then meant that lots of our patients that we would have operated on electively weren't coming in. It meant that a lot of other patients who would have come to outpatient clinics were either having video calls or remote consultations, but weren't having procedures carried out. So while some things like for example, cancer, chemotherapy, and cancer surgery pretty much carried on, there was a large drop in lots of our other activity, a large drop in elective imaging and all sorts of other things. So it is highly likely that the biggest impact on children of the COVID-19 emergency will be on the knock on bystander effects on other diseases. And we're only just starting to understand those now. So undoubtedly, there will be some cases of children who would have presented to their own doctor or an A&E department and turned out to have a malignant diagnosis. Who haven't presented because there was fear of catching COVID, or whatever it happens to be. So it has had an impact. The other thing, though, that has been beneficial, I think, has been the fact that, if COVID had come this time last year, it would have been very difficult for us because we would have been right in the process of implementing and switching on the new electronic patient record system. But actually, as it was this year, we had the EPR data, we had the EPR data from the previous year. We had the DRE platform, which meant that actually, for example, we within a day really had spun up a COVID-19 workspace within our DRE platform where every single patient that was then tested for COVID, whether it's negative or positive. We could de-identify their data and then look at the aggregate data in this workspace in real time. So we could see immediately what types of patients they were, what were their underlying vulnerabilities, was the kratom in distribution different between those that are positive versus negative. That was simply something that would have taken six months before, that literally took a couple of days, because you have the infrastructure. The sort of spin off thing from that is that you then realise that you can do that for things that are not COVID. So in fact, if you now wanted to look at, you can use the infrastructure you have, you can use the code, the visualisations, the statistical models, the shiny app, all of the stuff you've built. That's what we're doing now reporting that into other things. So we've actually ported that across into our patients with chronic renal dysfunction. So now you can literally just press a button and get daily updates on all of the lab results, all of the imaging, average time to next admission for any cohort you want. That's simply because if you can present data in a standardised way, in a standardised format, that the analytics tools can understand everything you build then can become scalable. As soon as you have to build something for one particular purpose, which is what we would have done before, we'd have ended up having the same results from the COVID thing, but we would have built it for that purpose, it would have taken a lot of time. We wouldn't have been able to easily use it for any other purpose. So in fact, it has made an impact. I think it's suddenly made the hospital also lots of the hospital kind of managers etc, suddenly realise, wow, Can you do that for us? I think so in a strange way, I think it's actually probably been quite beneficial in terms of data.

Hannah:

So to clarify, I'm thinking, what does this mean for a child or a parent or a clinician treating a child in layman's terms. What's going to happen differently now that wouldn't have happened before you had this learning?

Neil Sebire:

So there's, I think there's a couple of things. So on the specific point of the COVID, for example, it allowed us to quite quickly look at what proportion of our patients that were testing positive or in the different vulnerability groups that the government had produced. We saw relatively early that this generic, sort of category of immuno deficiency, we weren't actually seeing an increase in patients. In fact, that allowed us to then decide to manage the patients according to their underlying disease, which rather than put all the COVID patients together because the main complications for those patients are based on their underlying disease rather than and that's now been born out in other studies that are now set to come out in adults as well. So we're now realising that it's not immunodeficiency as such, it's particular types of immunodeficiency. So, for example, patients who were on steroids, actually, that's not that bad and in fact, some of the data would suggest is quite good. So those kinds of things around how we manage the hospital throughout that allowed us to get insights much quicker than we would have got otherwise. But I think more generally, what's going to happen from this is that it's the ability now to, firstly, to be able to present data in a very digestible way to a clinician in clinic. So that's the next bit for us is to be able to say, can we start presenting this aggregate data, so there's nothing on individual patients that a clinician will see, but they can see, on average, patients with this condition. What is their haemoglobin level compared to normal, what is their, whatever it happens to be, and that will, in the first instance, allow some insights from the conditions but my ultimate aim is to not to avoid the clinicians but to bypass the conditions to be able to get the insight before they even need to look at it. So once you can start to then build on some machine learning tools, you can start to flag and say, well, this interesting, because normally, patients with this condition, this is what this result normally looks like. But in these cases, it's higher or it's lower, or whatever. So we can start to get insights. For me that that's the ultimate exciting thing for you secondary use of data. it's moving away from using data to help clinicians to do what they do now a bit quicker, which is what everyone's sort of focusing on. That's okay, but a bit dreary, the real step change comes when you can say, can you start to get insights and understanding about disease from this data that you would never have got as a human? And that suddenly is what's going to translate into better outcomes for patients. Can you stratify people better, and that's going to be a very core thing. So what we tend to do at the moment, generally in all of healthcare is a patient is given a diagnosis. So a child will say that you have been diagnosed with so and so's syndrome. You're then in a bundled group of children with so and so syndrome. In fact, it is highly likely that for every single disease, depending on what other genetic mutations you have, your age, your gender, all sorts of other things, there will be sub clusters within that main group. If you can start to identify those sub clusters go well these patients who present at this age and are female, actually, they do better if you do this with them, rather than the standard treatment. That is an impossible question to answer with the old fashioned way of doing medical research. Because the old fashioned way is to have a hypothesis and test it. So you can test it and you say, no, that hypothesis was wrong, but it doesn't give you any new insights. Whereas the ability to look at all of the data and say find me associations, you can start to pull out those kinds of things. Now we're, we are still a reasonably long way away from

David:

Still to Come. We talked to Neil more about Dr. Was that because actually, computationally it's quite complex and one of the other issues is around how patients get labelled is a really difficult area. inadvertently asking him his least favourite question to answer and finally finding out what he would do if he could change anything in paediatric care. The purpose of drive, as I understand it, is really looking improving patient outcomes and experience. We've talked a lot about the data and to a greater or lesser degree, many hospitals, as you've pointed out, all industries are looking to work out how they can utilise data at the minute but why is it so important to do it with children in a children's hospital?

Neil Sebire:

There's a couple of things I think firstly, it's that if you are wanting to use a new technology, in terms of how you're going to interact with patients and parents or how you're going to look at a device, for example. Then you need patients or consumers in this sense, to want to use the thing. In fact, so children's hospitals are ideally suited because their population, both of their patients and their parents are young, so they tend to be tech literate. In general, parents are highly motivated about their child's health. They will be more motivated about their child's health and their own health. Certainly when they're older, they will be more motivated about their child or grandchild health on their own. So there's lots of tangible advantages about having the population of a children's hospital. The second advantage is that you can play a bit more. So there is this thing of working out around engaging, so how might you engage with patients and their families around a type of monitoring or giving information about their disease and how could a child if you want to try and get, you know, going back to the official term of patient reported outcomes. If you want to find out from a child who's five, how bad their tummy ache is, how's the best way of doing that probably isn't to have some dreary Likert scale, which is what we tend to do for adults, probably is best to do it with some kind of pictures or with some dial they can turn or some something else. That ability to, I mean this in a positive way to play a little bit to be a little bit more kind of a little bit more wide thinking about how you might do the interaction is much easier in a children's hospital. Then just the final bit is that whilst many children have rare and complex diseases, in general, the disease's tend to be more specific and more focused than in adults. So the difficulty is if you are an adult, and you're 75 years old and you have a bit of cardiovascular disease and you have a you've had a myocardial infarct, and you smoked and you've got some emphysema, and you've got whatever. Your breathlessness is actually quite hard to evaluate, because it could be due to almost anything of the above. If your three, your three year old is not breathless, due to any of the other things, they don't smoke, they don't do whatever, which means if they are breathless, it's either due to their lung disease or their cardiac disease. So in some ways, from the disease perspective, it's also a little bit cleaner. That is why the impact of things like genomic testing is so much more in children's hospitals. Because if you find a genetic mutation that causes this, most likely, that's the cause of the of this clinical presentation, because you haven't been smoking and drinking and doing all the other things that could also have led to a condition. So all of those things together, I think mean that there is a unique opportunity around children's hospitals. And also just to remember that what we are trying to do here and all of the hospitals that are involved in this kind of thing. Are trying to really say, how is healthcare going to look in 10 and 20 years time? That's what we're ultimately trying to do and the children that we have now are the patients of 10 to 20 years time, rather than the patients who are now 80. So it's really the ideal sort of scenario for doing this kind of work.

David:

You've preempted my next question, which was, you know, the ability, what we didn't have growing up and what our parents didn't have growing up was the ability to capture this data in a form that can then be looked at, over the course of one's life and referred back to and hopefully kind of start to help with whatever care you're going to need in the future. Whereas hopefully our children will have the ability to be able to go back to look at their records from when they were a child and use data analysis and you know, kind of help moving forward. I guess so you've touched on you, we've got the opportunity to use this in the future. But what do you fear? What do you see the future looking like? You know, at the minute, you are, as you said, you know, you've always been interested in computers. And so you work in basically, you've just created yourself a massive opportunity to continue to play with the latest and greatest technologies that are coming out with a real purpose. But what do you see that happening in the future? How do you think that's going to portray itself in the future?

Neil Sebire:

So the way that data I think, will impact healthcare, which is vastly underestimated by many of my colleagues, there are still so many people that go ahhhh, I'm a, whatever ologists it is, I'm a dermatologist, I'm not whatever, it's not gonna affect me, it is going to affect you enormously. So the first thing is, I think it's massively under estimating how disruptive this is all going to be. The way I kind of think about it is, is there are three or three kind of core areas if you like, where it's going to make a difference. The first is clinical decision support. So that is around, how do you use data to help the doctor possibly take over from the doctor, but for now help the doctor make better decisions about the care of an individual patient. Now, that could be making a diagnosis quicker, it could be stratifying care, it could be choosing a better type of drug. That's the first, there are loads of, you can split that up into lots of lots of different ways of doing this, image analysis and all sorts of things, but clinical decision support is one. The second one, I would call operational decision support, which has also been underestimated, which is, why is healthcare so frustrating for families? So if you come to Gosh and you see someone in the clinic and they say, right, we need to do a blood test now. Why does it take two hours to go and get a blood test and then you need an X ray, which you could have had whilst waiting for the blood test. But because the systems that are in place, meant that that couldn't happen. So if you look at other companies that are essentially logistics companies, which are any of the supply chain type companies, they will be flabbergasted that in any hospital, certainly almost any hospital in the UK, nobody knows where things are and by that, I mean, if you just said, where are all the patients now, where are all the beds, where are all the vials of this medication? Where are all the staff? We don't know. Because we don't track things, we don't track people, we don't track things, etc. And just from a logistics perspective, that could be transformative by just doing simple things to take what you have now and run them better. So there's a whole stuff around I think that that sort of operational decision support. Then the last part, I would put as being patient decision support, which is tools that you can then allow the patients and families to use themselves to manage their care better. The early examples of this have been things like, you know, monitors for diabetes, home diabetes monitoring, where, depending on the result, the monitor, or the system can give a nudge to work out the dose of insulin or whatever it happens to be. But there are many, many things you can think of, where actually devices that you have yourself, or your patient reported outcomes, or a urine dipstick or a Fitbit monitor. Where you don't need to see the doctor, where you can build in some relatively simple, even just rules actually about managing care better. I think we have not really started that process at all. That concept, particularly as a sort of patient decision support bit is something that is completely missing. So there are lots of health apps. Most of them are very bespoke and one of my personal bugbears for all of this stuff is the single. If someone said, why are we not doing this stuff? The single biggest problem is, we do not have well defined standards for almost anything. So if you don't have standards about how you record and transmit information, you can't do things. So imagine if every single mobile phone manufacturer had their own standard for how you make a call and how you got an app, you wouldn't be able to do anything, because you would only be able to call other people on the same phone as you, etc. But that is exactly the situation we're in with electronic patient record systems. What's happening now is you see health apps that are now coming and patient facing app. So an app that you can put some data in about your asthma or whatever it is, almost uniformly, these are also not using standards, they're coming up with their own ways of recording this information. You're ending up with just more and more silos of information. And until we do something about that as a concept, it's going to be very difficult to sort of move on with those things.

David:

If we think about So, if we have paediatricians or clinicians that are looking to move into this area, what would you say to them in terms of how they will be able to do that, or skills that they should think about acquiring?

Neil Sebire:

This is my single, most disliked question, because it's slightly embarrassing because I would really like to be able to say, you know what, because we do get contacted, particularly by medical students. The next generation of doctors, there's a whole generation that are very tech literate and code want to do informatics. So yeah, what can I do that? How do I train in this? It's embarrassing, because I have to say, well, you can't at the moment, because England, not just England, most places in the world don't have any training for this. Now, what I would say then, is that at the moment, there are opportunities can evolve through research. There are opportunities to do things like MSC's and projects. There is gradually increasing, although very rare still, funding for informatics based projects for clinicians or non-clinicians. But until fundamentally, we manage to get clinical informatics, seen as a specialty is going to be tricky. So I would urge them to contact the faculty of informatics who are essentially there to lobby, to try and get this to be recognised. But it is very frustrating, I can see that it's frustrating because there isn't a clear path. Therefore, at the moment, the people that are doing this are kind of doing it because they love it. But the flip side is. So having been a clinical academic for many years. The people who are in this space, are by far the most collaborative and generally jovial and fun groups of academic clinicians that I have met. Because almost by definition, they're all doing it because they like it. If someone in a different hospital shows something as oh, we built this thing, if it's an impressive thing, what you'll find is that everyone else goes, Wow, that's amazing. Can we have it? How did you do it? If you go to a normal academic conference, that generally does not happen. There is massive rivalry between people, which I think because it's very new, you don't see that anywhere near as much in this space. That perhaps is because it's not professionalised yet. So I think there's a flip side to these things. But in terms of just growing this more broadly, it is really important that this is addressed soon, because there is a whole generation of doctors, nurses, physios and computer scientists who wanting to work in healthcare, this idea that computer scientists will all go and want to work for a tech company and earn lots of money is actually not true. There are significant proportion that want to do that. But in fact, there are also a proportion that say, you know what, I actually like to do something that I feel is worthwhile. As long as I can earn enough, I quite like to do this for the NHS, how do I do it and until we have a clear path that we can filter all of these people from both sides. We're not going to move as fast as we could do and it is very frustrating, because there is actually quite a demand out there. It's just that we don't have a place to filter those people through.

David:

We will definitely need to get that link. So we can put it into the show notes so that people can go and lobby. In due course, Neil I feel like we could talk for, you know, hours and hours around this. But I'm also thinking that we should maybe look at doing a part two, because what we haven't touched on is actually, what are some of the solutions that you've managed to implement, at Gosh, and in other areas and what that's meant to the patient. So with, you know, maybe that's something that we can look at doing, you know, in the future. And I guess, to close from this point of view, the question that we're asking everybody is, if you could change anything in child health or paediatrics? What would it be?

Neil Sebire:

This will sound a strange answer. For people that know me, they won't think its that strange. Data standards, please standardise everything that we are doing. That's not just whenever you the trouble is, whenever you use that term, everyone kind of rolls their eyes, oh my god data and IT. But I mean that right from the perspective, as a profession, we should be teaching medical students and junior doctors, how to think properly about what words they are using, and how they are recording things. This is now starting to be recognised this sort of idea of phenomix. It's not enough to have a SNOMED CT term code that says this patient has type one diabetes, that in itself doesn't tell you anything about how the patient got the label. So who gave them that label? Was it the world expert in paediatric diabetes? Or was it that someone once thought they might have had this? How sure are we have the diagnosis? How severely are they affected? How long have they had it? What kind of complication have they had? Have they got a genetic mutation that's giving them that. So this idea that you just have a term, and that term is enough? It is just one example of almost anything you can think of in healthcare, where until we start to work out how we can systematise what we're doing and making computer readable, so that you can transmit that information to someone else and they will receive it with the same information. We are massively hobbled. That, for me will be the single thing because it's actually something we could do. It's not something that needs new things to be invented, it's not something that potentially needs a new piece of equipment is just the way that we work. It's frustrating that if you went to six or seven children's hospitals around the UK and you asked for exactly the same data out of their clinical systems. That will not be an easy task. When you eventually get the data, so forget all the government stuff, when you eventually get the data, the data will look different from all of those different hospitals. Even the same thing will probably be called a different thing, will have a different label, might be in different units, different formats. So to do what should take five minutes across six hospitals, will actually take six months.

Hannah:

I'm thinking about the impact of what you're saying there on patients and patient care. I think it sounds like it's to do with correct diagnoses or speed of diagnosis, or how a patient understands how they're being communicated to.

Neil Sebire:

It can be all of those, it can be all of those things.I think it also means that you can then start to build, we're talking about these tools like these patient decision support tools, etc. those tools only work or only work properly, when you are sure of the information that's coming out of them.It then means that if everyone was doing things in the same way, so you have this standardisation, doesn't matter which hospital you're at. Whereas at the moment, it does matter which hospital you're at, because, for example patients who come to Great Ormond Street for their specialist care but live in a different part of the country, and then go to a GP and then get admitted to their local hospital. We don't know that any of that happens routinely because all of the those three are using different systems. So even the those simple practical things around just patient care that the fact that you can transmit information for individual use of the patient, forgetting all of the more advanced stuff is all based around standardisation. Well, that's for another day I bang on about this a lot. But the trouble is people think it's a bit kind of unimportant. They want to get onto the sexy stuff. But ultimately, whatever your field is, if you can rely on the data that you're getting, and you know what it means, and you know what it looks like, you can build amazing things. If what you're getting is unreliable, and you can't trust it doesn't matter what or how amazing the thing is you've built, it won't work, it won't deliver the thing. That fundamental thing has to change.

Hannah:

Neil, well, that's been brilliant. Thank you very much. Indeed. It's been a super interesting conversation this morning.

David:

Absolutely. I think we've definitely gone into the future of what healthcare looks like. And as I said, if we can you know, get some time with you to do a part two, then I think that would be fascinating as well.

Neil Sebire:

Yeah, no, very good. Happy to do that. And it's always nice to get the opportunity to blabber on about data.

David:

Fantastic. Thanks so much. Thank you for joining us. We hope you enjoyed the

Neil Sebire:

Thanks very much. conversation with Neil as much as we did. And we're really looking forward to that part two. All details from today's podcast can be found in the show notes, including ways that you can get hold of us and also details of Drive and Great Ormond Street Hospital. If ever there is someone that you think that we should be talking to them, please get in touch and let us know. Over the next few weeks, we'll be moving out of the UK and talking to people further afield from both Denmark and Italy and understanding the innovations that they're putting in in some of their hospitals. We're really looking forward to until next time, thanks for joining