1 00:00:05,370 --> 00:00:11,160 Without further ado, please allow me to introduce our two main speakers today. 2 00:00:11,160 --> 00:00:15,870 So first up will be Dr Nick Scott RAM, MBA. 3 00:00:15,870 --> 00:00:21,080 He is the managing director of Life Sciences for Sense on Health. 4 00:00:21,080 --> 00:00:24,960 I was also director of the Oxford Martin School here at the University of Oxford. 5 00:00:24,960 --> 00:00:30,210 He's got over 25 years of experience in commercial and business developments in the life sciences sector, 6 00:00:30,210 --> 00:00:37,950 particularly of pharmaceuticals, as an MBA in natural sciences and a Ph.D. in the philosophy of Science from Cambridge University. 7 00:00:37,950 --> 00:00:43,950 He's worked in blue chip life science companies such as Oxford Biomedica, How the Jet Pharmaceuticals, 8 00:00:43,950 --> 00:00:47,820 where he was vice president for corporate affairs responsible for strategy. 9 00:00:47,820 --> 00:00:53,370 Previously, he was darch of the commercial developments of the Oxford Academic Health Sciences network, 10 00:00:53,370 --> 00:00:59,490 where he was responsible for the development of strategic partnerships with major pharmaceutical companies. 11 00:00:59,490 --> 00:01:05,310 He's led the Life Sciences Industry Position on Biotechnological Technology Patents Directive to the 12 00:01:05,310 --> 00:01:11,130 European Parliament in the 90s and in 2001 it was awarded the NBA for services to biotechnology. 13 00:01:11,130 --> 00:01:18,210 In 2001, after neck health halted his heels would come Dr Brian Heinz. 14 00:01:18,210 --> 00:01:26,040 Brian is the technical director for Global Non-woven for Campbell Kimberly-Clark and is responsible for driving key technology initiatives, 15 00:01:26,040 --> 00:01:33,210 intellectual property and developing strategic vision for the application of non-woven technologies within Kimberly-Clark. 16 00:01:33,210 --> 00:01:37,170 Prior to his current assignment, Dr Heinz has held various roles, 17 00:01:37,170 --> 00:01:45,570 including pilot facilities manager Mel Technical Team Leader and director of Frontend Innovation, so he knows everything about this during his career. 18 00:01:45,570 --> 00:01:48,030 He's obtained 43 US patents, 19 00:01:48,030 --> 00:01:57,840 16 trade secrets as the recipients of the I.A.E.A Lifetime Technical Achievement Award at the Society of Women Engineers Rob Rodney Dietrich Award. 20 00:01:57,840 --> 00:02:03,300 So the speakers are going to come one after another and after that, we're going to progress into Q&A. 21 00:02:03,300 --> 00:02:11,580 You can type your questions in the chat and they will be fed through the backstage process up to me and I will fire them off to speakers. 22 00:02:11,580 --> 00:02:16,560 So enough for me. I see that Nick has appeared on screen. Thank you very much. 23 00:02:16,560 --> 00:02:21,450 Welcome, Nick. Thanks, Phil. and good afternoon, ladies and gentlemen, 24 00:02:21,450 --> 00:02:26,070 it's a pleasure to be here and very much appreciate the opportunity to talk a 25 00:02:26,070 --> 00:02:32,430 bit about what happened to his stance on health during the COVID pandemic. 26 00:02:32,430 --> 00:02:40,110 Now there should be a slide deck that should be sort of appearing imminently in this so I can talk through some of the things that we did. 27 00:02:40,110 --> 00:02:45,180 But by way of background, I work for a company called Sunshine Health. 28 00:02:45,180 --> 00:02:51,270 We are about 150 people and our roots go back to a deep embedding within the University of Oxford 29 00:02:51,270 --> 00:02:59,460 and also the local hospital trust called Oxford University Hospitals NHS Foundation Trust. 30 00:02:59,460 --> 00:03:05,640 Next slide, please. Our model, which we set up about three years ago, 31 00:03:05,640 --> 00:03:13,260 was trying to balance a different type of relationship between industry and the health care system and particularly hospitals. 32 00:03:13,260 --> 00:03:16,770 And at the heart of that was really the notion of building a trusted, 33 00:03:16,770 --> 00:03:23,550 collaborative community that would enable us to get access to fully anonymized data and clinical 34 00:03:23,550 --> 00:03:29,550 AI to improve patient care and accelerate the discovery and development of new medicines. 35 00:03:29,550 --> 00:03:35,490 And in essence, what it was as a model that allowed NHS healthcare trusts to participate, 36 00:03:35,490 --> 00:03:41,130 become equity holders in the company to get investment in IT infrastructure, 37 00:03:41,130 --> 00:03:43,950 to support data access and analysis, 38 00:03:43,950 --> 00:03:51,580 and then to get a commercial return out of the partnerships that we would then generate through working with life science industry companies. 39 00:03:51,580 --> 00:03:59,610 And when we started this out, we really had to trust Oxford and Chelsea and Westminster as our partners. 40 00:03:59,610 --> 00:04:10,080 Next slide, please. And went out and did an initial public offering on AIM back in 2018, raising 60 million pounds. 41 00:04:10,080 --> 00:04:15,060 And since that time, when we had about access to two million unique patient records, 42 00:04:15,060 --> 00:04:21,270 that's now grown to over 8.5 million in the UK and a further nine million in the US. 43 00:04:21,270 --> 00:04:28,650 And you can see from this slide the partnerships we've built across the different hospital systems were working with the other side 44 00:04:28,650 --> 00:04:37,050 of the equation with being able to work with the pharmaceutical industry and effectively act as a docking station across NHS data, 45 00:04:37,050 --> 00:04:42,930 applying our in-house analytics capability and machine learning and computer vision expertise, 46 00:04:42,930 --> 00:04:47,550 and then partner with pharmaceutical companies, in particular therapy areas. 47 00:04:47,550 --> 00:04:56,460 So, for example, we've been working with bear and cardiovascular disease, the Brit with Bristol-Myers Squibb looking at blood cancers. 48 00:04:56,460 --> 00:05:00,000 The foundations of this is really around having close partnerships, 49 00:05:00,000 --> 00:05:05,090 and that's the key to the success in terms of being able to build up novel approach. 50 00:05:05,090 --> 00:05:12,710 So the words that Charles spoke earlier on in this, I think, are very pertinent in trying to find new models of collaboration and ways of doing this. 51 00:05:12,710 --> 00:05:17,720 Next slide, please. As a business, we operate across two domains. 52 00:05:17,720 --> 00:05:24,470 One is health care facing where we started out with a number of remote patient monitoring apps. 53 00:05:24,470 --> 00:05:30,380 A number of those which really were birthed within the University of Oxford collaborating with the 54 00:05:30,380 --> 00:05:37,460 hospital system while the first one for gestational diabetes through Lucy MacKillop and Lionel Tarasenko. 55 00:05:37,460 --> 00:05:39,230 And then following that through Send, 56 00:05:39,230 --> 00:05:47,090 which was a vital signs technology platform through Peter Watkinson at the hospital and again with Lionel Tarasenko. 57 00:05:47,090 --> 00:05:56,300 We then added onto that clinical algorithm engines, which is really about developing real time support mechanisms for healthcare practitioners, 58 00:05:56,300 --> 00:06:02,990 and we see this offer back into the health care systems where we don't make a profit out of this work 59 00:06:02,990 --> 00:06:09,110 as being key within the partnership that we have with different hospitals on the life sciences side. 60 00:06:09,110 --> 00:06:15,080 We focussed really around clinical AI enabled clinical trial optimisation and R&D of new medicines, 61 00:06:15,080 --> 00:06:21,590 particularly focussing around how you can actually optimise and improve clinical development in different areas. 62 00:06:21,590 --> 00:06:29,240 Next slide, please. At the point when the pandemic really started to hit in the UK, 63 00:06:29,240 --> 00:06:36,680 I remember very clearly about a week before we went into lockdown thinking that things were really going to change very, very dramatically. 64 00:06:36,680 --> 00:06:43,310 And as a company, we took very rapid steps in terms of ensuring that we could still continue to operate. 65 00:06:43,310 --> 00:06:51,590 One of those was being able to actually work remotely with the NHS data, which have been fully anonymized onto an information governance procedures. 66 00:06:51,590 --> 00:06:57,650 And we were grateful being able to do that. We also transitioned as a company completely to remote working. 67 00:06:57,650 --> 00:07:05,330 And what we found interesting was that actually this was very successful and as a company now we have a permanent remote working policy. 68 00:07:05,330 --> 00:07:12,080 So we have colleagues in different parts of Europe, in the US and further afield who all work remotely. 69 00:07:12,080 --> 00:07:16,970 And we have found that an effective way of working the two other main domains that we're 70 00:07:16,970 --> 00:07:21,290 really focussed on was what can we do in terms of helping the public and communities? 71 00:07:21,290 --> 00:07:26,030 And then also, equally importantly, was what can we do to help the NHS? 72 00:07:26,030 --> 00:07:30,200 And in this talk, I wanted to concentrate on a couple of these areas. 73 00:07:30,200 --> 00:07:36,920 The first is in Sign Cove, which is new near real time prediction of COVID 19 patients in the NHS. 74 00:07:36,920 --> 00:07:44,390 On the second is in Magnifier, which has a lateral flow reader technology platform that we developed as well. 75 00:07:44,390 --> 00:07:49,700 It's also fair to say that we worked closely with the NHS in terms of the clinical needs and the requirements, 76 00:07:49,700 --> 00:07:55,280 and we were very grateful for the continued engagement with our trust partners during the pandemic. 77 00:07:55,280 --> 00:08:04,190 Next slide. So this just touches on an app that we developed for the community, which was ready for remote monitoring of patients symptoms, 78 00:08:04,190 --> 00:08:11,270 quote CVM Health and here really, it was a question of being very nimble to build this out very, very rapidly. 79 00:08:11,270 --> 00:08:19,190 And that was done in a matter of weeks going into months to actually build the architecture and then build this up, which we then rolled out. 80 00:08:19,190 --> 00:08:27,110 Next slide, please. The other component was with the North Gestational Diabetes app, 81 00:08:27,110 --> 00:08:33,590 which is for remote monitoring of mothers in their third trimester with gestational diabetes. 82 00:08:33,590 --> 00:08:39,590 We took the decision very early on to offer this free to all NHS trust partners. 83 00:08:39,590 --> 00:08:47,720 And as that was rolled out over the last year or so, that's now been used in over 50 percent of trusts within the UK. 84 00:08:47,720 --> 00:08:52,550 So it's a fantastic testament to the team that we're actually responsible for doing this, 85 00:08:52,550 --> 00:08:57,680 but also the fact that we thought we could make a difference in supporting. 86 00:08:57,680 --> 00:09:04,550 A component on a pathway and make a difference in terms of the outcomes for diabetic mothers in the UK. 87 00:09:04,550 --> 00:09:08,090 Next slide. 88 00:09:08,090 --> 00:09:15,830 One of the key areas that we then started to focus on, and this started out very early on in relationship to one of our other trust partners, 89 00:09:15,830 --> 00:09:21,980 Chelsea and Westminster, where we were approached by one of the leading ICU consultants, 90 00:09:21,980 --> 00:09:27,800 which was how can we predict the only progression of disease and patients that are admitted into 91 00:09:27,800 --> 00:09:33,470 a hospital and try and understand what their clinical pathway through the hospital would be? 92 00:09:33,470 --> 00:09:37,670 And how can we actually manage that within the hospital setting? 93 00:09:37,670 --> 00:09:48,410 So when patients are presented, obviously, although in a hospital system about 14 percent, you may go into hospitals suffer severe symptoms. 94 00:09:48,410 --> 00:09:52,400 Of those, six percent would require critical care. 95 00:09:52,400 --> 00:10:00,950 And one of the challenges really early on in the pandemic was the availability of resources, access to mechanical ventilation. 96 00:10:00,950 --> 00:10:06,140 And there was evidence that really showed that the prolonged waiting time between the emergency department patients would 97 00:10:06,140 --> 00:10:13,850 be admitted and that admission into intensive care may lead to delayed intensive care treatment and increased mortality. 98 00:10:13,850 --> 00:10:15,170 Next slide. 99 00:10:15,170 --> 00:10:25,190 So we set out really to identify a solution that would enable better prediction to reduce waiting times for those patients which were most at risk, 100 00:10:25,190 --> 00:10:28,310 and then finding a way in which you could actually track that journey through the 101 00:10:28,310 --> 00:10:33,770 hospital and be able to then for the clinician to side on the basis of that prediction, 102 00:10:33,770 --> 00:10:40,700 which patients should be moved into intensive care and subsequently onto a ventilator if required. 103 00:10:40,700 --> 00:10:46,490 And on the back of that, we developed two app solutions, or clinical decision tool solutions. 104 00:10:46,490 --> 00:10:48,440 The first was signed Cove, 105 00:10:48,440 --> 00:10:56,360 which really was an algorithm that predicted three potential clinical outcomes for COVID 19 positive patients that was risk of intensive care, 106 00:10:56,360 --> 00:11:02,660 admission, risk of invasive mechanical ventilation and risk of in-hospital mortality. 107 00:11:02,660 --> 00:11:07,910 And then the second was an operational algorithm that provided near real time decision making 108 00:11:07,910 --> 00:11:13,430 support to hospital managers coping with the pressures and resources within the hospital. 109 00:11:13,430 --> 00:11:21,540 Next slide, please. This just gives you an overview in terms of the clinical question we were asking, 110 00:11:21,540 --> 00:11:29,100 which is kind of COVID 19 patients early clinical features predict the risk of severe disease and then 111 00:11:29,100 --> 00:11:34,890 in terms of the data that we access through the clinical data from any admission using that data, 112 00:11:34,890 --> 00:11:43,410 cleaning, curating it, applying that clinical algorithm approach to that and then being able to predict a risk of ICU admission, 113 00:11:43,410 --> 00:11:46,770 risk of ventilation and risk of mortality. 114 00:11:46,770 --> 00:11:55,620 And this was something that we started off at a weekend hackathon and very, very quickly moved into a prototype product within the following weeks. 115 00:11:55,620 --> 00:12:07,450 Next slide. So if you could just share this, this is a video with the leading clinician from Chelsea in Westminster. 116 00:12:07,450 --> 00:12:17,230 What's around the 27th of January? And I remember texting one of my colleagues from the clinical informatics team and I said, 117 00:12:17,230 --> 00:12:22,330 Look, it seems like, you know, this pandemic may affect us all. 118 00:12:22,330 --> 00:12:31,690 Part of it was the uncertainty. We really had no concept of the numbers of patients would appear with these symptoms, the severity of those symptoms. 119 00:12:31,690 --> 00:12:41,660 The impact on our own staff while we're trying to achieve our point was to prevent progression of disease because you get patients earlier, 120 00:12:41,660 --> 00:12:46,540 you intervene and you improve outcomes simply like that. 121 00:12:46,540 --> 00:12:56,590 The key here was, I know your risk of dying, your race of needed intensive care or your risk goal need mechanical ventilation. 122 00:12:56,590 --> 00:13:04,780 What's these machine learning algorithms can do is to take vast quantities of data and give it back to the clinician as quickly as possible, 123 00:13:04,780 --> 00:13:07,420 because that could be the difference between life and death. 124 00:13:07,420 --> 00:13:14,710 I don't think that artificial intelligence or algorithms will change the way that we practise medicine. 125 00:13:14,710 --> 00:13:21,940 What it does enhances it. Thanks. 126 00:13:21,940 --> 00:13:24,210 Next slide, please. 127 00:13:24,210 --> 00:13:31,680 So that just gives us sort of a snapshot and what was very interesting in this whole experience where we were developing the algorithm, 128 00:13:31,680 --> 00:13:40,530 testing it and validating it was really the close partnership we had with colleagues at Chelsea and Westminster on the ITU group. 129 00:13:40,530 --> 00:13:48,930 This enabled us to develop the algorithm, test it live within the hospital, look at outcomes and then do further work around validation. 130 00:13:48,930 --> 00:13:53,610 And what we found actually with the algorithm was that there was a increase in predicting 131 00:13:53,610 --> 00:14:00,460 performance by 24 points against clinical standards with a five fold increase in precision. 132 00:14:00,460 --> 00:14:10,740 And this then enabled the clinicians there to use that support tool for their clinical practise that they were delivering next slide. 133 00:14:10,740 --> 00:14:15,990 So some of the benefits around this work really were improving patient based clinical care. 134 00:14:15,990 --> 00:14:20,250 So earlier risk prediction would improve potentially clinical outcomes. 135 00:14:20,250 --> 00:14:23,400 All of this was run on a cloud based architecture, 136 00:14:23,400 --> 00:14:32,880 which was something that we were developing the Chelsea and Westminster and enables actually a much more rapid turnaround in terms of data analytics. 137 00:14:32,880 --> 00:14:38,490 This went through a regulatory process and also it informed that decision making as well 138 00:14:38,490 --> 00:14:45,120 all done within strict information governance requirements from the NHS side of this. 139 00:14:45,120 --> 00:14:54,390 And since that time, we've gone to work with the hospital in terms of developing improving this as the clinical pathways modified itself as well. 140 00:14:54,390 --> 00:15:02,770 Next slide. At the same time, in terms of the risk protection stratify patients, 141 00:15:02,770 --> 00:15:08,680 this is also being used as a support for operational decision makers to support understanding the resourcing, 142 00:15:08,680 --> 00:15:13,990 so you can then look at the expected numbers of admissions to patient to intensive care of patients. 143 00:15:13,990 --> 00:15:19,510 The number of patients that would require ventilation or the expected outcomes in terms of mortality, 144 00:15:19,510 --> 00:15:23,440 and the algorithm enables one to do that virtually in a real time setting. 145 00:15:23,440 --> 00:15:32,030 Next slide. And what we found with this was really that the hospital clinicians, 146 00:15:32,030 --> 00:15:38,240 administrators found this really an invaluable tool in terms of resource planning, and this is one of the benefits really. 147 00:15:38,240 --> 00:15:42,260 I think clinical decision support tools not only those related to COVID, 148 00:15:42,260 --> 00:15:50,360 but in other disease areas to really support and optimise planning in terms of patient journeys and what's actually required in doing that. 149 00:15:50,360 --> 00:16:02,820 Next slide. Turning now to a second development that we did, and again, this was done very much at a pace within the business and as a company, 150 00:16:02,820 --> 00:16:07,310 we have about 150 people working across different disciplines. 151 00:16:07,310 --> 00:16:13,920 And this was another example of bringing together different disciplines within the company to identify a problem 152 00:16:13,920 --> 00:16:19,830 which had been highlighted by the government and then to build a solution that actually could address that problem. 153 00:16:19,830 --> 00:16:24,690 And this was specifically with lateral flow testing for COVID 19. 154 00:16:24,690 --> 00:16:30,090 The number of you probably will have experienced using these lateral flow tests on one of the questions was actually, 155 00:16:30,090 --> 00:16:36,870 was there a way that we could actually automate and improve the reading performance of those tests against different backgrounds? 156 00:16:36,870 --> 00:16:43,260 And also then ultimately to be able to actually upload those results into the NHS? 157 00:16:43,260 --> 00:16:50,370 And this project started back in December and was really based on the computer vision 158 00:16:50,370 --> 00:16:54,450 team that we had within the company who had been doing more traditional imaging work, 159 00:16:54,450 --> 00:16:58,770 looking particularly around pathologies and lung disease. 160 00:16:58,770 --> 00:17:03,750 But very quickly they adapted themselves actually to work in a totally different domain, 161 00:17:03,750 --> 00:17:08,280 which is to develop novel algorithms to support better reading. 162 00:17:08,280 --> 00:17:15,630 And this really is a highly magnified, which is the product that was developed as a highly accurate set of deep learning algorithms, 163 00:17:15,630 --> 00:17:23,280 which really automates the reading lateral flow tests and enables you actually to do this beyond what the human eye could actually see. 164 00:17:23,280 --> 00:17:26,100 And also at the same time, this was a very, very fast, 165 00:17:26,100 --> 00:17:35,160 secure and scalable solution and enabled one actually to outperform the human by detecting faint lines and improve actually those readings. 166 00:17:35,160 --> 00:17:41,610 And we started working with a range of different datasets to actually prove this once we develop the algorithms. 167 00:17:41,610 --> 00:17:49,540 Next slide, please. One of the things at the start was to actually work out what the architecture was. 168 00:17:49,540 --> 00:17:58,990 And we developed a system which was able to read 500 plus tests per second and also sub to second machine to machine round trip response, 169 00:17:58,990 --> 00:18:04,630 which is pretty important. And also with the scalability to handle up to 40 million tests per week. 170 00:18:04,630 --> 00:18:11,710 So this was the fundamental building blocks were put in place, which we did in partnership with Microsoft on their zero platform. 171 00:18:11,710 --> 00:18:21,990 Next slide. What we then put together was a range of eight different algorithms built specifically for the lateral flow reading, 172 00:18:21,990 --> 00:18:26,880 which really would enable us to identify test which were positive, negative or void. 173 00:18:26,880 --> 00:18:31,440 There was also an anomaly or fraud detector and then quality control algorithms, 174 00:18:31,440 --> 00:18:37,620 which actually would enable one to actually review the brightness and actually understand what the test was of 175 00:18:37,620 --> 00:18:43,080 the context of the background lighting to be able to then understand whether you had a positive or a negative 176 00:18:43,080 --> 00:18:48,840 on that and that really sat at the heart in terms of the innovation that was brought into this for something 177 00:18:48,840 --> 00:18:54,840 that superficially similar looks quite easy in terms of getting better precision and lateral flow testing. 178 00:18:54,840 --> 00:19:01,980 But recognising that if you have to do that from lots of different backgrounds, then that is something that is much, much more complex. 179 00:19:01,980 --> 00:19:10,350 Next slide, please. At the same time, we found with this that actually we were able to support identity fraud as well, 180 00:19:10,350 --> 00:19:15,060 so that if it actually trained to seek out those things where there been some 181 00:19:15,060 --> 00:19:19,740 lines that were abnormal or fraudulent in the control and or the test line, 182 00:19:19,740 --> 00:19:26,160 and that's something then that could be uploaded using a QR code, which is an important part of this test. 183 00:19:26,160 --> 00:19:33,960 Next slide. So what we found really was in being able to easier read the test. 184 00:19:33,960 --> 00:19:40,440 Then there's the potential to use Magnifier to increase the use of lateral flow testings in a home setting or other 185 00:19:40,440 --> 00:19:47,400 settings that then gives greater confidence as those tests are done in terms of the readings that are generated. 186 00:19:47,400 --> 00:19:54,780 And this is something that we've seen is very, very significant. And the building and the development of this, which started in December, 187 00:19:54,780 --> 00:20:02,310 took us two to three months and really again comes back to having multidisciplinary teams coming together very, 188 00:20:02,310 --> 00:20:08,310 very quickly, having very strong project management and a very clear idea of what the end goal is. 189 00:20:08,310 --> 00:20:13,560 Charles earlier on was talking about what the objective is another example of actually 190 00:20:13,560 --> 00:20:19,410 having a shared objective that really everyone could work to and identify with. 191 00:20:19,410 --> 00:20:29,770 Next slide, please. We also found that we could build in this COVID 19 app for machine learning, for lateral flow testing. 192 00:20:29,770 --> 00:20:39,640 And this also working this of mobile phones was an important component of what we were doing with this next slide. 193 00:20:39,640 --> 00:20:44,740 Finally, then with this, there's the opportunity to create a digital certificate, 194 00:20:44,740 --> 00:20:49,390 which, if required, would support organisations to keep workplaces and events safe. 195 00:20:49,390 --> 00:20:53,650 And also in terms of understanding epidemiological trends as well. 196 00:20:53,650 --> 00:20:59,590 That comes with training videos and questions to ensure that the testing is done properly. 197 00:20:59,590 --> 00:21:09,130 As with all of this, the focus is on having proper regulatory framework around GDPR compliance on a quality management system. 198 00:21:09,130 --> 00:21:16,950 We see it's been absolutely essential and integral to the work that we've been doing on this next slide. 199 00:21:16,950 --> 00:21:21,600 All of this then gives the confidence to enable secure reporting, 200 00:21:21,600 --> 00:21:27,030 and then that is something that ultimately public bodies would be able to receive these reports and 201 00:21:27,030 --> 00:21:34,470 also can have importance in terms of supply chains into manufacturers and inventory management as well. 202 00:21:34,470 --> 00:21:43,730 Next Slide. So the final piece around this, which I've mentioned a little bit earlier, was around the regulatory piece, 203 00:21:43,730 --> 00:21:48,800 and there was a great deal of effort done by our regulatory team to ensure that the certification 204 00:21:48,800 --> 00:21:54,410 and regulatory compliance in this in terms of security controls over data handling and processing, 205 00:21:54,410 --> 00:21:58,460 tracing and auditing were actually done as well. 206 00:21:58,460 --> 00:22:05,090 And the magnifier has been approved for use for lateral flow tests, for professional use testing across Europe. 207 00:22:05,090 --> 00:22:12,170 And also it's being worked with. The test to go up is anticipated to receive regulatory approval in the future. 208 00:22:12,170 --> 00:22:18,690 So all of this was built into the whole process going forward. Next slide, please. 209 00:22:18,690 --> 00:22:28,350 So that's just sort of giving a snapshot on two particular projects that we undertook within the company in very short order. 210 00:22:28,350 --> 00:22:32,280 And this was really done against the backdrop of remote working. 211 00:22:32,280 --> 00:22:41,370 So everyone be familiar with back to back teams meetings, but really required a deeper collaboration within different disciplines with the company, 212 00:22:41,370 --> 00:22:46,170 within the company and also critically with our NHS partners. 213 00:22:46,170 --> 00:22:55,530 That really enabled us to do these things. Some of the lessons that came out of this for us was very much a greater emphasis on project planning, 214 00:22:55,530 --> 00:22:58,440 the importance of defining roles and responsibilities, 215 00:22:58,440 --> 00:23:04,200 which then flow into a shared vision of what the target output of everything that is being done. 216 00:23:04,200 --> 00:23:13,350 I think also what I witnessed was really a willingness and adapt to modify different development pathways and timelines repeatedly. 217 00:23:13,350 --> 00:23:21,060 So what we found in all of this is that the goalposts did change slightly because the disease progression, the impact of it, 218 00:23:21,060 --> 00:23:29,160 the decisions in terms of how things were managed also had an impact both within the NHS and in some of the other work that we did. 219 00:23:29,160 --> 00:23:33,210 I think one of the bigger things from a health care system perspective is that the 220 00:23:33,210 --> 00:23:40,290 COVID pandemic has shone a light on on the importance and impact of digital solutions. 221 00:23:40,290 --> 00:23:46,500 And certainly we saw a major sea change within our trust partnerships in terms of recognising now digital 222 00:23:46,500 --> 00:23:53,010 solutions could really have a future understanding of what could be done better within a healthcare system. 223 00:23:53,010 --> 00:24:01,620 So a shift from looking at things that hospital or moving to the hospital at home are all part of this where actually remote monitoring, 224 00:24:01,620 --> 00:24:10,080 vital signs tracking stay at home. A better understanding of patient tracking enables then clinicians to have more confidence 225 00:24:10,080 --> 00:24:14,790 in the decisions they make in terms of patient treatment is really the way forward, 226 00:24:14,790 --> 00:24:21,360 and this will continue to develop and grow. I think one of the challenges that remains within healthcare systems, though, 227 00:24:21,360 --> 00:24:26,580 is the disconnect between who benefits from the products, who pays for them, 228 00:24:26,580 --> 00:24:31,260 and there is a lag in the system in terms of changing those perspectives, 229 00:24:31,260 --> 00:24:36,960 not only at the system level, but also where you're trying to integrate new product solutions. 230 00:24:36,960 --> 00:24:42,390 And I think within the NHS, although great strides have been made in this direction, 231 00:24:42,390 --> 00:24:48,810 there's an awful long way still to go in terms of how to connect up the funding benefits with the clinical benefits, 232 00:24:48,810 --> 00:24:57,690 with the patient benefits across multiple stakeholders. And I think over the coming years, that's going to be an area of great concentration. 233 00:24:57,690 --> 00:25:02,580 The other thing in the final thing to say, and this is when there's a collective need and the collective requirement, 234 00:25:02,580 --> 00:25:07,050 it is incredible what people can do in terms of moving the dial and moving very, 235 00:25:07,050 --> 00:25:12,570 very quickly on the speed that we had in terms of building algorithms and apps. 236 00:25:12,570 --> 00:25:18,150 Validating and testing them within the business with our partners was something that was really incredible, 237 00:25:18,150 --> 00:25:24,210 and I really give great thanks to our partners in terms of the work and the commitment that they had. 238 00:25:24,210 --> 00:25:32,190 So for me, I think for the company changing to remote working, but being able to really partner and have a deep, 239 00:25:32,190 --> 00:25:42,870 engaged partnership is really been the key in the fundamentals to making some changes in the pandemic and the impact of digital solutions. 240 00:25:42,870 --> 00:25:50,357 Thank you very much.