1
00:00:00,540 --> 00:00:03,750
I say to them that you need to turn that off.

2
00:00:04,620 --> 00:00:09,180
So can you just, first of all, say your name and your current title?

3
00:00:09,930 --> 00:00:19,500
Yep. So my name is Thomas Kirk and I am now a research fellow at the University of Nottingham in biomedical engineering, specifically imaging.

4
00:00:20,340 --> 00:00:29,010
Terrific. Thanks very much. So without telling me every minute of your life story, can you just give me the headlines on your your career so far,

5
00:00:29,010 --> 00:00:37,350
starting really from how you first became interested in And I don't know whether the life sciences or the engineering side of it came first,

6
00:00:37,350 --> 00:00:48,299
but in those subjects. So I was I was always interested in engineering from from a relatively young age,

7
00:00:48,300 --> 00:00:54,270
but not interested in the sense that I knew from a long way out, I want to go to university and I want to be an engineer.

8
00:00:54,600 --> 00:00:58,110
I know some kids have not quite well formed life plan.

9
00:00:59,070 --> 00:01:04,530
That wasn't the case for me. However, when it came down to applying for university,

10
00:01:05,430 --> 00:01:11,430
I was at a very good school where pretty much everyone goes to university, so I was just going with the flow, as it were.

11
00:01:12,060 --> 00:01:20,130
And but that's the that was the first time where I seriously considered that I could go to Oxford or Cambridge, and now I had the grades to do that.

12
00:01:21,270 --> 00:01:28,649
I applied to Cambridge and again, actually I failed by interview and then I spent a year working as a maths teacher on

13
00:01:28,650 --> 00:01:36,389
a teach first scheme and then I reapplied fogs at this time and did much better.

14
00:01:36,390 --> 00:01:44,459
And I think the teaching experience I was doing probably helped a lot because my the interviewer at Trinity College,

15
00:01:44,460 --> 00:01:51,900
where I played in Oxford, his wife is a teacher, so we did spend part of the interview talking about the challenges of teaching.

16
00:01:51,900 --> 00:01:55,790
And I think that that probably helped put me at even. Anyway.

17
00:01:56,180 --> 00:02:02,239
And Oxford is a is a special place to do engineering in the UK like Cambridge.

18
00:02:02,240 --> 00:02:10,070
But not many other universities share this property of offering a general engineering course where you specialise as you go on.

19
00:02:11,660 --> 00:02:15,530
And initially biomedical didn't feature on my radar at all.

20
00:02:16,130 --> 00:02:22,290
I thought I'd probably end up as a civil engineer and then after having done one year of general engineering at Oxford,

21
00:02:22,290 --> 00:02:31,130
realised I absolutely hated civil engineering engineering, I ended up quite liking electrical, which I never thought would happen.

22
00:02:32,150 --> 00:02:42,559
And then that was the influence of one tutor in particular called Michael Chappell, who I remain working with now at Nottingham.

23
00:02:42,560 --> 00:02:45,740
He supervised my PhD and now I work with him at Nottingham.

24
00:02:47,420 --> 00:02:53,270
He is the person that steered me towards biomedical engineering as I went through my degree.

25
00:02:53,960 --> 00:03:01,640
So in second year he started tutoring me for control and then he supervised me on a small summer

26
00:03:01,640 --> 00:03:06,680
project in third year and then he supervised me for my Masters Year project in fourth year.

27
00:03:07,010 --> 00:03:12,139
And then he invited me to apply for a PhD with him and so on and so on.

28
00:03:12,140 --> 00:03:15,530
And here we are eight years later, and I'm still working with with Michael.

29
00:03:16,160 --> 00:03:21,590
And the subject to be the subject of my PhD is.

30
00:03:22,040 --> 00:03:30,380
So we work in the field of MRI imaging, which is a particular strong point of Oxford, particularly for the brain.

31
00:03:31,490 --> 00:03:41,000
And my Ph.D. was working with new types of MRI, in particular one called arterial spin labelling.

32
00:03:42,730 --> 00:03:53,530
And it's it's a very promising modality for investigating the function of the brain and moving towards neurodegenerative disease.

33
00:03:54,310 --> 00:04:03,520
And this is because it's measuring entirely different properties of the brain compared to conventional MRI imaging.

34
00:04:03,790 --> 00:04:09,550
So the image that you get out of the brain is very different. It tells you very different things compared to a normal MRI.

35
00:04:10,000 --> 00:04:15,250
But it is a it is an image rather than a spectrum that you then it remains an image.

36
00:04:15,330 --> 00:04:18,670
Yeah. The challenge is it's very noisy.

37
00:04:19,600 --> 00:04:24,000
I mean, you get very low quality, very blocky, grainy, noisy images.

38
00:04:24,640 --> 00:04:35,530
And therefore to get around that, you need to use sophisticated mathematics, in particular a framework called the variation or Bayesian framework.

39
00:04:36,370 --> 00:04:46,540
And so that was the crux of my Ph.D. is applying variational Bayes to arterials and labelling MRI to improve the image quality that you get.

40
00:04:46,960 --> 00:04:48,860
And here's a challenge for your teaching skills.

41
00:04:48,880 --> 00:04:56,230
Is it possible to explain very simple Bayesian approaches in a way that somebody who doesn't have the maps would understand?

42
00:04:57,910 --> 00:05:01,940
Yeah, it is, actually. It's an. How would you do it?

43
00:05:02,300 --> 00:05:08,700
So. So the crux of Variational Bayes is.

44
00:05:10,650 --> 00:05:25,110
There are good theoretical reasons why a Bayesian approach is the best strategy to use for this problem and for the problems of that statistic.

45
00:05:25,110 --> 00:05:31,229
Piece of statistical terminology. How exactly guilty Bayesian stands in contrast to.

46
00:05:31,230 --> 00:05:34,580
I might be scrambling my terminology here,

47
00:05:34,860 --> 00:05:43,380
but it stands in contrast to likelihood based approaches so that generally people will adopt a maximum likelihood approach.

48
00:05:44,640 --> 00:05:48,780
And a Bayesian approach is is similar.

49
00:05:48,780 --> 00:05:56,730
There's a subtle difference, but in certain circumstances, a Bayesian approach ends up being the same as a as a maximum likelihood approach.

50
00:05:57,090 --> 00:06:02,040
The difference is the Bayesian approach allows you to incorporate prior knowledge.

51
00:06:02,370 --> 00:06:11,880
So before you've even image to the patient, there are certain things that we know to be true about what answer will come back.

52
00:06:12,270 --> 00:06:20,970
And in particular outside of MRI measures, blood flow in the brain, we measure the rates of delivery of blood to different regions in the brain.

53
00:06:21,900 --> 00:06:26,190
And therefore we know ahead of time that these values will never be negative.

54
00:06:27,210 --> 00:06:31,410
And they could be zero. For example, a stroke would be an occlusion.

55
00:06:31,410 --> 00:06:36,210
So therefore some regions of the brain would have zero flow, but they will never be negative.

56
00:06:36,510 --> 00:06:41,850
And so you have this kind of private information of what is plausible from physiology.

57
00:06:43,140 --> 00:06:52,080
And when you incorporate the this prior knowledge into the Bayesian framework, you will end up with answers.

58
00:06:52,080 --> 00:06:56,740
That's a different. I knew you would obtain a few.

59
00:06:57,470 --> 00:07:01,370
And if you didn't adopt the Bayesian approach.

60
00:07:01,390 --> 00:07:09,220
If you use a likelihood based approach, what you're trying to achieve is some level of certainty about what you're looking at on the image.

61
00:07:09,730 --> 00:07:16,150
Hmm. Or in fact, actually, you you qualitatively, you are correct.

62
00:07:16,840 --> 00:07:21,460
But one of the really neat things about a Bayesian approach is not only does it give you

63
00:07:21,610 --> 00:07:25,840
the estimates of the things you're looking for and allows you to quantify uncertainty.

64
00:07:26,860 --> 00:07:34,750
So in the case of a particularly poor image, you might come back with an answer that says, okay, we believe the blood flow to be this value.

65
00:07:35,140 --> 00:07:37,600
But the uncertainty on that is very wide, actually.

66
00:07:38,020 --> 00:07:45,250
And that's the the maths telling you this was a very good acquisition, whereas in other cases you'll get a very tight uncertainty.

67
00:07:45,830 --> 00:07:56,650
So that's also quite a nice property of it. So you were working away on this Dphil very worthwhile, useful thing to be doing.

68
00:07:57,460 --> 00:08:04,660
And then at some point you will have heard that there was something going on in Wukan.

69
00:08:04,870 --> 00:08:13,989
Can you remember I'm asking everybody this. Can you remember when you first heard about the outbreak of of COVID in Wuhan,

70
00:08:13,990 --> 00:08:23,140
and how long was it before it became apparent to your group that actually this was something that you could jump in and and contribute to?

71
00:08:23,980 --> 00:08:30,020
So. Now I can remember very well Christmas 2019, reading about it.

72
00:08:31,370 --> 00:08:36,320
But even before that, actually, there's something interesting happened.

73
00:08:36,500 --> 00:08:43,520
To this day, I do not have an explanation for. I was in Shanghai late November 2019.

74
00:08:43,820 --> 00:08:50,810
Oh, really? Yeah. And I flew back with another Oxford Ph.D., actually.

75
00:08:51,560 --> 00:08:58,610
And we we flew back to the UK, and I've only ever flown to the Far East.

76
00:08:58,610 --> 00:09:01,960
That entire region of the world. I've only ever flown there once, which was this trip.

77
00:09:01,970 --> 00:09:07,460
So I don't know if this is normal, but I'm told it's not normal. On the way back home, it was a British Airways flight.

78
00:09:08,300 --> 00:09:16,910
At one point the cabin was sprayed down with a mist and all of the flight attendants put on respirators and masks,

79
00:09:17,390 --> 00:09:21,740
but they didn't give us any notice of what was about to happen. They didn't give us any masks.

80
00:09:21,740 --> 00:09:25,730
And then for about 20 seconds, the entire cabin was sprayed down with something.

81
00:09:26,480 --> 00:09:33,410
And at the time, I assumed it was some sort of infection control, pest control measure.

82
00:09:34,220 --> 00:09:39,020
Never has it been explained. I'm led to believe it's not exactly common on flights.

83
00:09:39,020 --> 00:09:41,990
I've done a couple of long distance flights from the US, and that hasn't happened.

84
00:09:42,320 --> 00:09:50,960
And we now know in hindsight that the first confirmed case in Shanghai was a few days after we left Shanghai.

85
00:09:51,200 --> 00:09:55,010
Actually, they've managed to date the first COVID case in Shanghai. It's just afterwards.

86
00:09:55,520 --> 00:10:02,600
So for all we know, while we were there, it already started spreading around South Korea spread.

87
00:10:03,470 --> 00:10:08,780
And then when when it started gathering pace in the UK.

88
00:10:11,490 --> 00:10:15,390
We never once did we discuss as a group whether we could contribute.

89
00:10:16,290 --> 00:10:20,969
It's not really that's not really the way my group operates.

90
00:10:20,970 --> 00:10:24,840
We're a computational group. We deal with quite hardcore imaging.

91
00:10:25,710 --> 00:10:29,340
In particular, MRI is not going to be used in a COVID context.

92
00:10:30,300 --> 00:10:38,970
It would never be used in an emergency situation. The doctors would always reach for X-ray C.T., things like that that are much quicker to operate.

93
00:10:39,870 --> 00:10:44,660
And however, as it was becoming increasingly clear there would be a lockdown,

94
00:10:44,670 --> 00:10:50,040
I did email the Jenner Institute, who were working in the floor above me.

95
00:10:50,490 --> 00:10:53,640
I was based at the Old Road campus and the IBM,

96
00:10:54,420 --> 00:10:59,610
so I emailed the Jenner Institute and said I'd be happy to help out for data analysis

97
00:11:00,300 --> 00:11:04,050
because I presumed at some point they would have reams of data to be crunching through.

98
00:11:04,830 --> 00:11:10,530
So they said they would bear the offer in mind and they added it to a list.

99
00:11:10,530 --> 00:11:14,819
I believe they had a list of people who'd offered likewise and said, We'll get in touch if we need you.

100
00:11:14,820 --> 00:11:19,860
But that never did happen. And then very soon after I made that offer.

101
00:11:19,860 --> 00:11:23,790
Anyway, Oxfam's came in to my my path.

102
00:11:24,870 --> 00:11:27,540
Right. So let's tell that story. So what?

103
00:11:30,810 --> 00:11:39,459
Clearly, after looking at what happened in Italy, there was always concern in the UK about hospitals being overwhelmed,

104
00:11:39,460 --> 00:11:45,450
a shortage of beds, shortage of personnel and and also, crucially, crucially, shortage of ventilators.

105
00:11:46,320 --> 00:11:53,760
So how did colleagues of yours come to think that that was an area where they could make a contribution?

106
00:11:55,080 --> 00:11:58,650
Are you on Twitter, by any chance? Yes. Yeah, I'm not.

107
00:12:00,780 --> 00:12:06,749
Medical Twitter is is a very active subset of Twitter and medical.

108
00:12:06,750 --> 00:12:10,290
Twitter, I believe, plays an important role in this question.

109
00:12:10,620 --> 00:12:18,390
So the the person who really kicked off her was Robert Starrett, who is also a PhD.

110
00:12:18,600 --> 00:12:24,750
He's is just about finishing his Ph.D. now, actually, and he's also a military medic.

111
00:12:25,740 --> 00:12:30,000
Do you know any background about Robert or I did wait up a little bit about him as well.

112
00:12:30,000 --> 00:12:38,850
Yes. So. So it was it was him who was looking at the news coming out of Italy.

113
00:12:39,750 --> 00:12:47,130
And in particular, he's very active on Twitter and he's he's plugged into all the communities of health care, people on Twitter and.

114
00:12:48,520 --> 00:12:55,510
It is my understanding he was the first person to join the dots and think this is going to hit the UK like it like a train.

115
00:12:55,780 --> 00:13:02,740
This is going to be really, really bad. And so he started thinking about emergency ventilation.

116
00:13:03,460 --> 00:13:07,690
And then I believe the next person he approached was Andrew Farmer.

117
00:13:09,570 --> 00:13:22,740
Within anaesthetics. Yes. Correct. And Rob's wife and their stroke is a she's an a niece anaesthetist who works in intensive care as well.

118
00:13:23,250 --> 00:13:26,480
So I believe she might have made the introduction.

119
00:13:26,490 --> 00:13:27,550
I'm not sure on that.

120
00:13:27,570 --> 00:13:37,170
But either way, Rob has this background in plastics and surgery and that his wife has the background in intensive care and certainly is more,

121
00:13:37,470 --> 00:13:45,330
more familiar with that part of the medical world. And very soon after Rob started having these ideas, he was put in touch with Andrew Farmer,

122
00:13:46,080 --> 00:13:55,200
I believe Andrew Farmer being the kind of person he is, which is to say is a wonderful free thinker.

123
00:13:55,350 --> 00:13:59,610
Actually, he's he's an excellent person to to work with.

124
00:13:59,610 --> 00:14:08,880
It's really good fun interacting with him. And because he he is just a curious experimental type of person, which is thrilling to be around.

125
00:14:09,720 --> 00:14:13,620
So I suspect he was already having many of the same ideas anyway.

126
00:14:14,460 --> 00:14:19,260
And then between the pair of them and my understanding is that a.

127
00:14:20,410 --> 00:14:26,560
They quickly built a prototype and they looked at a few other people to this effort.

128
00:14:27,250 --> 00:14:33,490
And then the prototype itself was assembled by the engineering workshop at the IBMA.

129
00:14:35,480 --> 00:14:42,110
And particularly the two technicians whose names I'm struggling to remember right now.

130
00:14:42,210 --> 00:14:44,600
I can pull them up later, if that would be important.

131
00:14:46,640 --> 00:14:52,880
And what was the what was the I mean, clearly there are designs of ventilator that are out there already.

132
00:14:53,330 --> 00:14:56,540
What were the advantages of this? Of the oxidant?

133
00:14:56,960 --> 00:15:04,130
What became the oxidant? What was what were they kind of new challenges that it was trying to solve?

134
00:15:05,200 --> 00:15:13,070
That's a good question. So I. Cheap and simple above all else is the is the mantra.

135
00:15:13,490 --> 00:15:16,670
I presume that was the guiding principle from the off.

136
00:15:17,420 --> 00:15:23,960
Cheap and simple. And the initial vision was that we would.

137
00:15:24,680 --> 00:15:32,030
I was involved in the project at this point, but it would be designed and then the design would be open sourced and then it could

138
00:15:32,030 --> 00:15:37,280
be assembled by anyone around the country who had a sufficiently equipped workshop.

139
00:15:37,910 --> 00:15:42,320
And in particular they were thinking university workshops could probably channel these out,

140
00:15:43,340 --> 00:15:54,950
which is why the two technicians at the IBM Me workshop, who are magicians, they have a long history of creating incredible apparatus.

141
00:15:55,360 --> 00:15:59,120
It is why they were such a crucial part of this early effort.

142
00:16:00,110 --> 00:16:02,899
And so the idea was that using their practical knowledge,

143
00:16:02,900 --> 00:16:07,250
they could refine the design and then you chair anyone could start manufacturing from that point.

144
00:16:07,970 --> 00:16:10,820
So it was simplicity all the way.

145
00:16:11,390 --> 00:16:20,690
And around this time, news must obviously have been spreading internally within Oxford that some people were working on this project.

146
00:16:21,790 --> 00:16:30,800
And then the next key persons come on board, in my recollection is Harrison.

147
00:16:30,820 --> 00:16:38,830
Harrison Steele is called and Harrison is now an associate professor at the Department of Engineering.

148
00:16:38,830 --> 00:16:42,380
He was not at the time. I don't believe he is another surgeon.

149
00:16:42,460 --> 00:16:50,350
These people, he's he's a genius. And Harrison is probably the most capable individual on the entire project.

150
00:16:50,890 --> 00:16:54,250
He really is operating at a level above everyone else.

151
00:16:55,420 --> 00:17:00,390
And he. His expertise in controlling electronics.

152
00:17:01,170 --> 00:17:10,860
And he approached Rob and Andrew Farmer and basically said, I will take care of whatever electronics and control is required for this,

153
00:17:11,400 --> 00:17:18,780
and to which I think their response was along the lines of that's actually we hadn't really thought anything on that topic.

154
00:17:19,860 --> 00:17:25,200
So if you're willing to just take care of that, that's great because that is not something they are familiar with.

155
00:17:26,300 --> 00:17:31,190
And then my involvement at this point, I think I've been at home for.

156
00:17:32,260 --> 00:17:43,180
Four or five days. The start of lockdown. And then, yeah, my involvement came in because I had previously imagined that's his name.

157
00:17:44,090 --> 00:17:48,440
I had previously worked with Jim, the technician at the IBM workshop.

158
00:17:48,460 --> 00:17:55,450
I'd done a few projects with him and I was once again, I don't know how Rob came to learn this,

159
00:17:55,450 --> 00:18:05,620
but basically my name was thrown out as someone who would be a safe pair of hands to liaise with Jim because like many workshop technicians,

160
00:18:05,620 --> 00:18:13,550
the world over and. There's a certain way of dealing with the people who own the workshop.

161
00:18:13,970 --> 00:18:20,810
And in particular, at this point, the project was gaining momentum and lots of people were just agglomerated around the nucleus of the project,

162
00:18:21,290 --> 00:18:29,570
but they were all congregating in the workshop. And so there were, you know, a dozen or two dozen people floating around the workshop,

163
00:18:29,690 --> 00:18:38,030
getting in the way of the technicians and just generally asking difficult questions and not not being,

164
00:18:38,150 --> 00:18:41,809
let's say, linearly minded in the way the workshop.

165
00:18:41,810 --> 00:18:49,070
People like to be very linearly minded. And so my role, Rob approached me to basically say, look,

166
00:18:49,070 --> 00:18:56,330
I need to get everyone out of the workshop and I need them to stop asking silly questions to the technicians.

167
00:18:56,690 --> 00:19:04,099
And I need someone to be the liaison between the ideas people and the workshop people,

168
00:19:04,100 --> 00:19:08,270
someone who can speak the right language to them and just insulate them from all the chaos.

169
00:19:09,380 --> 00:19:13,450
And that was my role initially. Mm hmm.

170
00:19:13,880 --> 00:19:21,380
I just wanted to say. Yeah. And you weren't the only graduate student involved in this project.

171
00:19:21,410 --> 00:19:26,840
Oh, no. Very much. No. Yeah, Lots of others accumulated.

172
00:19:27,260 --> 00:19:31,520
I'm happy to say quite a big chunk of them were Trinity.

173
00:19:31,670 --> 00:19:35,390
As it happens at this point, I am now maudlin.

174
00:19:36,200 --> 00:19:39,770
As a PhD student, I went over to modelling college, and that's where Rob was.

175
00:19:40,550 --> 00:19:47,510
But by sheer coincidence, a lot of the key foods happened to be Trinity,

176
00:19:49,010 --> 00:19:57,800
including the the student Andrew, or who was on the news report, which I hope is now in the archives.

177
00:19:59,240 --> 00:20:04,040
When Sky News came in to film the prototype operating the ventilator,

178
00:20:04,640 --> 00:20:14,180
it was Andrew or who was put forward as our representative to demonstrate the prototype working, and he started receiving fan mail.

179
00:20:14,300 --> 00:20:21,710
I believe the various people managed to track him down and he received quite a lot of emails.

180
00:20:22,880 --> 00:20:28,310
So see, I hope that piece of footage can be located and put in the archives is very entertaining.

181
00:20:29,480 --> 00:20:33,110
So what was the how was the prototype being demonstrated?

182
00:20:33,110 --> 00:20:36,680
So at this stage it hadn't actually been used on a on a patient, presumably.

183
00:20:36,920 --> 00:20:43,030
No. And. To my knowledge, I don't think it ever was.

184
00:20:43,700 --> 00:20:54,400
Yeah, at no stage. And the closest they got was either dummy patients or I'm not sure that's the correct term for it.

185
00:20:58,990 --> 00:21:06,030
Now at this stage it was operating on an iron lung or a rubber lung, rather,

186
00:21:07,290 --> 00:21:14,140
which is an ancient piece of apparatus that is covered in rust, but I believe dates from the fifties or sixties.

187
00:21:14,520 --> 00:21:18,669
You'd have to ask an Andrew summary. It's my understanding, Andrew Farmer,

188
00:21:18,670 --> 00:21:25,550
he has a collection of eclectic equipment and apparatus that should probably have been on the scrap heap a long time ago.

189
00:21:25,560 --> 00:21:29,130
But he's here in the museum. Yeah, exactly.

190
00:21:29,580 --> 00:21:32,960
And that he's somehow, you know, saved from the scrap heap.

191
00:21:32,970 --> 00:21:37,590
So we were using this and yeah, it demonstrates the principles.

192
00:21:39,630 --> 00:21:45,930
Which. Which are simply we should you know, I've been kind of assuming making assumptions that people would know what we're talking about.

193
00:21:45,930 --> 00:21:51,330
But it it delivers oxygen to a patient who's not able to breathe for themselves.

194
00:21:51,420 --> 00:22:00,000
Yeah. Yeah, well, no. Good point. So a crash course in ventilation and so.

195
00:22:01,900 --> 00:22:05,440
Yeah, I actually just delivers air. That's the main thing.

196
00:22:06,040 --> 00:22:13,720
It is compensating for the patient's inability to breathe in sufficiently under their own power.

197
00:22:15,610 --> 00:22:18,970
It is common that you will also enrich it with oxygen.

198
00:22:20,430 --> 00:22:27,780
But you probably wouldn't do pure oxygen and that would be a quite severe case if they were on pure oxygen.

199
00:22:28,140 --> 00:22:34,950
I think for COVID, most of the time you could, you know, they just needed help breathing with a slightly enriched mix.

200
00:22:35,920 --> 00:22:47,520
And there are different modes of ventilation. So the way our one operated is the most basic and called a volume controlled ventilator.

201
00:22:47,970 --> 00:22:53,010
So the operator would set a specific volume of air.

202
00:22:54,060 --> 00:23:00,990
Fernando, This will be about 400 millilitres and the ventilator will provide 400 millilitres

203
00:23:00,990 --> 00:23:07,800
in one go at a certain frequency and the patient just has to get with it.

204
00:23:08,070 --> 00:23:15,210
Here the patient has to sync themselves up with the right cause because the ventilator will do it no matter what.

205
00:23:15,240 --> 00:23:20,570
It will deliver that 400 at that frequency. And there are more sophisticated modes.

206
00:23:20,580 --> 00:23:25,440
So there is a mode that's called and pressure controlled,

207
00:23:25,440 --> 00:23:34,709
and that's where the ventilator will provide the access sets and pressure, whereas in volume controlled,

208
00:23:34,710 --> 00:23:43,320
the ventilator will basically provide that's not volume 400 millilitres, whatever pressure is necessary to get it in up to a certain point.

209
00:23:43,710 --> 00:23:52,350
And the reason why pressure becomes so important is because in acute respiratory distress syndrome, AIDS,

210
00:23:52,740 --> 00:24:02,760
which is the the and the state in which COVID patients end up in when they have difficulty breathing.

211
00:24:02,760 --> 00:24:06,060
That was called ARDS respiratory distress syndrome.

212
00:24:06,600 --> 00:24:12,330
And I believe there's there's a stiffening of the the respiratory organs.

213
00:24:13,800 --> 00:24:18,810
And so to to inhale the same volume of air,

214
00:24:18,900 --> 00:24:28,230
a much greater force is required from the patient because that their membranes and the lungs themselves have become stiffer because of this syndrome.

215
00:24:28,590 --> 00:24:33,810
And so that's why the pressure's required to achieve the same volume might change during ARDS.

216
00:24:34,230 --> 00:24:35,520
And then finally, the.

217
00:24:36,440 --> 00:24:45,830
And the most sophisticated forms of ventilation will entirely remember how they work, to be honest, because it's not not at all my background.

218
00:24:46,180 --> 00:24:57,309
And I suppose the key takeaway here is to note that the the ventilator was designed to capture

219
00:24:57,310 --> 00:25:05,390
the the most basic end of the spectrum of patients that might plausibly be puts on it.

220
00:25:05,990 --> 00:25:12,080
It was never designed to be a comprehensive ventilator that could handle every situation that was thrown into it.

221
00:25:12,500 --> 00:25:19,760
It was designed to be a ventilator that could handle 80% of the bottom 80% of cases,

222
00:25:19,760 --> 00:25:26,960
the simple ones, and therefore the top 20%, a complex that would require very advanced ventilation.

223
00:25:27,620 --> 00:25:35,240
You'd be able to put them on the existing high standard ventilators and free up the capacity, basically.

224
00:25:35,750 --> 00:25:41,900
So this was it was a deliberately simple design, and we knew that it would never be able to take care of all situations.

225
00:25:43,200 --> 00:25:47,140
But the idea was that it would be implemented during this pandemic.

226
00:25:47,160 --> 00:25:51,150
I mean, there was there was an urgency about developing it.

227
00:25:51,570 --> 00:25:58,410
Absolutely. So. So this now leads into phase two of the project, the the ventilator challenge.

228
00:25:59,130 --> 00:26:03,830
Does that ring any bells? It's yes, it's beginning to.

229
00:26:03,840 --> 00:26:10,049
But talk me through it. So for posterity, for the future, people listening.

230
00:26:10,050 --> 00:26:18,780
I will also say at this point that there was a lot of media coverage around the ventilator challenge and in particular towards the end of it,

231
00:26:19,290 --> 00:26:23,430
I believe this would have been April, May, June of 2020.

232
00:26:24,090 --> 00:26:36,240
And the Financial Times published quite a thorough investigation of behind the scenes in the Investigates ventilator challenge and.

233
00:26:37,270 --> 00:26:47,830
To my knowledge, that was a very well resourced story because they were reporting things that I we knew to be true but were very much secret.

234
00:26:48,040 --> 00:26:51,790
And B, they then reported a whole lot of things that we have no idea about ourselves.

235
00:26:52,510 --> 00:26:59,409
So so for anyone looking in the future to, to understand the background to that F.T. article,

236
00:26:59,410 --> 00:27:02,620
I don't know the name of the journalist, unfortunately, but it was very well resourced.

237
00:27:03,690 --> 00:27:11,940
So I'm. So this was a this was a challenge that was thrown out by the government, was it to develop the.

238
00:27:11,940 --> 00:27:18,089
Correct. So so coming late, late February, early March,

239
00:27:18,090 --> 00:27:25,620
I'm not entirely sure the state of play was that in Oxford we had put together a benchtop prototype that was operating.

240
00:27:28,610 --> 00:27:35,809
Very basic ventilator and we were thinking we could share it around the country and people could crank out a few hundred at a time.

241
00:27:35,810 --> 00:27:38,480
And it would all be part of this wonderful grassroots effort.

242
00:27:39,290 --> 00:27:50,000
And lots of other people had come up with similar ideas and in particular after after he went on Sky News.

243
00:27:51,690 --> 00:27:56,969
And other people who'd had similar ideas started sending their ideas to us and saying,

244
00:27:56,970 --> 00:28:01,670
Oh, maybe you could incorporate this, this and this, which was very nice.

245
00:28:01,680 --> 00:28:06,570
But we ourselves, we didn't really have the capability to do anything with those ideas.

246
00:28:07,770 --> 00:28:11,190
We weren't able to start manufacturing in large quantities ourselves.

247
00:28:12,450 --> 00:28:14,309
And this is where the ventilator challenge came in.

248
00:28:14,310 --> 00:28:25,170
So the government decided to launch this challenge and they called up all industry in academia to to respond to the ventilator challenge.

249
00:28:25,860 --> 00:28:34,710
And they set out a specification. It was called the AR and the document, the rapidly manufactured ventilator specification.

250
00:28:35,920 --> 00:28:41,890
And that sets out the requirements for a basic ventilator device.

251
00:28:42,490 --> 00:28:50,190
They must provide at least this volume, at least this respiratory rate, and must operate for this length of time, etc., etc.

252
00:28:51,460 --> 00:29:02,970
And then. Teams were called to basically submit their prototype designs to me, that specification, and then that would be a shortlisting stage.

253
00:29:04,130 --> 00:29:13,040
And then that was a preparation for manufacture stage where ideas were assigned to to other.

254
00:29:14,640 --> 00:29:22,709
Other companies who could help accelerate them. And I'm not really sure of all the the logistics anymore in anyway.

255
00:29:22,710 --> 00:29:30,810
It doesn't really matter because the the whole thing constantly kept changing in motion anyway because it was such a stressed, rushed situation.

256
00:29:31,560 --> 00:29:37,350
So I, I believe Oxford wasn't originally on the shortlist to go through,

257
00:29:37,350 --> 00:29:42,630
but then after the Sky News report, I believe we were added to the shortlist at the last minute.

258
00:29:43,410 --> 00:29:48,150
At this point I was still only vaguely involved in the project, so that's why I'm not entirely certain.

259
00:29:50,880 --> 00:30:00,060
And then once we were on the shortlist, what then happened is Smith and Nephew take a medical manufacturing firm in the UK.

260
00:30:00,630 --> 00:30:08,820
They had submitted their own idea to the ventilator challenge that was not shortlisted.

261
00:30:09,570 --> 00:30:11,670
I can't remember the principles.

262
00:30:11,670 --> 00:30:19,860
I think that has worked on negative pressure, which is logical for Smith and Nephew because they're their main product.

263
00:30:19,860 --> 00:30:25,020
The thing that generates most of their revenue is it's called negative pressure therapy for wounds.

264
00:30:25,950 --> 00:30:35,640
And it's a it's a way of keeping serious wounds clear while they heal and it dramatically reduces reinfection, etc.

265
00:30:36,240 --> 00:30:39,299
So they've already mastered that technology for wound healing.

266
00:30:39,300 --> 00:30:42,900
So it's quite logical that they approach ventilation in a similar manner.

267
00:30:43,500 --> 00:30:47,960
And so Smith and Nephew had put that design forward.

268
00:30:47,970 --> 00:30:53,640
It wasn't shortlisted and so they were then match made with us by the government, the powers that be.

269
00:30:55,420 --> 00:31:01,139
So now the situation was we Oxford and King's College, London.

270
00:31:01,140 --> 00:31:04,890
They were at this point a load of anaesthetists and intensive care.

271
00:31:04,890 --> 00:31:08,820
People had come on board from King's College London.

272
00:31:10,150 --> 00:31:15,370
We were matched up with Smith and Nephew to prepare the design for manufacturing.

273
00:31:15,820 --> 00:31:21,990
And so around this time. Eight of us were.

274
00:31:22,990 --> 00:31:32,740
We went drove up from Oxford to Hull, which is where Smith and nephew have their main UK manufacturing facility.

275
00:31:33,550 --> 00:31:38,410
And yeah, we moved up there to prepare for the next phase of the project.

276
00:31:38,410 --> 00:31:42,690
And your role was going to change then?

277
00:31:42,700 --> 00:31:45,970
At this point you were no longer simply the gatekeeper of the workshop. You were.

278
00:31:46,450 --> 00:31:56,589
Yeah. Yeah. So at this point I was Tim Dennison, who was one of the professors at the Oxford end of things,

279
00:31:56,590 --> 00:32:00,100
who has a lot of experience and regulatory side of things.

280
00:32:00,970 --> 00:32:05,740
He used to work in industry at Medtronic, which is a huge American medical devices fund.

281
00:32:06,880 --> 00:32:12,310
He was taking on more of the sort of regulatory strategic direction of the project.

282
00:32:13,900 --> 00:32:19,180
Yeah, I was never officially assigned a role, but Tim Dennison described me as systems integration lead.

283
00:32:19,750 --> 00:32:22,960
Yes, I saw that. I saw that phrase. And I was going to ask you what it meant.

284
00:32:23,420 --> 00:32:33,700
Yeah. So it systems integration is, you know, a little bit about everything and you make sure that it all fuses together.

285
00:32:34,330 --> 00:32:40,120
So, you know a little bit about the control system, you know a little bit about the manufacturing side of things,

286
00:32:40,120 --> 00:32:43,150
you know a little bit about the battery, the power, the sensors.

287
00:32:43,750 --> 00:32:48,820
And your responsibility is to make sure everything matches up nicely.

288
00:32:51,080 --> 00:32:54,620
So yeah, that's a very big responsibility, isn't it?

289
00:32:55,400 --> 00:33:04,250
Yeah. Yeah, it is. And yeah, so I was I mean, I was quite pleased with that because I'm, I am a generalist.

290
00:33:04,250 --> 00:33:08,980
I think that is, that is what I do best. So I thought it seems to me quite well.

291
00:33:08,990 --> 00:33:17,270
But, and though that was the title that someone used to describe what I was doing, it was always very flexible.

292
00:33:17,330 --> 00:33:24,140
Anyway, we didn't really ever have much time to sit down rigidly, define who was doing what with I.

293
00:33:24,470 --> 00:33:27,620
You know, if we were doing things again, that would probably be a good thing to do.

294
00:33:28,280 --> 00:33:37,000
But it was more to do with. Interpersonal relations management, then you weren't actually doing computation.

295
00:33:37,360 --> 00:33:43,190
You weren't doing this. You weren't deploying your skills in maths and computation on this project yourself?

296
00:33:44,390 --> 00:33:52,130
No. A lot of the time there were parts where I did need to do some certain maths, in particular around calibration,

297
00:33:52,730 --> 00:33:56,570
so that that would be the most technical contribution I made was around the calibration.

298
00:33:57,230 --> 00:34:05,810
But no, for for the most part you're entirely right. It was is it's a team management and common sense role.

299
00:34:06,410 --> 00:34:10,580
So in particular, lots of troubleshooting if if a problem arises in a subsystem.

300
00:34:12,890 --> 00:34:18,980
We had lots of issues with lots of issues with the electronics, the batteries, sensors, the things like this.

301
00:34:18,980 --> 00:34:24,860
And so an important part of it is can we redesign around that problem?

302
00:34:24,920 --> 00:34:33,260
Can we tweak the design such that that problem goes away and which more often than not is common sense as opposed to specialist knowledge?

303
00:34:33,860 --> 00:34:45,950
And yeah, so, so off we went to Hull and at this point we were given letters by the engineering department and letters that very importantly stated,

304
00:34:46,730 --> 00:34:52,160
you know, to whom it may concern this person is engaged in a mission of national importance.

305
00:34:53,060 --> 00:35:02,719
You must let them pass, because we we didn't know if we'd meet police roadblocks and all of this we never did once made a police roadblock.

306
00:35:02,720 --> 00:35:07,190
Never once did I have to use my my important letter or the whole of code, which is a shame.

307
00:35:07,970 --> 00:35:14,360
And and the other people who came up that that initial.

308
00:35:15,710 --> 00:35:19,760
Trip to Tahoe were key people and.

309
00:35:21,030 --> 00:35:31,030
So she could. There was a project manager who might have to look her name up in a minute.

310
00:35:31,550 --> 00:35:38,860
Actually, she was very important. She was one of Andrew family's trainee medics.

311
00:35:40,300 --> 00:35:51,820
And I had previously had a career in the RAF working as an engineer and looking after helicopters and had made a change not late in life by any means.

312
00:35:51,820 --> 00:35:59,980
She's only three or four years older than me, I think. But certainly she'd made a change too, to become a graduate doctor.

313
00:36:00,670 --> 00:36:09,520
And so at this stage she was going through the first years of her medical training and Andrew Farmer, who was one of our tutors and.

314
00:36:12,160 --> 00:36:15,560
So Tyler Edwards was the name I was looking for.

315
00:36:17,290 --> 00:36:21,670
And yeah, she came on as an as project manager.

316
00:36:22,180 --> 00:36:28,890
That was one of the few roles that we did properly defined, and that was a role.

317
00:36:28,900 --> 00:36:35,530
She was magnificent and she was an excellent, excellent manager.

318
00:36:36,190 --> 00:36:47,230
And her RAAF experience proved to be very handy because two times did we then draw on her RAAF contacts in the future.

319
00:36:48,220 --> 00:36:52,270
So one of them was to activate this thing called the Civilian Air Corps.

320
00:36:53,470 --> 00:37:03,310
Which is a network of volunteers who I presume live in way of the day that they will be called up,

321
00:37:03,310 --> 00:37:08,320
is the kind of Thunderbirds situation that they've probably been waiting for their whole lives.

322
00:37:09,040 --> 00:37:16,329
And The Cove, it was one of those times when you could activate the civilian echo and they yeah,

323
00:37:16,330 --> 00:37:23,799
they are volunteers who will undertake missions at times of great need in the national interest.

324
00:37:23,800 --> 00:37:33,070
And so one of the times we activated the civilian echo was to fetch Paul Gula, who is a professor in engineering at Oxford.

325
00:37:34,150 --> 00:37:40,720
And we needed him up in Hull very, very short notice to to trouble fix some software issues we were having.

326
00:37:41,680 --> 00:37:54,790
And so and I believe from the time that we decided we needed him in the hull to a taxi being sent for him to his house, only an hour elapsed.

327
00:37:54,790 --> 00:37:57,640
You know, he got a phone call, pack a bag, a taxi. He's coming to get you.

328
00:37:58,090 --> 00:38:02,710
He's driven to a field somewhere outside Oxford where there's a little propeller plane waiting for him.

329
00:38:03,520 --> 00:38:08,499
And then about an hour and a half later, and he's told to pack light as well, because there's not much space in the cabin.

330
00:38:08,500 --> 00:38:10,990
And if it's too heavy, the plane won't take off.

331
00:38:11,830 --> 00:38:19,450
And then an hour and a half later, he's in Hull and a volunteer civilian volunteer who had no idea of their name or where they came from,

332
00:38:19,870 --> 00:38:25,960
had just flown him up on his own initiative to Hull and deposited them for us.

333
00:38:26,950 --> 00:38:30,390
So that was that was quite an interesting little vignette.

334
00:38:30,400 --> 00:38:37,420
And the the other time in which her RAAF experience was very handy was obtaining the Queen's helicopter,

335
00:38:37,570 --> 00:38:43,629
which I'm not sure if you'd heard of that either, and a similar situation.

336
00:38:43,630 --> 00:38:55,600
We needed to move some circuit boards that had just been manufactured in Redding and by a by really fascinating British company.

337
00:38:55,840 --> 00:38:59,260
I forget their name at the moment, but quite an innovative,

338
00:38:59,260 --> 00:39:05,950
established company who manufacture printed circuit boards, custom custom interconnects, that sort of thing.

339
00:39:06,580 --> 00:39:13,870
And, and that yeah, we needed these circuit boards to be with us in Hull as soon as possible so we could put them in

340
00:39:13,870 --> 00:39:21,339
some ventilators that would then be sent out for evaluation testing by the healthcare regulator,

341
00:39:21,340 --> 00:39:25,750
the MHRA. Once, once again, time was of the essence.

342
00:39:26,440 --> 00:39:37,030
So through Chantelle's contacts and she managed to obtain permission for the Queen's helicopter to do the job.

343
00:39:37,690 --> 00:39:41,800
And so it was dispatched off to Redding, picked them up and flew them up to Hull.

344
00:39:41,920 --> 00:39:49,600
And we did all downed tools at that point and we all went to to the airfield by the helicopter was coming in and took some pictures.

345
00:39:49,600 --> 00:39:52,690
I think all of the teams, about 50 of us in total,

346
00:39:53,350 --> 00:40:02,230
went for that because that was partly things were no longer in the supercritical stage at that time.

347
00:40:02,890 --> 00:40:05,890
So we just needed to plug the circuit boards in.

348
00:40:05,890 --> 00:40:13,600
But a lot of the other work had been done and partly I think we all needed the morale boost at that moment in time because.

349
00:40:14,650 --> 00:40:23,980
We'd been in the deep, dark tunnel for quite a while, which is something I will now loop back to, to explain.

350
00:40:24,970 --> 00:40:30,520
So. So we arrived in.

351
00:40:30,520 --> 00:40:39,950
How? Initially the eight of us who'd come up from Oxford and let's call them the the core eight.

352
00:40:39,950 --> 00:40:43,760
We were the boots on the ground and the brief.

353
00:40:45,130 --> 00:40:54,080
Was. Well, I guess the brief was never explained as clearly as it could have been to everyone.

354
00:40:54,740 --> 00:41:02,180
My understanding was that the purpose of going up there was to hand over the project.

355
00:41:03,170 --> 00:41:07,040
It was you guys have you you've done a lovely effort.

356
00:41:07,040 --> 00:41:10,880
You've made your prototype design. Thanks very much.

357
00:41:11,180 --> 00:41:14,360
We've teamed you up with some professionals now who know what they're doing.

358
00:41:15,020 --> 00:41:20,900
So go up there, you know, explain it to them, hand it over, and just let them get on with it.

359
00:41:21,140 --> 00:41:24,600
Now they'll do the rest. And.

360
00:41:26,540 --> 00:41:31,520
I'm not sure what Smith and Nephew were expecting of from their side.

361
00:41:31,760 --> 00:41:35,990
And at any rate. It wasn't quite.

362
00:41:37,170 --> 00:41:44,200
But it wasn't quite that neat. Clean separation as this is now your project.

363
00:41:44,220 --> 00:41:49,140
I hand it over to you. Thanks very much. Good luck. It didn't quite work out like that.

364
00:41:49,740 --> 00:41:50,250
And.

365
00:41:51,420 --> 00:42:02,100
So we and the reason why that's important is because we initially went up thinking we'd be there for two or three days and so had to act accordingly.

366
00:42:03,060 --> 00:42:07,049
I remember I live quite close to the IBM at the time,

367
00:42:07,050 --> 00:42:16,200
so I remember sprinting home and picking up two sets of underwear in a shot and not set thinking we'd only be there for a few days.

368
00:42:17,070 --> 00:42:26,820
Of course, predictably we ended up staying for three and a half weeks, during which time I was still on massive underwear and t shirts and that's it.

369
00:42:29,290 --> 00:42:32,690
And. Yeah, well, what happened?

370
00:42:33,170 --> 00:42:40,940
The way things did play out was it was a handover, but it remained a much more collaborative effort.

371
00:42:41,360 --> 00:42:54,460
And in particular. The the system needed major redesigns in pretty much every respect before it could be manufactured at scale.

372
00:42:55,390 --> 00:43:05,610
However. Smith and nephew on their own would not have been able to achieve those redesigns in a quick enough timeframe.

373
00:43:05,850 --> 00:43:12,270
Let's say so in particular, the thing that Smith and Nephew excelled at was designed for manufacture.

374
00:43:12,840 --> 00:43:23,250
So that is, how can we redesign the device such that it becomes extremely easy to manufacture?

375
00:43:23,610 --> 00:43:30,090
So one specific example of this was the case of the device, the physical enclosure.

376
00:43:30,720 --> 00:43:35,129
And they very quickly through through what we'd done out the window,

377
00:43:35,130 --> 00:43:44,730
and they went for a single sheets of thin steel that is cut with a laser and the template is scored into it using a laser,

378
00:43:44,970 --> 00:43:48,210
and then it's robotically folded into a box.

379
00:43:49,110 --> 00:43:58,350
And and so just like that, that single step, you've dramatically simplified the assembly of the whole thing and it can all be automated.

380
00:43:59,160 --> 00:44:02,280
And that's it's a far superior solution.

381
00:44:02,520 --> 00:44:08,040
And so that kind of knowledge was, you know, was invaluable.

382
00:44:08,040 --> 00:44:15,540
And we would never have made those same decisions because we don't manufacture anything university engineering departments.

383
00:44:15,550 --> 00:44:19,860
We sit around having ideas. We don't actually ever have to build anything ourselves.

384
00:44:20,400 --> 00:44:30,870
With the exception of gym in the workshop. So so a lot of that design effort went into addressing those kinds of questions.

385
00:44:31,260 --> 00:44:34,409
And that's something I was involved in a bit as well.

386
00:44:34,410 --> 00:44:46,430
I remember we had. Quite a fun session in all of that boardrooms and where they had the heads of the manufacturing teams were all there.

387
00:44:47,450 --> 00:44:50,870
Four or five people, very senior in manufacturing were there.

388
00:44:51,590 --> 00:44:55,040
And then it was just me and one other from the Oxfam team.

389
00:44:55,730 --> 00:45:03,260
And we we were basically taking apart the prototype device we had and then reassembling it in different orders.

390
00:45:03,410 --> 00:45:11,690
So we kept taking apart and then trying to assembler in this order, in that order and, and time it how long it took each, each different route.

391
00:45:12,110 --> 00:45:21,380
And that's the kind of basic hands on work that they do to work out what is the optimal route to, to assemble it.

392
00:45:21,680 --> 00:45:26,330
And the reason that they were putting so much effort into this is because they'd already cleared out one of their

393
00:45:26,570 --> 00:45:35,750
warehouses on the site and they'd started putting together a manufacturing line for it with I believe they're about.

394
00:45:37,180 --> 00:45:40,900
That stations 30 or 40 stations on this manufacturing line.

395
00:45:41,290 --> 00:45:46,000
They already had a team of volunteers who were going to work the line.

396
00:45:46,780 --> 00:45:51,040
And so they were really pouring money and resources into this.

397
00:45:51,070 --> 00:45:55,450
In total, they spent in excess of £6 million, I think, on the project.

398
00:45:56,290 --> 00:46:00,849
Yeah. And they were really getting ready to to manufacture at scale because because

399
00:46:00,850 --> 00:46:06,340
that was what we were led to believe was going to happen by the government.

400
00:46:09,550 --> 00:46:21,160
But then for the other parts of the. Other aspects of the project in Oxford, it turns out, needed to remain much more involved,

401
00:46:21,580 --> 00:46:27,190
and in particular, the thing that caused by far and away the most headaches was the control software.

402
00:46:28,300 --> 00:46:32,470
So there's there's two types of software on the device.

403
00:46:33,160 --> 00:46:43,450
There's a basic interface, which is to say there's a small screen that has buttons on it and there are menus and you

404
00:46:43,450 --> 00:46:47,980
can scroll through the menus and you can change the settings and you do all of this.

405
00:46:49,240 --> 00:46:58,060
And it also measures in reports key metrics about the patient so that oxygen saturation and things like this.

406
00:46:59,200 --> 00:47:04,300
So that's one subset of the system of the system's software.

407
00:47:05,430 --> 00:47:10,560
And then the other subset of the system software is the the control software.

408
00:47:10,920 --> 00:47:21,180
So if the user has requested deliver 400 millilitres at 12 breaths per minute to this patient and do not exceed 25 centimetres of water pressure,

409
00:47:22,200 --> 00:47:27,710
what software actually. Takes responsibility for ensuring that happens.

410
00:47:27,730 --> 00:47:34,510
That's the control software. And unfortunately, it turns out the Smith and Nephew,

411
00:47:35,200 --> 00:47:43,089
I think this was a learning experience for them that surprised that in-house software team, unfortunately,

412
00:47:43,090 --> 00:47:53,470
just did not have the capability to address these problems at speed, and that particularly important in the context of a medical device because.

413
00:47:54,750 --> 00:47:58,970
You know, universities these days, a lot of what we produce is we we churn out software.

414
00:47:59,600 --> 00:48:05,060
A lot of my degree was churning out software algorithms that are now used in other universities,

415
00:48:05,750 --> 00:48:09,890
but they are far from perfect and they have bugs in them. They have bugs that I don't know about.

416
00:48:10,640 --> 00:48:20,030
And when a bug arises, someone can email me and hopefully they bring my attention to it and then I can fix it.

417
00:48:20,360 --> 00:48:24,290
But the bugs are never safety critical because the software I have written is never

418
00:48:24,290 --> 00:48:28,790
being used in a life or death situation to make a decision on someone's treatment.

419
00:48:30,570 --> 00:48:38,490
A the software that goes into a ventilator when someone's in acute respiratory distress is 100% safety critical.

420
00:48:39,090 --> 00:48:49,380
And therefore, there are extremely strict rules about how you can write software that goes on to such a device and how you have to test the software.

421
00:48:50,280 --> 00:48:55,860
And they make and they make the process very slow and difficult.

422
00:48:57,180 --> 00:49:02,730
It's a laborious process to write software in the correct manner in that kind of application.

423
00:49:05,040 --> 00:49:10,260
And so that this is where that was a very unfortunate skills gap.

424
00:49:10,380 --> 00:49:19,710
One would have assumed the Smith and Nephew in-house software team had the necessary skills, and they'd be able to do that quickly at speed.

425
00:49:20,580 --> 00:49:26,670
And unfortunately, Smith and nephews in our software capability was just not very strong at all.

426
00:49:28,400 --> 00:49:33,020
Largely, I think, because their existing product lines do not make heavy use of software.

427
00:49:33,030 --> 00:49:35,630
It's just not something they've had to do very much previously.

428
00:49:36,990 --> 00:49:50,930
And we Oxford people, we have a lot of capability in writing software, but we never have to do it in such a constrained safety critical manner.

429
00:49:51,980 --> 00:50:01,670
And so this was by far and away the biggest problem we faced in that second phase of the project when we were preparing to manufacture is with Nephew.

430
00:50:01,680 --> 00:50:07,340
We have continual issues with getting the software to work, verifying it was safe.

431
00:50:08,380 --> 00:50:14,140
And being in regulatory compliance. And that's why we had to to fly up.

432
00:50:14,860 --> 00:50:20,710
Paul Gula from Oxford, a very short notice. He was leading the control side of things.

433
00:50:21,460 --> 00:50:27,640
And also special mention should go at this point to Michael Garcia, who is another Ph.D. student.

434
00:50:29,080 --> 00:50:34,300
He was he's in Paul's group or was in Paul's group at the time, also at Trinity.

435
00:50:35,190 --> 00:50:44,170
And he he was the the magician of get it's called get is a software tool.

436
00:50:44,740 --> 00:50:49,180
And that in this context allows people to collaborate.

437
00:50:49,510 --> 00:50:56,770
So get is a mechanism by which you can have ten people working on different aspects of the same project.

438
00:50:57,420 --> 00:51:00,540
But I think it's GitHub. I've heard of GitHub. Exactly.

439
00:51:00,550 --> 00:51:05,580
Yeah. Yeah. And it allows yeah,

440
00:51:05,590 --> 00:51:10,329
you can have ten people working on ten different sections of the same project and get

441
00:51:10,330 --> 00:51:16,600
is the mechanism that synchronises it all back together creates a centralised copy.

442
00:51:17,020 --> 00:51:26,320
It's a fabulously powerful tool but is quite difficult to use, particularly the advanced features.

443
00:51:27,100 --> 00:51:30,520
And Michael Garcia was a magician at this.

444
00:51:30,550 --> 00:51:35,170
He he was coordinating everything from his kitchen table.

445
00:51:35,200 --> 00:51:38,830
He never did come up to actually, he he was always remote,

446
00:51:39,730 --> 00:51:45,820
but I only learned relatively late in the project he was the guy taking responsibility for that.

447
00:51:47,060 --> 00:51:51,760
And in fact, in so doing, he inspired me after this to go out and don't get properly.

448
00:51:51,790 --> 00:51:56,380
Because at that point, I'd only ever done in a very superficial manner by now.

449
00:51:56,680 --> 00:52:05,170
And a bit better at that. But that's purely through having seen first hand what it's capable of in the hands of someone competent by Michael Gaskin.

450
00:52:06,760 --> 00:52:16,260
So, yeah, I think. That surmises the the further development work we had to do and how.

451
00:52:17,300 --> 00:52:21,080
And then the next the next phase.

452
00:52:21,470 --> 00:52:26,260
And also just to say it was really fun working with the team at Smith and Nephew.

453
00:52:27,950 --> 00:52:31,630
They. They were all volunteers.

454
00:52:31,650 --> 00:52:36,840
They didn't have to do it. And. It's not staff members, but volunteers.

455
00:52:37,180 --> 00:52:40,020
Yeah. Yeah. They were all stuff of Smith nephew, but.

456
00:52:41,990 --> 00:52:49,129
But all of you know, all the Smith nephew stuff had been sent home at this point, all the non-critical stuff.

457
00:52:49,130 --> 00:52:52,220
And so this was a group of people that had self-selected.

458
00:52:52,940 --> 00:52:57,410
We want to keep coming in to work on this.

459
00:52:58,250 --> 00:53:03,500
And there were there were lots of COVID controls on site, understandably so.

460
00:53:03,500 --> 00:53:12,390
In particular, they they split up their team into two halves and kept them completely isolated from each others.

461
00:53:12,420 --> 00:53:22,070
There was a design team and a manufacture team, and we only ever met the design team in person so we could interact physically with the design team,

462
00:53:22,970 --> 00:53:27,650
but we'd never met the manufacture team physically for COVID safety, of course,

463
00:53:27,650 --> 00:53:32,480
because we were worried that if COVID spread into that group, then they'd never be able to manufacture.

464
00:53:32,990 --> 00:53:36,770
So. So even though we're all on the same side, we were in buildings next to each other.

465
00:53:36,770 --> 00:53:45,740
We had to we had to video conference the whole time to the people next door, which which is a real pain when you're trying to.

466
00:53:46,780 --> 00:53:51,960
Yeah. If you're trying to troubleshoot a problem, particularly a problem that has such a physical, tangible form.

467
00:53:52,770 --> 00:53:53,579
For example,

468
00:53:53,580 --> 00:54:01,710
there's not enough space in the box to put the battery here and then the circuit board is going to hit it at this kind of really basic issue.

469
00:54:03,600 --> 00:54:07,830
Yeah, it's such a pain to have to be doing that over online the whole time.

470
00:54:08,940 --> 00:54:17,040
It's just not how people are meant to work. And where were you living during this period while you were in the.

471
00:54:18,500 --> 00:54:22,790
I think as a Holiday Inn Travelodge. Holiday Inn. Okay. So, yeah, a budget hotel?

472
00:54:23,180 --> 00:54:29,130
Yeah, budget hotel. It's right next door. We had a fleet of minis.

473
00:54:29,140 --> 00:54:36,410
I remember Mini Oxford gave us a load of minis too to run around in because I didn't have a car.

474
00:54:36,430 --> 00:54:40,830
Only a handful of people on the project had a car, so that was nice.

475
00:54:40,840 --> 00:54:45,550
Then we brought these minis out from Oxford. We were driving between the hotel and the.

476
00:54:46,780 --> 00:54:50,110
And like the Italian job. Yeah, pretty much.

477
00:54:50,770 --> 00:54:57,240
But I don't know why we would trade between those inside, because they're basically next door to each other as a five minute walk.

478
00:54:57,860 --> 00:55:07,540
And then towards the end, I started. I started walking anyway, because I needed time out, because I've used the phrase long, dark tunnel.

479
00:55:07,540 --> 00:55:16,300
I'm just waiting for it to get back to you. So to go to it. So, I mean, at this point, we're we're fairly in the tunnel because.

480
00:55:18,330 --> 00:55:24,220
It was really punishing at that. The days were really, really long.

481
00:55:25,210 --> 00:55:32,530
And we were regularly starting at seven or eight in the morning and regularly going past midnight.

482
00:55:33,070 --> 00:55:36,970
And then as things reached the crunch time. Yeah.

483
00:55:37,900 --> 00:55:44,330
The thing that I think really ramped up the pressure was when Boris Johnson went into

484
00:55:44,350 --> 00:55:49,240
hospital and he went into Guys and Thomas's and he was in the intensive care ward.

485
00:55:50,050 --> 00:55:58,450
And it so happened that because we had kings, people involved in the project and we,

486
00:55:58,450 --> 00:56:07,210
we had inside knowledge from the, the ward guys and Thomas's of what his status was.

487
00:56:08,080 --> 00:56:12,940
And it is my understanding that the news reports at the time, quite understandably,

488
00:56:13,090 --> 00:56:19,760
were airing on the more catastrophic side of things compared to what his situation actually was.

489
00:56:19,780 --> 00:56:30,280
It is my understanding, based purely on what I heard Second-hand from clinicians in the team, that he would probably be okay.

490
00:56:32,170 --> 00:56:36,790
Nevertheless, it was at that moment that.

491
00:56:38,290 --> 00:56:46,930
The yeah, the magnitude of things really started to sink in because the isolation was really kicking in at this point.

492
00:56:48,070 --> 00:56:53,190
We'd been up there about a week and a half. You're working in a very small team of people.

493
00:56:53,210 --> 00:56:56,830
You're not seeing the outside world. Your horizon has shrunk enormously.

494
00:56:58,720 --> 00:57:05,950
Also, I happened to be reading a book at the time. That is an excellent book, but it was quite bad timing.

495
00:57:06,730 --> 00:57:10,030
It's a sci fi book about the world ending in slow motion.

496
00:57:11,050 --> 00:57:20,230
It's called 7/8 by Neal Stephenson. But I got to the bit where the world starts to end, just as COVID was really kicking off.

497
00:57:21,130 --> 00:57:29,870
So I'm hoping it's out there reading this book before going to sleep, which probably wasn't helping my spirits enormously.

498
00:57:31,330 --> 00:57:36,639
But when Boris Johnson went into hospital, really started to, you know, thinking, wow,

499
00:57:36,640 --> 00:57:44,290
this could be much bigger than than we initially thought, because up to that point.

500
00:57:45,310 --> 00:57:51,010
Probably naively, given that I was in a slightly privileged position with the information, I had access to.

501
00:57:51,190 --> 00:58:02,050
The fundamental message that the government was still putting out was this will pass, will have quite a few difficult weeks.

502
00:58:02,560 --> 00:58:08,260
Lots of people will go to hospital, but it will pass and in a few months time we'll be out of this.

503
00:58:08,260 --> 00:58:12,400
And there's nothing, you know, nothing catastrophic is going to happen.

504
00:58:12,850 --> 00:58:19,210
And it's at that point when the magnitude of things really started landing for me, that it could be quite catastrophic.

505
00:58:19,870 --> 00:58:24,810
And then the thing that. The other important effect it had.

506
00:58:26,000 --> 00:58:34,910
Boris Johnson going into intensive care was at this point the ventilator challenge became complete chaos.

507
00:58:35,660 --> 00:58:39,770
Everything just the wheels came off at a national level.

508
00:58:40,850 --> 00:58:46,370
And I remember one day we we had a call from.

509
00:58:47,800 --> 00:58:50,490
I believe it's Lord Agnew. Phil Agnew. He was cool.

510
00:58:50,980 --> 00:59:00,310
He was the person in the Cabinet Office who, along with Michael Gove, was sort of taking responsibility at Cabinet level for the ventilator challenge.

511
00:59:01,510 --> 00:59:06,160
And in if we were waiting at this point for the go signal,

512
00:59:06,670 --> 00:59:15,610
we'd been told and the government had placed an order for 6000 units minimum, that was an option for 6000 more.

513
00:59:16,060 --> 00:59:21,070
And we'd been told to prepare to manufacture a rate of 5000 a week, I believe.

514
00:59:22,470 --> 00:59:24,900
So. Yeah. We,

515
00:59:25,440 --> 00:59:33,030
we've been told we we're going to need some upfront and we're probably going to need you to maintain a high rates of production for a period of time.

516
00:59:34,290 --> 00:59:40,410
And, and so in response to that, Smith and nephew, you know, started spending a lot of money.

517
00:59:41,710 --> 00:59:47,740
Getting hold of components became a huge challenge, which I'll also come back to in a minute,

518
00:59:47,740 --> 00:59:54,190
because this was one of the most chaotic and poorly thought out aspects of the ventilator challenge.

519
00:59:54,190 --> 00:59:57,940
It was shockingly coordinated.

520
01:00:00,540 --> 01:00:06,570
But when Boris Johnson went into hospital, we we got a phone call and we were candidly told off the record.

521
01:00:07,410 --> 01:00:14,219
While Boris is in intensive care, no one is going to make any decisions whether you should proceed or not.

522
01:00:14,220 --> 01:00:24,080
Proceed. Decision making stopped. Everything stopped and it was Michael Gove or Theo Agnew who told us that.

523
01:00:24,500 --> 01:00:30,650
And. And I think at this point, the politics becomes unavoidable.

524
01:00:31,040 --> 01:00:35,809
Something that was said about Boris Johnson's cabinet from day one way before COVID started,

525
01:00:35,810 --> 01:00:41,060
but something that was said about his cabinet was his cabinet of lightweights who have been put that because,

526
01:00:41,180 --> 01:00:44,390
you know, they're not going to rock the boat. They don't stand up to him. They agree with him.

527
01:00:45,170 --> 01:00:53,390
And yet we very clearly saw as soon as the man at the top was out of action, everything just grinds to a halt.

528
01:00:53,420 --> 01:00:56,630
No one was going to make a decision until he was back in action.

529
01:00:57,560 --> 01:01:06,530
And and that for me, and that's when I came to realise there are no adults in the room.

530
01:01:07,610 --> 01:01:14,409
That was the. The main take home of COVID for me, which hit me really hard, actually.

531
01:01:14,410 --> 01:01:20,550
But. Yeah, it was shocking. There's such chaos and there were no adults in the room.

532
01:01:21,660 --> 01:01:29,790
You expect when things get really bad, you expect that someone somewhere is going to take the lead and make it okay.

533
01:01:30,660 --> 01:01:34,050
This is a very childlike belief. It's very, you know, innocent belief.

534
01:01:34,830 --> 01:01:40,410
And it's these moments like that, you come to realise, no, that is not how the UK operates.

535
01:01:40,740 --> 01:01:45,150
I don't know about other countries. Very few countries came through COVID in a good way.

536
01:01:46,880 --> 01:01:51,800
Certainly in the UK there were no adults in the room at this point. It was complete pandemonium.

537
01:01:52,670 --> 01:01:58,930
And then. Because we had been led to believe we needed to manufacture thousands of units.

538
01:01:59,770 --> 01:02:02,170
We started going through our supply chain.

539
01:02:02,920 --> 01:02:12,430
So this was an exercise in contacting everyone who supplied components for the ventilator and and asking them how many parts have you got available,

540
01:02:12,440 --> 01:02:16,300
how quickly can you provide them? And there were a few bottlenecks.

541
01:02:16,780 --> 01:02:22,750
And were they all UK suppliers or was some of them overseas completely global?

542
01:02:23,170 --> 01:02:28,990
So so in some respects we were quite innovative and resourceful.

543
01:02:29,020 --> 01:02:35,170
So for example, our oxygen sensor that we used in us was a scuba diving oxygen sensor.

544
01:02:36,220 --> 01:02:42,940
And I believe that we were the only people to use a scuba diving oxygen sensor.

545
01:02:42,940 --> 01:02:48,190
But that was quite fortuitous for us because I was the one to actually make the call.

546
01:02:48,730 --> 01:02:57,970
And I called the supplier up one morning and told them that will buy your entire stock place and put it on the Smith and Nephew credit card.

547
01:02:58,540 --> 01:03:05,740
And that was half a million pounds of purchase in one go that I was making on behalf of Smith nephew.

548
01:03:06,610 --> 01:03:13,209
And that was that was one of those moments where you it's a lot of responsibility that I'd never previously had before.

549
01:03:13,210 --> 01:03:15,280
And you come to realise that, you know,

550
01:03:15,280 --> 01:03:24,220
Smith and Nephew really taking a huge amount of trust on this because they're letting 20 year old PhDs call up and just spend money on their behalf.

551
01:03:25,310 --> 01:03:31,150
And so that was something where the supply chain was good because no one else was using those devices.

552
01:03:31,930 --> 01:03:46,840
And so therefore we had a free run at it. Great. One of the really horrendous bottlenecks was nest hoses that caused NIST and in any NHS hospital.

553
01:03:48,530 --> 01:03:54,980
There's a when you need to access high flow oxygen or pressurised air,

554
01:03:55,550 --> 01:04:01,340
that is a plug on the wall and it has a special type of sitting on it like a garden hose almost.

555
01:04:01,970 --> 01:04:09,140
And you can just jam your pipe on the air and it will click and then you do whatever you needed to do.

556
01:04:09,440 --> 01:04:16,160
And that fitting is called a nest fitting and it's standardised across all hospitals in the U.K. I don't know if it's used elsewhere.

557
01:04:17,270 --> 01:04:27,250
I'm. And so therefore, for a ventilator, you need a hose to plug into the wall and then plug into the the device itself.

558
01:04:28,030 --> 01:04:31,140
And. By.

559
01:04:32,560 --> 01:04:37,389
But this is one of the lessons that we've learned globally since COVID is around

560
01:04:37,390 --> 01:04:42,969
supply chain concentration and the fact that it turns out in in a few key areas,

561
01:04:42,970 --> 01:04:47,410
the world is hugely dependent on the tiny number of supplies in the UK.

562
01:04:47,440 --> 01:04:51,970
To my knowledge, there's only one supplier of doses.

563
01:04:53,020 --> 01:04:57,790
And that's probably because day to day there is a tiny demand for nest houses.

564
01:04:58,630 --> 01:05:02,720
There's only so many hospitals in the country. They only need so many houses.

565
01:05:02,740 --> 01:05:06,940
Every now and then they probably break and replace them. But that's it.

566
01:05:08,140 --> 01:05:12,220
And what was not planned for is that.

567
01:05:13,150 --> 01:05:16,540
One day something like the ventilator challenge would come along.

568
01:05:17,660 --> 01:05:23,660
And the government would go round to all of British industry and say.

569
01:05:24,660 --> 01:05:28,860
And, you know, what's the maximum number of ventilators you can prepare?

570
01:05:29,940 --> 01:05:36,990
And start buying the components to do so. And they gave the green light to many teams, and I believe.

571
01:05:38,850 --> 01:05:46,380
I believe that the total demand for ventilators in the UK was projected at 50 or 60,000.

572
01:05:46,890 --> 01:05:52,680
That was the number that was required to get through the first wave, give or take.

573
01:05:53,070 --> 01:05:57,860
And how how many are then in normal times in operation in hospitals?

574
01:05:58,230 --> 01:06:05,220
Probably down to ten or 20,000. I believe I am less certain on that particular figure.

575
01:06:05,480 --> 01:06:10,150
And given it's not my not my day to day area of expertise.

576
01:06:12,360 --> 01:06:18,610
There was this. But this fateful phone call that happened between all the teams,

577
01:06:18,640 --> 01:06:24,360
the ventilator challenge with the Cabinet Office, I wasn't on the call, but then I heard about it.

578
01:06:24,780 --> 01:06:31,980
And one by one, and the person hosting the call for the Cabinet Office went round all the different teams and said,

579
01:06:32,700 --> 01:06:41,339
okay, how many ventilators can you produce? And the numbers that were coming back, I mean, for us, I think we we answered between ten and 15,000.

580
01:06:41,340 --> 01:06:51,060
We were one of the bigger numbers. But all the the big companies that come on board, companies like Ford, I believe Ford were doing them at Dagenham.

581
01:06:51,420 --> 01:06:58,440
Airbus had come on board. Dyson was still claiming that they would be able to manufacture lots.

582
01:06:58,440 --> 01:07:01,830
But I'm firmly of the opinion Dyson was only ever a publicity stunt.

583
01:07:02,670 --> 01:07:06,300
I do not believe that was ever a serious effort on that part.

584
01:07:07,140 --> 01:07:18,330
And the problem is that the total number of ventilators that were claimed could be delivered, I believe came back over 200,000.

585
01:07:20,120 --> 01:07:23,210
And so there was now this issue of we.

586
01:07:24,500 --> 01:07:25,730
We've cried wolf.

587
01:07:25,970 --> 01:07:35,600
We've got all of industry to put down their tools and focus on this problem all at the same time with zero coordination between the different groups.

588
01:07:35,990 --> 01:07:44,360
We've put more headlong onto this, and now actually the number that they've come back with is way over what is required.

589
01:07:45,590 --> 01:07:51,980
And yet they still weren't willing to coordinate the supply of resources.

590
01:07:52,790 --> 01:07:58,220
And this comes into this becomes visible.

591
01:07:59,380 --> 01:08:10,680
The morning I called the man who manufactures nest hoses in the UK, of which every single ventilator manufactured by any team would need a nest hose.

592
01:08:10,690 --> 01:08:15,540
That is a party you absolutely cannot design around. And I asked him, How many hoses have you got?

593
01:08:15,550 --> 01:08:19,030
How many can you supply? And he just broke down in tears.

594
01:08:19,450 --> 01:08:22,460
And he said, you're the fourth call I've had this morning.

595
01:08:23,740 --> 01:08:27,200
I've already been asked to provide tens of thousands.

596
01:08:27,220 --> 01:08:31,360
I have a thousand in stock. There are none in coming any time soon.

597
01:08:32,380 --> 01:08:36,610
And you guys keep calling me up and ordering tens of thousands.

598
01:08:37,550 --> 01:08:42,230
And then I said, okay, this is you know, this is a huge issue.

599
01:08:42,260 --> 01:08:46,790
Thank you for telling me. And I tried to find out more about the problem and what I was planning to do.

600
01:08:47,860 --> 01:08:54,100
Was relayed back to the Cabinet Office because we'd been given instructions to go out and check your supply chain.

601
01:08:54,400 --> 01:09:01,870
And if there are any issues, report them back to the Cabinet Office and we will try and coordinate centrally and try and get involved.

602
01:09:02,470 --> 01:09:11,020
And I said this to the guy and I said, I will I will report this back to the Cabinet Office and hopefully they can tidy up.

603
01:09:11,020 --> 01:09:17,470
And he said, The Cabinet Office have already called me and they just screamed at me and told me to, you know, make it happen, find a way.

604
01:09:18,220 --> 01:09:23,200
So there was no I don't know who placed that call from the Cabinet office.

605
01:09:24,400 --> 01:09:33,160
Maybe it was the Cabinet office, maybe it wasn't the Cabinet Office. Maybe it was someone trying to speak with the the authority of the Cabinet Office

606
01:09:33,160 --> 01:09:37,000
because there were dozens of consultants involved in the project at this point,

607
01:09:37,000 --> 01:09:39,940
unsurprisingly. Who knows what value they were adding?

608
01:09:40,180 --> 01:09:46,630
Certainly the consultant that was added to our project by the Cabinet Office was largely useless.

609
01:09:46,930 --> 01:09:52,330
Every morning they would tell us, Let me know whatever problems you have and I'll help you fix them.

610
01:09:52,900 --> 01:09:57,309
And every time a problem was brought to his attention, he'd say, Coralie, I hope you're not one.

611
01:09:57,310 --> 01:10:04,930
I'm afraid. So I don't know if this person who called up the next hose supplier was one of those consultants or not.

612
01:10:04,930 --> 01:10:10,990
But at any rate, the guy was pretty much having a breakdown on the phone.

613
01:10:11,350 --> 01:10:21,110
And all because. All because basically every team at this point genuinely thought they would be called on to deliver thousands of devices.

614
01:10:22,400 --> 01:10:28,900
And so were double and triple booking the same hoses. There was no rationalisation, no coordination.

615
01:10:29,950 --> 01:10:37,940
And then one of the final. Well, the final resourceful things we did in this phase to get hold of parts was.

616
01:10:39,740 --> 01:10:45,350
The central component of our ventilator was a rubber balloon, effectively a glorified rubber balloon.

617
01:10:45,770 --> 01:10:49,910
And when you squeeze that balloon, that's what pumps are into the patient.

618
01:10:50,840 --> 01:10:53,900
And the balloons are manufactured in China.

619
01:10:55,250 --> 01:11:03,350
And at this point, pretty much every country in the world had put down an export restriction on any medical product of any nature.

620
01:11:04,130 --> 01:11:10,010
And so we had. I think it was 50,000 of these.

621
01:11:11,810 --> 01:11:17,480
We managed to successfully buy 50,000 of them, and they were in Shanghai, in China,

622
01:11:18,020 --> 01:11:22,640
but we couldn't get them out of Shanghai because of the export restrictions.

623
01:11:24,380 --> 01:11:29,360
And so for a while, lots of lots of creative things were being looked at.

624
01:11:30,140 --> 01:11:34,850
Someone I can't remember who, but someone even had a relationship to Richard Branson.

625
01:11:35,300 --> 01:11:41,300
And so at one point we spoke to Richard Branson about could he send one of his jets over that to pick it up for us.

626
01:11:42,140 --> 01:11:46,790
And in the end, that didn't come off the way we got them out.

627
01:11:46,940 --> 01:11:51,050
This was Shontell as well. Johnson Edwards was doing a lot of this work.

628
01:11:51,470 --> 01:11:58,610
I believe somehow we persuaded someone on the ground in China to label them as footballs.

629
01:12:00,030 --> 01:12:03,960
And then Smith and Nephew were able to send one of their planes out to pick them up.

630
01:12:04,800 --> 01:12:11,310
We managed to get them out, but it was purely through creative accounting, let's call it.

631
01:12:12,090 --> 01:12:23,490
And so these are the kinds of issues that that were faced in trying to find enough components to to build tens of thousands of devices.

632
01:12:24,150 --> 01:12:28,020
But it was it was complete chaos all over the world. Complete chaos.

633
01:12:31,250 --> 01:12:38,380
So. Yeah. A very unusual way of spending what should have been the final year of your dphil?

634
01:12:39,520 --> 01:12:47,200
And so, yes, you ended up spending, what was it, three and a half weeks in in Holland.

635
01:12:47,200 --> 01:12:53,040
All that happened during that time? Yeah, it was very well, I mean, ultimately cut to the chase.

636
01:12:53,050 --> 01:12:56,350
Ultimately, the 6000 audit were never made.

637
01:12:56,470 --> 01:13:01,180
Is that is that correct? Yeah. So. So what happened is.

638
01:13:02,800 --> 01:13:13,990
Oh, device was sent off for testing in Birmingham and the first batch of tests came back that the devices and now it's fundamentally operable.

639
01:13:14,530 --> 01:13:22,120
The device broadly meets what it's required to do, but it's one or two areas require major improvement.

640
01:13:22,510 --> 01:13:33,760
And could you also add this thing called respiratory controlled ventilation, which basically is a slightly different mode of operation?

641
01:13:35,230 --> 01:13:40,360
Where the ventilator will wait for the patient to start breathing again.

642
01:13:40,360 --> 01:13:44,440
And once it senses the patient has started breathing again, it will then help them along.

643
01:13:45,790 --> 01:13:50,570
So. And all of this.

644
01:13:51,600 --> 01:13:54,830
Came about what struck us as slightly odd.

645
01:13:55,280 --> 01:13:58,490
We weren't surprised that one or two areas were quite improvement.

646
01:13:58,610 --> 01:14:08,350
That was entirely predictable. But the fact that they the requirements had suddenly increased quite substantially to

647
01:14:08,360 --> 01:14:14,330
we would now like it to be as an assisted ventilation where it assists the patient.

648
01:14:15,200 --> 01:14:24,950
That was quite a big step up and had not previously been a requirement in the specification and it all was a little bit sneaky behind the back.

649
01:14:27,940 --> 01:14:36,580
That's the first red flag in hindsight. We then went ballistic for four or five days.

650
01:14:38,830 --> 01:14:44,590
Trying to trying to implement that change, which happily we could do that entirely in software.

651
01:14:44,590 --> 01:14:48,040
It didn't require a physical redesign of the device.

652
01:14:48,460 --> 01:14:51,850
It just required reworking the software quite substantially.

653
01:14:53,110 --> 01:15:02,830
The problem with that, though, is, as I already mentioned, writing software for a medical device at high speed is extremely complicated,

654
01:15:03,490 --> 01:15:09,280
let alone when you have a very, you know, stretched team in place to do it.

655
01:15:10,030 --> 01:15:16,689
I never wrote any software for the device. It's not something that's in my skill set or already a handful of people that

656
01:15:16,690 --> 01:15:22,360
were capable of doing so and they had already been worked to the to the bone.

657
01:15:24,280 --> 01:15:28,909
You at one point I. I can't remember what Paul Goulart did,

658
01:15:28,910 --> 01:15:38,209
but I effectively instructed Paul to leave his computer and go and sleep in the factory because he'd been at

659
01:15:38,210 --> 01:15:44,390
his computer for 22 hours or something non-stop and was visibly trembling and he couldn't type any more.

660
01:15:45,500 --> 01:15:52,970
So I noticed that on your CV you said that you had the title of Chief Morale Officer.

661
01:15:53,210 --> 01:15:57,080
Yeah. That was that official? Or was that just a role you took?

662
01:15:58,190 --> 01:16:03,300
It was. It was what? Someone described me as rich, which, yeah, I'm very proud of that.

663
01:16:03,370 --> 01:16:06,920
So. Yeah. Yeah.

664
01:16:09,670 --> 01:16:13,330
So anyway, we went to a great speed to do all of that.

665
01:16:14,710 --> 01:16:19,330
And this was the point when. Tim Denison described it as a death march.

666
01:16:20,870 --> 01:16:25,910
It was just a death march when you felt you knew the project was going nowhere?

667
01:16:26,450 --> 01:16:33,020
Yeah, because we we were in limbo. We'd been told, you need to be ready for thousands of devices ASAP.

668
01:16:33,260 --> 01:16:36,800
And the order never came in. We never got a go order.

669
01:16:37,400 --> 01:16:41,060
And for the first part, it was what's happening to Boris Johnson.

670
01:16:41,750 --> 01:16:46,310
No one's going to make a decision. So he comes out and then he came out and still no decision was coming.

671
01:16:47,270 --> 01:16:50,780
And so we didn't really know what was happening and.

672
01:16:51,830 --> 01:16:58,130
At this point, I think one or two people that had had to leave Hull and gone back to Oxford.

673
01:16:59,630 --> 01:17:03,110
I just kept breaking down, crying at random moments.

674
01:17:03,860 --> 01:17:09,920
I started walking in and out of the factory between the hotel just because I needed the time out.

675
01:17:09,930 --> 01:17:16,420
But I remember just crying all the time. It was a very intense period and.

676
01:17:17,680 --> 01:17:24,130
And and then we managed to get the device ready to go back for testing.

677
01:17:25,730 --> 01:17:29,750
Second time in Birmingham, and this time they came back.

678
01:17:30,620 --> 01:17:33,770
The test results came back. It met the requirements.

679
01:17:34,280 --> 01:17:39,670
It was a sufficient device. I forget the exact terminology that was used on the report.

680
01:17:39,680 --> 01:17:43,850
I have the report myself. We we all got a copy of the report, but it meets the requirements.

681
01:17:45,520 --> 01:17:50,839
And then we got a letter signed by Michael Gove shortly afterwards saying, you know,

682
01:17:50,840 --> 01:17:55,940
thanks so much for all the effort you put into it, but unfortunately your device is deficient.

683
01:17:56,300 --> 01:18:00,890
It just doesn't meet the requirements and so you can all pack up and go home.

684
01:18:03,140 --> 01:18:11,630
And that was that. And what effectively happened and this is why about that big article I mentioned earlier in the FT about the ventilator challenge.

685
01:18:13,210 --> 01:18:19,120
What happened is the government got everyone really wound uptight about it.

686
01:18:19,750 --> 01:18:24,700
Industry responded in a more forceful way than than was anticipated.

687
01:18:25,890 --> 01:18:33,060
And to put it bluntly, I think the government didn't want the embarrassment of going around shutting people down.

688
01:18:34,350 --> 01:18:40,410
And so they were doing it by the backdoor. By shifting the goalposts, they kept shifting the goalposts.

689
01:18:41,510 --> 01:18:45,680
Such that they could find an excuse to then say, Oh, your device doesn't matter anymore.

690
01:18:46,520 --> 01:18:50,750
And we were naive and stupid and we didn't read between the lines.

691
01:18:51,260 --> 01:18:56,810
The actually, it didn't matter whether we kept meeting the new requirements,

692
01:18:57,500 --> 01:19:02,540
the government would have just kept doing something to dismiss the whole project.

693
01:19:03,590 --> 01:19:07,430
Then there was a brief moment when we.

694
01:19:08,700 --> 01:19:11,219
I think Smith and Nephew were making the case that, look,

695
01:19:11,220 --> 01:19:17,280
we've we've already sunk huge sums of money into this and we've already acquired thousands of parts to build.

696
01:19:17,280 --> 01:19:20,490
Then let us build them and send them to India.

697
01:19:21,450 --> 01:19:24,479
So then there was this phase of the project of looking for alternative markets.

698
01:19:24,480 --> 01:19:28,470
So India was one of them. We looked into Mexico quite a lot.

699
01:19:29,550 --> 01:19:33,959
Interestingly, A, it could have gone to Mexico,

700
01:19:33,960 --> 01:19:43,680
but it would have struggled a little bit more because the the maximum volume that the ventilator can provide is 600 millilitres.

701
01:19:44,940 --> 01:19:50,610
And by the volume of that, you need to provide scales with the body mass of an individual.

702
01:19:51,210 --> 01:19:54,930
So you normally do a millilitres per kilogram of body mass figure.

703
01:19:56,220 --> 01:20:02,370
Interestingly in Mexico that if needed a ventilator up to about 800 and because average weight

704
01:20:02,460 --> 01:20:08,520
is higher in Mexico basically got heavier people so so it wouldn't have been so well suited,

705
01:20:08,520 --> 01:20:16,229
but it could have been perfectly operable for India. And and that's where the the Oxford Foundation comes in.

706
01:20:16,230 --> 01:20:23,310
It's effectively it's and people trying to find some use.

707
01:20:23,640 --> 01:20:32,350
So it's come out of the device. It could you know, it could be relevant for a future epidemic.

708
01:20:32,890 --> 01:20:37,930
Unfortunately, this is not going to be the last respiratory pandemic the world sees.

709
01:20:38,620 --> 01:20:46,360
So so it could be used for that. And also the the other argument that we various people tried to make to the government is.

710
01:20:47,720 --> 01:20:54,290
You could just spend the money now and buy 10,000 of them and stick them in a warehouse for the next time,

711
01:20:55,100 --> 01:20:59,390
because there will be a next time and top 10,000 emergency devices would be quite handy.

712
01:20:59,990 --> 01:21:07,030
To my knowledge. I believe at the time when the whole ventilator challenge disintegrated,

713
01:21:07,090 --> 01:21:10,390
I'm not going to say it was shut down because it was never shut down properly.

714
01:21:10,400 --> 01:21:17,440
It just disintegrated in a completely chaotic, uncontrolled fashion at that moment in time.

715
01:21:17,980 --> 01:21:25,240
I believe Oxford with Smith, the nephew, was the project that had the capacity to manufacture the most ventilators.

716
01:21:25,930 --> 01:21:34,870
It was the cheapest. It was about £1,000 per unit, whereas the cost of a typical ICU ventilator device is about £20,000.

717
01:21:35,590 --> 01:21:43,030
So it's an incredibly cheap device. Smith and Nephew had managed to get the manufacturing time down to 90 seconds per unit.

718
01:21:44,150 --> 01:21:48,260
And it was the only one that could be manufactured at thousands per week.

719
01:21:48,440 --> 01:21:55,890
None of the others could do thousands per week. But according to Michael Gove doesn't meet requirements.

720
01:21:56,370 --> 01:22:00,150
It was a deficient device according to him,

721
01:22:00,540 --> 01:22:07,200
although he doesn't probably know that we we have the report from the regulator ourselves anyway, so we know that's complete nonsense.

722
01:22:07,720 --> 01:22:12,660
Now you go and I mean, with hindsight, I mean,

723
01:22:12,660 --> 01:22:17,219
I think I've heard it said that it actually turned out we didn't need so many ventilators

724
01:22:17,220 --> 01:22:23,160
after all that that once they were redistributed from one hospital to another,

725
01:22:23,730 --> 01:22:29,160
there were more or less enough to go round. I mean, like the Nightingale, all these Nightingale hospitals that were set up.

726
01:22:29,160 --> 01:22:33,690
And in the end they turned out not to be needed. Yeah, no.

727
01:22:34,560 --> 01:22:42,750
Broadly you are absolutely correct. I think, you know, moments in time it would have been useful.

728
01:22:42,750 --> 01:22:47,280
Thompson Some more ventilators. I'm sure there were moments in time when.

729
01:22:48,820 --> 01:22:52,690
Hospitals were uncomfortably tight on ventilators, capacity.

730
01:22:54,040 --> 01:23:00,670
But broadly speaking, you are correct. We did not require tens of thousands of devices in the UK, thank God.

731
01:23:02,380 --> 01:23:06,610
But that was more than anything. It just turned out that way.

732
01:23:06,790 --> 01:23:14,740
With regards to the Nightingales, an interesting little fact about that is because we we designed our device for nightingales.

733
01:23:15,550 --> 01:23:22,600
So during the design process we because we had loads of clinicians on the team, which is a fantastic attribute to have.

734
01:23:23,080 --> 01:23:31,440
We were talking to the recently appointed heads of the Nightingales people setting them up and they were telling us what they wanted from our device.

735
01:23:31,450 --> 01:23:35,859
So for example, a standard ICU ventilator comes on the stand.

736
01:23:35,860 --> 01:23:41,229
It's got wheels, It's it is its own thing. Ours is a box that you can put on the floor.

737
01:23:41,230 --> 01:23:44,980
You can clip it to a bed, you could clip it to another thing.

738
01:23:46,210 --> 01:23:49,630
It's a big steel box so that if you walk into it, it doesn't matter.

739
01:23:50,290 --> 01:23:56,589
A is designed to be left lying around in a completely chaotic environment

740
01:23:56,590 --> 01:24:01,239
because that's what the heads of Nightingales that were calling us up saying,

741
01:24:01,240 --> 01:24:05,380
I want this on, I want bats on it. And we were saying, Yeah, we can do that.

742
01:24:05,960 --> 01:24:13,840
And interestingly with the Nightingales it is highly debateable whether they would, even if they had been required,

743
01:24:14,320 --> 01:24:18,940
it is highly debateable if they would have ever been useful because they would

744
01:24:18,940 --> 01:24:23,860
have been staffed by nurses who were already allocated to other hospitals.

745
01:24:24,010 --> 01:24:27,940
Actually, they just weren't enough people to fill them. Yeah, yeah.

746
01:24:28,780 --> 01:24:29,800
But in the meantime,

747
01:24:30,310 --> 01:24:40,270
you and your colleagues have designed and prototyped and developed the manufacturing process for a device that sounds like a very good device.

748
01:24:40,780 --> 01:24:45,940
And does it does oxidant still exist as an entity and what are its plans for the future?

749
01:24:46,690 --> 01:24:51,520
So oxygen does still exist. It's a social enterprise.

750
01:24:52,720 --> 01:24:57,640
I myself, I'm not involved in it running Oxfam.

751
01:24:57,820 --> 01:25:06,010
I'm not involved in any regard. Actually in Oxfam, it's mainly it's mainly the professors who are involved in that.

752
01:25:06,940 --> 01:25:12,040
And I believe they they envisage going into.

753
01:25:14,530 --> 01:25:18,549
Other markets where ventilator capacity is more constrained.

754
01:25:18,550 --> 01:25:26,380
So broadly speaking, there are many contacts in the developing world where it'd be really handy to have a ventilator.

755
01:25:26,770 --> 01:25:32,739
So not just for respiratory disease, but, for example, in childbirth.

756
01:25:32,740 --> 01:25:35,680
I believe under some circumstances, if there's a complication,

757
01:25:35,680 --> 01:25:42,430
it can be very advantageous to have ventilator available and there's many aspects of the oxygen really nice.

758
01:25:42,820 --> 01:25:51,850
You can run it on a compressor, an industrial air compressor, whereas a conventional hospital ventilator requires a medical grade.

759
01:25:51,850 --> 01:25:58,870
Air supply is called. To put it simply, if you are in many regions of the world,

760
01:25:58,870 --> 01:26:04,509
you're not going to find a medical grade air supply, but a little industrial compressor, no problem at all.

761
01:26:04,510 --> 01:26:08,860
You'll be able to find that. So so that that's the vision for it.

762
01:26:08,860 --> 01:26:13,239
But I don't know if they've made any great strides in that direction.

763
01:26:13,240 --> 01:26:19,490
I believe not at this moment in time. So for you.

764
01:26:21,260 --> 01:26:27,370
All that drama in Hull was over by, what, the end of May?

765
01:26:28,180 --> 01:26:34,040
Yeah. Yeah. And then we. He came back down to Oxford.

766
01:26:34,460 --> 01:26:42,930
My housemates had been. Having a great time that's watched every day.

767
01:26:43,200 --> 01:26:50,640
Basically, they'd got baked part into the vernacular and been watching Lord of the Rings back to back.

768
01:26:51,540 --> 01:26:56,130
And they they'd been having a lovely time because really Sunday it was a beautiful spring.

769
01:26:57,720 --> 01:27:01,640
And it was a very jarring experience.

770
01:27:01,680 --> 01:27:04,770
I. To this day, I've lived through lockdown.

771
01:27:04,770 --> 01:27:08,430
Subsequently, of course, I have no idea what that first lockdown was like.

772
01:27:10,010 --> 01:27:14,089
I hear so much has been written and said about that lockdown.

773
01:27:14,090 --> 01:27:18,350
And people were wondering, Oh my God, the world could be different. The world could forever be different.

774
01:27:18,590 --> 01:27:20,570
What will it be like when the dust settles?

775
01:27:20,600 --> 01:27:31,730
There was a lot of curiosity and hope around that first lockdown, and I did not experience any of that because we were in the tunnel and.

776
01:27:32,770 --> 01:27:39,100
Yeah. And it took a long time to. I didn't work again until.

777
01:27:40,600 --> 01:27:46,680
So August that year. I mean, I was trying to work, but I wasn't really.

778
01:27:47,310 --> 01:27:53,910
What did you have left to do? Well, you just writing up your thesis or did you still have my my final.

779
01:27:54,390 --> 01:27:58,740
That was the start of my third year. So I was still about a year to go.

780
01:27:58,740 --> 01:28:02,160
I finished in the end, I actually finished entirely on time.

781
01:28:02,400 --> 01:28:11,550
And it just so happened that when COVID hit, I was ahead of schedule and then the effects of COVID wiped out that buffer.

782
01:28:12,360 --> 01:28:17,070
So in the end, I finished June of 21 and.

783
01:28:19,180 --> 01:28:23,170
I was very angry for quite a long time.

784
01:28:24,470 --> 01:28:28,310
I still I'm quite angry at various bowlers. It comes and goes.

785
01:28:29,810 --> 01:28:38,330
Yeah, just really, really angry. And I can't quite put my finger on why, but it is an incredibly demoralising experience.

786
01:28:40,340 --> 01:28:46,280
But you were, you were taken advantage of. I mean, all of you were in a very cynical way, it sounds like to me.

787
01:28:46,910 --> 01:28:49,790
Yeah, we were, you know, we were happy to do so.

788
01:28:51,670 --> 01:28:59,800
And when the rest of you I mean, how did you share your feelings with the other seven people in the group who'd been through all the.

789
01:29:01,340 --> 01:29:07,340
Yeah. I can't speak for everyone, but I know some people incredibly hard.

790
01:29:07,870 --> 01:29:11,930
Mm hmm. Yeah, it hit some people incredibly hard.

791
01:29:12,200 --> 01:29:18,170
Definitely. Yeah. Yeah.

792
01:29:18,410 --> 01:29:21,590
And then. And then there was Kofi, project number two.

793
01:29:21,830 --> 01:29:27,110
Yes. Which was a lot more fun and with a with a better outcome and lovely feedback from.

794
01:29:27,960 --> 01:29:39,030
Yeah. So tell me about that. Probably it's a more it's a less rollercoaster story, but it's still you know,

795
01:29:39,030 --> 01:29:44,270
it still unfortunately contains some pretty damning indictments of of the system.

796
01:29:46,250 --> 01:29:59,510
So that one starts again on Twitter. And it's also Rob Starke because he just cannot get enough of Twitter though it's early 2021.

797
01:30:00,700 --> 01:30:04,120
And the vaccination campaign is about to launch.

798
01:30:05,080 --> 01:30:15,750
And. There were the nine priority groups that had been identified by the Joint Committee on Vaccination and Immunisation.

799
01:30:16,910 --> 01:30:23,330
And I believe priority group number three or four was housebound individuals.

800
01:30:24,620 --> 01:30:27,680
And so they would not be able to come to a vaccination centre.

801
01:30:28,610 --> 01:30:34,010
The GP or nurse would need to go to the person's house and vaccinate them on site.

802
01:30:35,900 --> 01:30:45,530
And. The the original target was that those top four priority groups, the high vulnerability people,

803
01:30:46,910 --> 01:30:52,020
they had to be done within six weeks of the vaccination campaign starting.

804
01:30:52,040 --> 01:30:55,160
That was the first objective. We need all of them done in six weeks.

805
01:30:56,710 --> 01:31:01,690
And in order to incentivise GP's.

806
01:31:03,550 --> 01:31:08,770
But I should say also that unlike the majority of the vaccination campaign where.

807
01:31:10,110 --> 01:31:15,060
Individuals who could do so would go to a central vaccination site.

808
01:31:15,620 --> 01:31:21,240
You know, one of the big massive vaccination centres and they were organised centrally.

809
01:31:22,570 --> 01:31:29,530
And so. The NHS is all about layers of responsibility and who's looking after which patient.

810
01:31:30,100 --> 01:31:37,960
And so for a lot of people they could go to a central vaccination site and so the authority lies with some central body within the NHS.

811
01:31:39,520 --> 01:31:49,720
Housebound patients were only the responsibility of their local GP and there was no other way that they were going to be vaccinated.

812
01:31:50,260 --> 01:31:57,070
It was purely the responsibility of the GP surgery to make sure that housebound patients were taking care of.

813
01:31:59,270 --> 01:32:04,900
And. And that burden falls extremely asymmetrically.

814
01:32:05,020 --> 01:32:11,890
So if you look at the number of patients that are assigned, talking about all patients here,

815
01:32:12,850 --> 01:32:20,980
the number of patients assigned to each GP surgery in the country, I believe the figures range from about 1500.

816
01:32:21,940 --> 01:32:25,330
In some regions to over 3000. In other regions.

817
01:32:25,750 --> 01:32:31,660
So the density of patient to GP can vary by 100%.

818
01:32:33,830 --> 01:32:41,150
And so you can imagine when it comes to that and treating housebound patients, some regions of the country, in particular the South-West,

819
01:32:41,570 --> 01:32:51,140
Devon, Cornwall, the entire South-West region will have a much higher density of older patients who are frail and likely to be housebound.

820
01:32:52,010 --> 01:32:57,480
And so and who probably also live further apart from the very rural area.

821
01:32:57,890 --> 01:33:04,040
So there's this burden of you need to go and look after these patients is incredibly asymmetric.

822
01:33:04,760 --> 01:33:10,000
And also, astonishingly, no one has any idea how many there are.

823
01:33:11,880 --> 01:33:16,560
No one keeps the numbers I've made for Freedom of Information requests at this point.

824
01:33:16,710 --> 01:33:21,960
Two different authorities within the U.K. for the total number of housebound patients.

825
01:33:23,370 --> 01:33:32,280
The only figure I've been given that has any let's say credibility is a lower bound of 250,000.

826
01:33:33,150 --> 01:33:38,150
I'm led to believe the number is probably about 500, but no less than 250.

827
01:33:39,270 --> 01:33:42,380
But shockingly, no one actually has a number for it.

828
01:33:42,390 --> 01:33:43,620
No one's keeping track of that.

829
01:33:45,330 --> 01:33:53,670
And we've been told by someone at the Royal College of GPS that this this illustrates how the problem even exists in the first place,

830
01:33:54,270 --> 01:33:58,040
because no one has responsibility for them in the central NHS.

831
01:33:58,050 --> 01:34:01,320
And therefore, why would you count them? They're not my problem.

832
01:34:02,340 --> 01:34:06,900
These are patients that consistently fall through the cracks of the system

833
01:34:06,960 --> 01:34:13,590
because no one is charged with looking after them for essential points of view.

834
01:34:14,430 --> 01:34:23,700
So. In order to incentivise GP's to get through vaccinating these people as quick as possible.

835
01:34:24,150 --> 01:34:29,100
And they started putting incentives down. So there was a £10 uplift it was called.

836
01:34:29,730 --> 01:34:38,520
So for every housebound patient a GP vaccinated, they would get an extra £10 on top of the existing right as an incentive for doing so.

837
01:34:39,720 --> 01:34:45,480
The missing link was that there was no technology in place to do so.

838
01:34:47,250 --> 01:34:52,650
You can throw money at the problem. You can tell people I'll pay you twice as much money to do it,

839
01:34:53,280 --> 01:34:58,620
but unless you really give them the means to do so, that's not actually going to make the problem go away.

840
01:34:59,830 --> 01:35:05,889
And this is what Rob became aware of on Twitter in particular.

841
01:35:05,890 --> 01:35:12,130
There was one GP in Wales outside Merthyr Tydfil, who was.

842
01:35:13,330 --> 01:35:18,400
He was tweeting about the challenges he faced and he managed to find a plug in

843
01:35:19,090 --> 01:35:24,910
for Google Sheets whereby if he entered a list of postcodes into Google sheets,

844
01:35:25,510 --> 01:35:30,610
it would automatically locate all of them on a map on Google Maps.

845
01:35:31,210 --> 01:35:36,160
And then he could draw a line between the postcodes and he could make a route.

846
01:35:36,550 --> 01:35:40,780
And then he knew, okay, to do these ten patients, this is the route I'll follow.

847
01:35:41,380 --> 01:35:46,840
And he was very pleased with what he'd done and that he was tweeting about it so that he could share it with other GP's.

848
01:35:47,410 --> 01:35:56,860
And it was a very labour intensive solution, but it was a very crude solution to the problem of how do you efficiently do this?

849
01:35:57,580 --> 01:36:02,830
And so Rob brought this to me and said, surely we could automate this, surely we could do better.

850
01:36:03,760 --> 01:36:09,040
It's a travelling salesman problem. Yeah, is exactly that.

851
01:36:09,370 --> 01:36:19,899
Yeah. And, and because it's the travelling salesman problem solutions that exist, this is a trivial problem to solve.

852
01:36:19,900 --> 01:36:24,920
Now there are many solutions out there. And so, so.

853
01:36:25,180 --> 01:36:28,720
So I got on with that and 48 hours later.

854
01:36:30,020 --> 01:36:36,560
I worked two days and two days, and then we put the basic version online.

855
01:36:38,010 --> 01:36:43,710
And yeah, if it went, it gained a lot of use very quickly.

856
01:36:44,610 --> 01:36:52,800
People started sharing it, word of mouth. So on Twitter and there's loads of WhatsApp groups for GPS.

857
01:36:53,130 --> 01:36:58,890
So in particular someone would come across it and then they would put it on a WhatsApp group and they would share it on.

858
01:36:59,820 --> 01:37:05,190
We kept the site free to use. So at no point have we charged for it and.

859
01:37:06,800 --> 01:37:15,410
Partly because we. Partly because we didn't want to and partly cause it to set up a whole payments processing system, which is really complicated.

860
01:37:15,770 --> 01:37:21,860
It's not the kind of thing you want to do at high speed, as much more interested in just solving the problem than getting it out there.

861
01:37:23,170 --> 01:37:28,750
But that was in my rush to do so. That was one thing that would sting in the tail a little bit further on.

862
01:37:29,590 --> 01:37:41,390
And so within. I think within two or three weeks we'd hit 100,000 patients had gone through the site.

863
01:37:41,750 --> 01:37:47,090
You know, at one point about 10,000 patients a day were running through the service.

864
01:37:48,380 --> 01:37:53,600
So how do you focus? Because people would upload their patients to the site so you could go out?

865
01:37:54,650 --> 01:38:02,510
Yes. So the doctors. The patients, Yeah. The doctors would export an Excel spreadsheet and then you upload the Excel spreadsheet.

866
01:38:02,510 --> 01:38:09,080
And so we can we can count. And all it was doing, it was an elaborate frontend to.

867
01:38:10,220 --> 01:38:15,080
To Google Maps. Basically, I would take the patients.

868
01:38:16,300 --> 01:38:20,460
And and the addresses and I would sort them into smaller groups.

869
01:38:20,470 --> 01:38:23,080
So some of ized. Obviously you didn't have the names.

870
01:38:23,800 --> 01:38:30,580
Yeah, no names and also postcodes, because postcodes are anonymized, you can't identify an individual from a postcode.

871
01:38:30,580 --> 01:38:38,650
It's too broad. But these kinds of data questions just capturing that, unfortunately not in a constructive way.

872
01:38:41,800 --> 01:38:47,500
And so say you upload 50 patients, you're not going to go off and vaccinate 50 at once.

873
01:38:47,500 --> 01:38:51,150
You split them into five groups of ten, for example. So that's what the service does.

874
01:38:51,160 --> 01:38:58,180
You can choose the group size, it splits them off into these clusters and then for each cluster, it goes off to Google Maps,

875
01:38:58,180 --> 01:39:04,150
it gets Google Maps to do the travelling salesman for you, and then it sends the result back to the user.

876
01:39:05,220 --> 01:39:08,760
This is incredibly simple and that's how it was able to be built so quickly.

877
01:39:09,750 --> 01:39:12,870
Nothing in there is novel or unique.

878
01:39:13,440 --> 01:39:21,510
It's just gluing together existing pieces of code already out there and it's going to give it its name.

879
01:39:22,080 --> 01:39:27,030
So you've got it's got a name fancy map facts about.

880
01:39:29,050 --> 01:39:32,490
Yeah. My name, actually. I thought I was quite loud and.

881
01:39:34,050 --> 01:39:41,160
And then, yeah, we had 100,000 patients really quickly. The problem with this is it was all running on my credit card.

882
01:39:42,620 --> 01:39:51,980
So the server costs you highly you pay to to have a server, but that's not actually that expensive.

883
01:39:52,340 --> 01:39:59,149
It's about $7 a month. That's fine. The thing that racks up is API costs code and that's that's a fee.

884
01:39:59,150 --> 01:40:02,420
You pay Google for every address that you look up and you route.

885
01:40:03,350 --> 01:40:06,590
So you as an individual, when you go on Google Maps, they don't charge you.

886
01:40:06,980 --> 01:40:13,460
But if you're starting to do large scale stuff programmatically, they start charging you for that.

887
01:40:14,120 --> 01:40:19,400
And I believe I worked out the final figure at some point was about.

888
01:40:20,390 --> 01:40:24,410
Is it? A couple of pennies per patient was the total cost we were paying.

889
01:40:24,740 --> 01:40:29,210
But then when you consider that for some reasons, I'm not entirely sure.

890
01:40:29,780 --> 01:40:31,939
There's really interesting usage patterns around it.

891
01:40:31,940 --> 01:40:38,900
What people tend to do with Vaccine App is they tend to upload their patients and then they'll generate some routes.

892
01:40:39,200 --> 01:40:44,960
They'll look at them, decide whether they like them or not, and then they'll change some settings and then they'll regenerate.

893
01:40:45,170 --> 01:40:48,620
So they'll actually they'll run the same request we call it.

894
01:40:48,920 --> 01:40:55,490
They'll run that four or five times before they're happy, and then they'll they'll save the final one and use that.

895
01:40:56,570 --> 01:41:00,800
But it means that you've paid four or five times to generate the salaries.

896
01:41:01,280 --> 01:41:09,380
And so long story short, very, very quickly, my credit card was £2,000 down in debt and.

897
01:41:10,940 --> 01:41:15,890
And so the logical thing at this point is you think, well, you'd go to the NHS and say,

898
01:41:16,700 --> 01:41:20,390
would you mind awfully just giving us a bit of money to cover this?

899
01:41:21,020 --> 01:41:25,940
And they say, absolutely not. And they don't want to know.

900
01:41:26,540 --> 01:41:33,100
And this. Just became a Kafkaesque mystery.

901
01:41:33,610 --> 01:41:41,019
We were calling around anyone and everyone trying to persuade them to get involved.

902
01:41:41,020 --> 01:41:46,460
So happily Modelling College said they would pay at least £2,000.

903
01:41:46,480 --> 01:41:49,600
Ideally, could we find someone else as well to match them?

904
01:41:51,220 --> 01:41:56,260
We spoke to Oxford University Innovation. The spin out of the uni.

905
01:41:56,260 --> 01:42:05,710
They they loved it and they were happy to put in, I think up to 15, but that the problem with Oxford funding is it's mainly benefits in-kind.

906
01:42:06,400 --> 01:42:14,020
So it's money towards lawyers fees and things like that. They don't really want to just give you a lump sum of cash and also.

907
01:42:15,840 --> 01:42:19,469
The project and you could call it a spin out,

908
01:42:19,470 --> 01:42:26,130
but it wasn't really a spin out in any meaningful sense of the word because we weren't really planning to monetise it in any sense.

909
01:42:26,610 --> 01:42:38,190
It's a very time limited problem that, you know, it might be a useful service for a year or two, but then thereafter it's not.

910
01:42:38,670 --> 01:42:42,049
You know, COVID vaccinations are not going to be going on large scale next year.

911
01:42:42,050 --> 01:42:44,580
And so therefore. Is there a business case for it?

912
01:42:45,380 --> 01:42:51,060
And we tried to get the news involved, you know, as if the news would publish some stories about it.

913
01:42:51,510 --> 01:42:54,629
But, you know, they didn't they didn't want to take it at all. They didn't care.

914
01:42:54,630 --> 01:42:59,220
They didn't even return emails, which I was quite disappointed about.

915
01:42:59,230 --> 01:43:02,370
I thought it would have been a really nice, nice little thing.

916
01:43:05,130 --> 01:43:08,280
And then at this point, we we had another really interesting idea.

917
01:43:08,280 --> 01:43:11,810
This wasn't my idea, actually. Another individual got involved to this point.

918
01:43:11,820 --> 01:43:20,550
A mutual friend of both myself and Robert called Adam Barker, and they had an idea together for.

919
01:43:22,390 --> 01:43:26,440
A a service to reallocate spare vaccine doses at the end of the day.

920
01:43:27,160 --> 01:43:35,290
So I don't know if you remember, people were queuing up outside GP's to try and get a spare shot and this was happening all over the world.

921
01:43:35,530 --> 01:43:39,760
It's this basic principle of at the end of the day, we just need to get the vaccine in someone.

922
01:43:39,770 --> 01:43:45,210
It doesn't matter who. And they quite, quite understandably had the idea.

923
01:43:45,220 --> 01:43:48,280
Many other people around the world had the same idea of.

924
01:43:49,840 --> 01:43:58,510
If you could make a system, an online website where you an individual member of the public at the end of the day, logs onto the website,

925
01:43:58,510 --> 01:44:08,559
puts in their address and gets told there are three spare doses this GP or they are to over that as you could do this really short notice allocation.

926
01:44:08,560 --> 01:44:15,340
If the people who are completely flexible on time, who are happy to just stop work whenever and walk out the door,

927
01:44:15,880 --> 01:44:19,630
it would be an ideal system and it would avoid people having to queue for hours.

928
01:44:20,770 --> 01:44:27,850
And so we we actually at this point started building a prototype of that, even though vaccine itself was still completely unfunded.

929
01:44:28,870 --> 01:44:35,110
We started building a prototype of this other one. And we we spent a few days on that.

930
01:44:35,860 --> 01:44:41,950
And then when we first brought it to the attention of the NHS, they basically said no,

931
01:44:42,610 --> 01:44:53,350
it would require a level of coordination that is not possible in the NHS at this time, which is basically the answer that we got.

932
01:44:53,860 --> 01:44:59,260
In particular, a company called Cure X were handling all the vaccine booking.

933
01:45:00,290 --> 01:45:09,670
They are a private company and they want the contract to manage the vast majority of vaccine appointment booking.

934
01:45:10,150 --> 01:45:16,540
And so I think for our system to have worked, it would have required a deal of integration with secure access.

935
01:45:17,320 --> 01:45:19,920
That was just not going to happen.

936
01:45:19,940 --> 01:45:30,520
Basically, it is a great shame it didn't happen because someone in France had the same idea and he built it and he's called David Meadows,

937
01:45:30,760 --> 01:45:39,400
which means quickly my dose. And he was awarded the Order of Merit for it very soon afterwards.

938
01:45:39,880 --> 01:45:43,930
So when I read that paper, then that was all right.

939
01:45:44,620 --> 01:45:51,470
We wanted to do that anyway and seek some recognition for both your project.

940
01:45:51,500 --> 01:45:55,270
So we should say that because it said a lot about compensation.

941
01:45:55,270 --> 01:46:03,220
So far it's been a little on the downbeat side, but that there have been awards of that not that was an award for the Oxford U.S.

942
01:46:04,120 --> 01:46:09,340
Institute of Engineering and Technology Global Challenge Grand Innovation Award.

943
01:46:09,610 --> 01:46:17,160
I believe it was called which which was a lovely surprise because we, I myself did not know we were in the running for that.

944
01:46:17,170 --> 01:46:21,420
So that was nice. And the vaccine at.

945
01:46:23,300 --> 01:46:27,140
No, not to my knowledge. And I must have been thinking of the Austin fun.

946
01:46:27,380 --> 01:46:35,810
Yeah, not to my knowledge. And so we were having these conversations with with different parts of the NHS.

947
01:46:37,260 --> 01:46:47,010
This is a important aspect of the way the UK health care system is structured is, you know,

948
01:46:47,290 --> 01:46:53,880
we talk about the NHS, but in reality there is no NHS, there is no monolithic NHS.

949
01:46:54,480 --> 01:47:01,200
What there is, is a large number of closely related but slightly independent groupings.

950
01:47:01,920 --> 01:47:05,880
So you have NHS England, for example, you have NHS Digital.

951
01:47:05,910 --> 01:47:11,760
You'd have thought that that was a very obvious place for for this vaccine project to be parked.

952
01:47:11,790 --> 01:47:21,000
NHS Digital, that was NHS X, which is the innovation section of the NHS and introductions were made for us.

953
01:47:21,330 --> 01:47:33,060
Lots and lots of people were. Very generously exerting themselves on our part to try and get the powers that be to take notice of this thing.

954
01:47:33,540 --> 01:47:36,570
And we weren't even trying to sell it to the NHS.

955
01:47:36,600 --> 01:47:42,850
We were not trying to sell the whole system to them. We would have happily gifted it to them.

956
01:47:43,630 --> 01:47:49,840
We were simply looking for a sum of around £10,000 to just cover the costs.

957
01:47:50,110 --> 01:47:57,360
Because I did some calculations and I calculated the £10,000 would more than cover all the costs we could foreseeably hit.

958
01:47:58,510 --> 01:48:03,280
And that's not even including any payment for the time that Rob and I put into it.

959
01:48:03,310 --> 01:48:07,360
It just covers the actual costs of the running it and.

960
01:48:08,360 --> 01:48:13,340
Eventually even. They went all the way to Patrick Vallance.

961
01:48:13,760 --> 01:48:20,030
I have an email from Patrick Vallance on my phone that I got forwarded and forwarded saying,

962
01:48:20,030 --> 01:48:26,270
look, can can vaccination not resource this and he just sent forwarded to vaccination needs.

963
01:48:26,300 --> 01:48:30,920
Thanks for bringing my to my attention and unfortunately nothing ever came of that.

964
01:48:31,460 --> 01:48:38,540
And in essence the way it went is every single conversation we had with people in the

965
01:48:38,540 --> 01:48:43,880
NHS would be along the lines of This is lovely and we really should resource this.

966
01:48:44,360 --> 01:48:57,169
But it's not in my purview to do so. No one was willing to take responsibility and I think a big part of that is because because conventionally

967
01:48:57,170 --> 01:49:03,240
primary care is actually one of the bits of the NHS that is more privatised than the rest.

968
01:49:03,260 --> 01:49:05,780
GP surgeries are actually independent businesses.

969
01:49:06,920 --> 01:49:16,610
So when you have something that is very explicitly geared towards primary care but has this national level strategic value,

970
01:49:18,110 --> 01:49:21,259
no one knows what to make of it and they don't know where to fit it in.

971
01:49:21,260 --> 01:49:30,860
And it obviously it didn't make sense to go around asking individual GP's to pay for it on an individual basis because look, I'm not going to do that.

972
01:49:31,430 --> 01:49:36,169
I'm not going to set up shop as an individual, calling around, making sales and booking all of it.

973
01:49:36,170 --> 01:49:42,229
I've got better things to do. So that's why we just wanted someone in the middle to give us the lump sum of money.

974
01:49:42,230 --> 01:49:54,020
And then then the more we spoke to the NHS, the more they started asking difficult questions around data security and and privacy.

975
01:49:54,470 --> 01:49:59,990
And from the beginning, we've been very I had been very careful around the way I wrote the code.

976
01:50:00,440 --> 01:50:05,620
At this point. I'd already had the experience with Oxfam, so I was much more aware of the,

977
01:50:06,100 --> 01:50:11,750
the implications of writing code that is used in a large scale medical context.

978
01:50:12,770 --> 01:50:22,700
So I believe we'd made many I had made many reasonable attempts to take care of the data privacy and the security aspects of it.

979
01:50:23,770 --> 01:50:33,129
And furthermore, there is guidance online from from the Information Commissioner's Office that quite clearly states that there are no privacy

980
01:50:33,130 --> 01:50:39,670
concerns around what we were doing because we were just using postcodes and postcodes and not regarded as identifying information.

981
01:50:40,900 --> 01:50:45,130
The problem is again dealing with an organisation like the NHS.

982
01:50:46,000 --> 01:50:53,410
You will get asked this question over and over and over again and people don't really they're not willing to take yes for an answer.

983
01:50:54,100 --> 01:51:01,360
They'll take they'll ask the question because, you know, it is their responsibility to do so, which is understandable.

984
01:51:01,660 --> 01:51:05,709
They must always be critical of a new technology that comes in front of them.

985
01:51:05,710 --> 01:51:08,410
They must look at it quite carefully and you can understand that.

986
01:51:09,400 --> 01:51:18,670
But none of them are really willing to take yes for an answer, because if you take yes for an answer, you're that implicitly taking responsibility.

987
01:51:18,910 --> 01:51:27,250
You yourself, you are the person who said, okay, I believe this is acceptable and they don't really want that decision tied to them.

988
01:51:28,330 --> 01:51:33,790
So I then spend hours joining phone calls with various information,

989
01:51:33,790 --> 01:51:39,310
leads across different sections of the NHS, answering the same question again and again and again.

990
01:51:40,690 --> 01:51:45,310
But they would never respond to my answers and say, okay, yes, I'm satisfied.

991
01:51:46,120 --> 01:51:50,760
They would just. They basically just let the conversation die.

992
01:51:51,870 --> 01:51:57,840
And that is why at no point did any section of the NHS ever encourage the uptake of vaccine out.

993
01:51:58,170 --> 01:52:01,590
Never did they condone it, Never did they resource it?

994
01:52:01,590 --> 01:52:05,640
Never. Did they recommend it to be used at primary care level?

995
01:52:06,030 --> 01:52:12,750
Instead, we were adopted entirely by a grassroots effort, word of mouth.

996
01:52:13,770 --> 01:52:17,220
And in the end the people who funded it were the UK Army.

997
01:52:17,490 --> 01:52:22,440
The Ministry of Defence put the money in because they had seconded.

998
01:52:24,030 --> 01:52:29,580
Military medics had been seconded to vaccination teams that were roving around the country.

999
01:52:30,620 --> 01:52:36,680
That was it was called Operation Script was the the Army's contribution to COVID response.

1000
01:52:37,430 --> 01:52:44,720
And because all of the military vaccination teams had adopted vaccine, that they were all using it of their own initiative.

1001
01:52:45,990 --> 01:52:50,900
Eventually, the person who heads up medical innovation within the army,

1002
01:52:52,250 --> 01:53:02,210
it was him who he who took the initiative and he found the money for us because it was so useful to what the army were doing.

1003
01:53:04,460 --> 01:53:08,030
Mm. So if you want a decision taken, go to the army.

1004
01:53:09,850 --> 01:53:13,160
No, They are in the business of making decisions. Yes, indeed. Yeah.

1005
01:53:13,640 --> 01:53:18,120
No, he was. Who knows? Maybe I'll cross paths with him again.

1006
01:53:18,830 --> 01:53:25,190
But it was. He's one of the few people who is willing to stick his neck out and make a decision.

1007
01:53:26,060 --> 01:53:32,150
And it just baffles me. And by the time the money landed, I think I was close to £3,000 in debt.

1008
01:53:32,690 --> 01:53:35,870
Me as an individual, unpaid drawing.

1009
01:53:35,870 --> 01:53:39,770
No salary. Not having a job. Just having a stipend.

1010
01:53:40,010 --> 01:53:43,130
And I'm bankrolling part of the UK's vaccination effort.

1011
01:53:43,580 --> 01:53:48,590
And no one cares. No one wants to listen within the NHS.

1012
01:53:49,160 --> 01:53:52,350
No one wants to do anything. And the sums of money that are involved.

1013
01:53:52,370 --> 01:53:57,049
A trivial. Absolutely. And when you hear these two, it well,

1014
01:53:57,050 --> 01:54:01,640
it must make you absolutely furious when you hear these stories of multi-billion

1015
01:54:01,760 --> 01:54:06,140
pound contracts handed out to private companies of one kind or another.

1016
01:54:06,320 --> 01:54:13,620
So yeah, we spoke to there was an individual at the Royal College of GPS, the Informatics Lead at the Royal College of GPS.

1017
01:54:14,120 --> 01:54:18,049
And he he made some good efforts to help us. And it was very telling.

1018
01:54:18,050 --> 01:54:23,330
At one point he said, you know what, if you were a consultancy, we buy this off for you for 5 million.

1019
01:54:24,560 --> 01:54:29,240
But you're not a consultancy, so we're not going to buy off you and nor are we going to deal with you.

1020
01:54:30,200 --> 01:54:35,810
And the NHS just does not have the the processes are not in place for engaging.

1021
01:54:37,080 --> 01:54:41,280
Like that. They're very familiar with dealing with consultants.

1022
01:54:41,280 --> 01:54:46,889
They know how to do that. The commercial department will draw up a contract and legal will swing in and the

1023
01:54:46,890 --> 01:54:51,390
process is there and they do it all day long and they waste loads of money by doing so.

1024
01:54:52,380 --> 01:54:56,400
But as soon as there's any kind of grassroots innovation, particularly ironically,

1025
01:54:56,400 --> 01:55:05,270
given that Rob is a military medic, so Rob is half in the NHS anyway, that kind of process just does not exist.

1026
01:55:05,280 --> 01:55:08,520
They do not know how to comprehend it and how to engage with it.

1027
01:55:09,550 --> 01:55:16,510
Mm hmm. So now you're you've moved on to the next stage in your in your life?

1028
01:55:17,050 --> 01:55:20,740
Yes. Yeah. What are you working on now?

1029
01:55:21,640 --> 01:55:27,490
So now I am working with Michael Chappell, my supervisor from Oxford,

1030
01:55:28,060 --> 01:55:37,210
and we are in the process of spinning out a company, a medical imaging company, which is related to.

1031
01:55:38,430 --> 01:55:42,569
Well, part of the work I did during my Ph.D. will be in the company,

1032
01:55:42,570 --> 01:55:50,670
but what I did in my Ph.D. is just the latest in a long line of research efforts that have been made in the group.

1033
01:55:51,180 --> 01:55:59,850
And and it's that long line, that body of historical research that we've done that we're we're now looking to commercialised sites.

1034
01:56:00,450 --> 01:56:05,579
It is an imaging company mainly focussed on the brain at this moment in time.

1035
01:56:05,580 --> 01:56:17,010
But maybe we'll be able to spread out in the future and we will be addressing to, to areas of application.

1036
01:56:17,460 --> 01:56:20,650
So. One of them will be clinical trials.

1037
01:56:21,640 --> 01:56:29,080
So hopefully we believe our type of MRI can be very useful in drug discovery.

1038
01:56:30,480 --> 01:56:36,420
And in fact, the day we are recording this, the 6th of December 2022.

1039
01:56:37,710 --> 01:56:42,510
Just recently there's been a lot of discussion around the first Alzheimer's drug.

1040
01:56:42,780 --> 01:56:45,750
Yes. That has significant positive evidence behind it.

1041
01:56:46,380 --> 01:56:56,580
And and this is very interesting because Alzheimer's is one of the holy grails in disease research at the moment.

1042
01:56:57,300 --> 01:57:00,000
And. We hope.

1043
01:57:01,050 --> 01:57:08,850
And when I say we, we as the collective community of people around the world who work on asylum, because there are many, many, many of us.

1044
01:57:09,570 --> 01:57:17,730
And we hope that ESL will be extremely useful in the context of Alzheimer's and.

1045
01:57:18,620 --> 01:57:24,440
We hope that it will be very sensitive to the early signs of Alzheimer's.

1046
01:57:25,730 --> 01:57:31,219
And this this is quite interesting in the context of the drug that's just come

1047
01:57:31,220 --> 01:57:38,780
through because that drug has been entirely developed around pet pets imaging.

1048
01:57:39,980 --> 01:57:44,780
So the drug on emission tomography, I guess we should spell it out.

1049
01:57:45,110 --> 01:57:48,120
Yes. Yes.

1050
01:57:48,120 --> 01:57:53,600
It's a nuclear imaging techniques that involves the use of radioactivity and.

1051
01:57:54,670 --> 01:57:59,170
The the drug that's just been just been reported on.

1052
01:57:59,170 --> 01:58:04,870
I think it's called like minimap lacuna map. I'm terrible with drug names.

1053
01:58:05,080 --> 01:58:08,920
I'm glad I did. Memorable, aren't they?

1054
01:58:09,370 --> 01:58:12,879
Yeah. And there is a reason. And now there are monoclonal antibody.

1055
01:58:12,880 --> 01:58:18,420
That's. That's. Most of the new ones seem to do that right where you have that knowledge.

1056
01:58:18,430 --> 01:58:23,320
I did not notice that you will be able to qualify that in the debate of drug discovery myself.

1057
01:58:23,980 --> 01:58:33,610
And so the one that's just going through, it's it targets the the plaques, the amyloid plaque.

1058
01:58:34,480 --> 01:58:41,290
And and this is important because it fits under the amyloid hypothesis,

1059
01:58:41,290 --> 01:58:46,719
which is remains controversial, although now this is more evidence for I guess.

1060
01:58:46,720 --> 01:58:52,420
But the the amyloid hypothesis is one explanation of how Alzheimer's comes about.

1061
01:58:53,560 --> 01:59:00,790
And therefore, under the amyloid hypothesis, if your drug can remove the amyloid plaques from the brain,

1062
01:59:01,270 --> 01:59:04,480
that will alleviate the symptoms of the disease.

1063
01:59:04,690 --> 01:59:10,059
That's the thinking. And the drug does that and it reduces the amyloid.

1064
01:59:10,060 --> 01:59:17,020
And I think the figure they reported is about 27% improvement in cognitive function.

1065
01:59:18,130 --> 01:59:28,330
And the reason why this is all interesting in the context today is so the MRI we're working with is the missing link in all of this.

1066
01:59:28,330 --> 01:59:34,900
Good news is that for the treatment to be successful, it has to start early.

1067
01:59:35,620 --> 01:59:47,949
Yes, to start really quite early. And at the moment, there remains no good systematic way of screening for the early signs of Alzheimer's.

1068
01:59:47,950 --> 01:59:55,630
The way that they do it at the moment is either doing a CSF sample, which involves taking small amounts of fluid from the body.

1069
01:59:56,110 --> 02:00:02,720
So that's an invasive procedure. Or they do a pet image of positron emission tomography.

1070
02:00:03,350 --> 02:00:07,520
And that also involves injecting the body with a radioactive compound.

1071
02:00:07,610 --> 02:00:15,260
So these techniques are they're well understood that say they're not challenging per say,

1072
02:00:15,860 --> 02:00:21,259
but they're labour intensive and you could never run a screening program at national scale on that.

1073
02:00:21,260 --> 02:00:26,510
And they said it's just too much work. And so this is where the big hope is for MRI.

1074
02:00:26,510 --> 02:00:33,320
That MRI could hopefully be used to run a screening program to pick up the early signs of Alzheimer's,

1075
02:00:33,800 --> 02:00:37,460
and then you could put them on a treatment pathway using this new drug.

1076
02:00:38,480 --> 02:00:45,260
So so the fact that we now have a treatment in place or it looks like we have a fast treatment in place, Alzheimer's,

1077
02:00:45,830 --> 02:00:54,020
it may increase demand for the screening technology to then identify the patients, because at the moment,

1078
02:00:54,350 --> 02:01:01,610
one reservation that you hear in the UK, it's quite a cynical and depressing points of view,

1079
02:01:01,610 --> 02:01:10,220
but is understandable is there is no point screening for Alzheimer's because if we find it, we can't do anything.

1080
02:01:10,580 --> 02:01:17,480
Yep. So why would you bother? But now maybe the calculus has shifted in the other direction.

1081
02:01:17,560 --> 02:01:21,610
So this is an interesting project. Companies very small at the moment.

1082
02:01:22,030 --> 02:01:32,870
It's just me and the supervisor. We're applying for funding early next year, so if we get that, then the company will exist and.

1083
02:01:34,180 --> 02:01:38,340
And I think it will remain. Who knows?

1084
02:01:38,350 --> 02:01:40,989
I mean, maybe it will be very successful and take off very quickly.

1085
02:01:40,990 --> 02:01:47,890
But I think it will always remain quite a tight knit company, actually, because we have a company called Quantified Imaging.

1086
02:01:48,670 --> 02:01:51,670
It's the name of the company, but we haven't even registered yet. All right.

1087
02:01:52,180 --> 02:01:55,270
The company is just a twinkle in our collective eyes at this moment in time.

1088
02:01:57,940 --> 02:02:05,230
Yeah, it will be interesting to work in a team again and I do enjoy working in the same room.

1089
02:02:05,560 --> 02:02:09,070
That was definitely the upside of.

1090
02:02:10,800 --> 02:02:17,060
The upside of oxtail. And the upside of having done vaccine was that.

1091
02:02:18,620 --> 02:02:23,720
I gained ten years worth of career experience and three months.

1092
02:02:24,700 --> 02:02:29,460
The the level of work that I had to just suddenly jump up to.

1093
02:02:30,130 --> 02:02:39,760
I personally authored many of the reports that went to the MHRA regulator in the UK, and I did a lot of the data analysis.

1094
02:02:39,760 --> 02:02:43,150
I believe some of them went to an FDA submission in the US.

1095
02:02:43,900 --> 02:02:50,710
So yeah, I mean, in a very short period of time I gained ten years worth of career experience and.

1096
02:02:52,730 --> 02:02:56,870
I learned a lot about how much I'm worth, actually, as well in the context of work.

1097
02:02:57,320 --> 02:03:00,860
Not that I'm a millionaire at this instance. I'm definitely not that.

1098
02:03:01,640 --> 02:03:08,150
But yes, that. I'm sorry that I shouldn't use the word unique because it's probably not unique,

1099
02:03:08,150 --> 02:03:14,930
but being able to acquire that level of experience is what he had experience working in an academic context.

1100
02:03:14,930 --> 02:03:24,889
And then you had to liaise with both clinical researchers and then clinical practitioners, and then also with a medical device manufacturing company.

1101
02:03:24,890 --> 02:03:28,310
So you see something about the commercial world and then also with government and

1102
02:03:28,310 --> 02:03:37,400
regulatory bodies that that universe of experience must be extraordinarily unusual for,

1103
02:03:37,400 --> 02:03:40,670
as you say, a 20 year old in the second year of that field.

1104
02:03:41,030 --> 02:03:46,030
Yeah, I think I'm. You're right. It's not completely unique, but it's almost unique.

1105
02:03:46,570 --> 02:03:50,110
A few other individuals were in that similar situation.

1106
02:03:51,100 --> 02:03:55,840
There was undoubtedly there was a silver lining to go that for me.

1107
02:03:56,500 --> 02:04:06,280
I also I didn't get COVID and to Christmas of last year, actually, I managed to avoid it for almost two years before I finally got COVID.

1108
02:04:06,820 --> 02:04:14,379
Now, I usually ask everybody how frightened they were of getting it, and the onset tends to scale by age.

1109
02:04:14,380 --> 02:04:20,050
So obviously people who are older or who have co-morbidities are more likely to have been quite frightened about getting it,

1110
02:04:20,050 --> 02:04:24,150
but in younger people, less so. I don't know how you felt initially.

1111
02:04:25,000 --> 02:04:32,320
Yeah, initially I wasn't too frightened. So that really was the very first period.

1112
02:04:32,440 --> 02:04:37,870
And then and then by the time I was doing oxygen and stuff, I was like, Fair enough.

1113
02:04:37,910 --> 02:04:43,140
It doesn't look. It doesn't look as harmless as was initially made out.

1114
02:04:44,160 --> 02:04:48,960
And then by the time I did get it, I was I was vaccinated at that point.

1115
02:04:49,440 --> 02:04:56,310
And I almost certainly had I believe I was the data was the wave going through at the time.

1116
02:04:57,090 --> 02:05:03,180
And I was supposed to go to France for Christmas because I'm French.

1117
02:05:03,180 --> 02:05:12,750
Half of my family is in France and I was supposed to go out there and then just before going, I finally fell ill with it and.

1118
02:05:14,170 --> 02:05:20,020
I'm pleased that I didn't get it earlier because it definitely knocked me out for a few days.

1119
02:05:20,290 --> 02:05:23,759
It's not. Not a pleasant experience. Who knows?

1120
02:05:23,760 --> 02:05:28,650
Maybe on the very first wave. The severity would have been different.

1121
02:05:28,660 --> 02:05:34,240
But but by the time I was vaccinated and I had it, yeah, I really knew I had it.

1122
02:05:34,580 --> 02:05:39,380
Yeah. Did you have experience of friends, colleagues, or family members having it?

1123
02:05:39,500 --> 02:05:43,460
Having severe illness with it? Being badly affected.

1124
02:05:43,850 --> 02:05:48,230
I am. So I think.

1125
02:05:49,850 --> 02:05:53,630
Yeah. I believe my mum lost her sense of smell for a while.

1126
02:05:57,340 --> 02:06:00,620
And. I only know of one.

1127
02:06:01,930 --> 02:06:08,090
Person. Let's say an acquaintance who he was a model in college.

1128
02:06:09,090 --> 02:06:15,360
Who died. And. That was very sad, actually, because a.

1129
02:06:17,720 --> 02:06:25,580
He was a person. For which he reacted very badly to two lockdowns.

1130
02:06:27,170 --> 02:06:30,500
And I became incredibly anxious.

1131
02:06:32,170 --> 02:06:36,760
And didn't handle it well at all. And then he puts on a lot of weight.

1132
02:06:37,990 --> 02:06:43,090
We're talking 30, 40% of his body weight he gained in that period of time.

1133
02:06:43,990 --> 02:06:52,960
And then tragically, it was a week before his first vaccination was due April.

1134
02:06:53,940 --> 02:06:56,100
21 may be May 21.

1135
02:06:56,970 --> 02:07:06,570
He caught COVID, and by this point, I think he was so obese, you know, he'd taken on so much weight that he died very soon afterwards.

1136
02:07:06,990 --> 02:07:12,300
It's a classic case of the co-morbidity COVID and obesity is quite well established, I believe.

1137
02:07:12,990 --> 02:07:18,180
And I didn't know him as an individual in any meaningful sense.

1138
02:07:18,180 --> 02:07:21,720
I just knew him as someone around college. But.

1139
02:07:23,500 --> 02:07:30,120
Yeah I'm very much of the opinion he. He's one of those people that lockdown killed inadvertently.

1140
02:07:31,200 --> 02:07:36,810
Which is not. I'm not going to turn that into a broader argument of lockdown was the wrong course of action to take.

1141
02:07:36,840 --> 02:07:42,110
I genuinely don't know. That question is is too big of a question for me to to pronounce upon.

1142
02:07:42,120 --> 02:07:48,600
I'm not qualified that some he I firmly believe he's one of those inadvertent.

1143
02:07:49,600 --> 02:07:58,450
Costs of lockdown. So I think we really need to just think what in summary,

1144
02:07:58,450 --> 02:08:05,829
what would you say are the key lessons that you're going to take forward in in how you plan

1145
02:08:05,830 --> 02:08:11,200
your life and career in future from the time you spent working on these COVID projects?

1146
02:08:11,770 --> 02:08:16,330
I think the vaccine map taught me that nothing is ever so simple and so obvious.

1147
02:08:17,380 --> 02:08:27,730
That it hasn't already been tried. I unless Rob had been able to point to people on Twitter saying, I need help for this right now,

1148
02:08:27,940 --> 02:08:32,409
I would have never even started working on that because I would have thought, no, it's a solved problem.

1149
02:08:32,410 --> 02:08:35,410
It's done, not needed. But there you go.

1150
02:08:35,410 --> 02:08:38,710
Nothing is ever so simple and so obvious that it shouldn't be tried.

1151
02:08:40,150 --> 02:08:45,570
And from Oxfam's. Yeah.

1152
02:08:45,610 --> 02:08:49,460
From Oxford. There are no adults in the room. I firmly stand by that conclusion.

1153
02:08:49,480 --> 02:08:56,920
And in the future, I. I believe I want to leave the UK for a period of time and go and live and work abroad.

1154
02:08:57,220 --> 02:09:01,760
And. And experience something a bit different.

1155
02:09:01,770 --> 02:09:07,499
I don't think I'm going to find a utopia anywhere. I don't think anywhere is a paragon of that being well-run.

1156
02:09:07,500 --> 02:09:18,240
But I think the UK is afflicted by an inability to to critically look in upon itself and address the problems that are here,

1157
02:09:19,140 --> 02:09:26,850
where we're very good at deflecting and blaming everything but ourselves, I believe, and in COVID that really came home to roost.

1158
02:09:32,100 --> 02:09:32,580
Great.