1
00:00:00,870 --> 00:00:04,410
So could you just give me your name and what your current position is?

2
00:00:05,070 --> 00:00:12,450
So my name is Christine Roman. I'm a postdoctoral research assistant at the Institute of Biomedical Engineering in London through Sancho's Group.

3
00:00:13,020 --> 00:00:16,680
Great. Thanks very much. So first, tell me a little bit about yourself.

4
00:00:16,920 --> 00:00:20,970
Where did you come from originally? How did you first get interested in engineering?

5
00:00:21,510 --> 00:00:28,680
So I'm born in Romania and I moved with my parents when I was 12 in France.

6
00:00:29,580 --> 00:00:37,560
My dad got a position to work as a computer scientist, so we moved when I was 12.

7
00:00:38,190 --> 00:00:46,950
I spent nine years in France doing my middle school, high school and first three years of university and then as part of the engineering degree.

8
00:00:47,850 --> 00:00:50,970
The final did you. Why did you pick engineering? Was that your father's influence?

9
00:00:51,960 --> 00:00:53,780
Well, no, I think I was.

10
00:00:54,420 --> 00:01:05,940
I always enjoyed maths at school is probably the the subject where I found that I didn't have to put as much effort as I did in the others.

11
00:01:07,020 --> 00:01:10,919
And then naturally I did I did follow in that in that path.

12
00:01:10,920 --> 00:01:15,390
I remember growing up with computers at home all the time because they were always there.

13
00:01:15,870 --> 00:01:18,900
Even in the early nineties, we had different computers at home.

14
00:01:20,940 --> 00:01:24,500
But my my parents never influenced the decision of what to do.

15
00:01:24,510 --> 00:01:29,010
I just remember one discussion when I thought about going into finance,

16
00:01:29,010 --> 00:01:34,030
and then my dad had a serious conversation with me saying, maybe I should rethink this.

17
00:01:34,050 --> 00:01:38,220
The next day, I decided to continue with the engineering.

18
00:01:40,950 --> 00:01:44,609
So I, I then went to the US to do a one year exchange.

19
00:01:44,610 --> 00:01:51,870
As part of this five year course, I thought I should take advantage of this mobility that was offered.

20
00:01:52,290 --> 00:01:59,519
And then I went to the US for one year in San Francisco and then for the last year there was another opportunity to do an exchange at Oxford.

21
00:01:59,520 --> 00:02:07,300
Brookes and I decided to go for it, and so that's how I ended up in Oxford.

22
00:02:07,330 --> 00:02:14,580
That was in 2011. So the plan was to come and stay in Oxford for one year, and now 11 years later, I'm still here.

23
00:02:15,660 --> 00:02:23,430
And by that time, had you begun to specialise in a particular branch of engineering at all, or were you still doing a general engineering?

24
00:02:23,910 --> 00:02:30,660
So it was already very specialised in computer science and I, I didn't enjoy programming during university,

25
00:02:30,660 --> 00:02:35,370
so I chose the other option which was available, which was telecommunications.

26
00:02:36,390 --> 00:02:39,760
Rather than doing the programming, I would just do very minimal programming,

27
00:02:39,760 --> 00:02:46,350
but do the application of how do you make things communicate between them, what we currently call Internet of Things.

28
00:02:48,210 --> 00:02:52,890
So I did that, and the course at Brookes was broadband networks for my MSE.

29
00:02:55,590 --> 00:03:03,990
Now, while I was at Brookes, I always had this idea of going back to the US and continuing working in a Start-Up in San Francisco.

30
00:03:05,880 --> 00:03:16,500
But I got offered a scholarship for a Ph.D. and given I didn't know what it was, I decided to take it and it was more difficult than I anticipated.

31
00:03:17,040 --> 00:03:20,399
But it was it was probably worth it.

32
00:03:20,400 --> 00:03:24,420
And I did enjoy the the experience. So what was the subject?

33
00:03:24,630 --> 00:03:31,950
And the subject so is the subject of the PhD was scheduling algorithms in autonomous vehicles.

34
00:03:32,400 --> 00:03:40,980
So what technologies could vehicle use to communicate effectively between them to share information so that they can do less computation

35
00:03:41,490 --> 00:03:48,990
or less prediction on board and therefore reduce cost and have this cooperative system where they just communicate between them.

36
00:03:50,030 --> 00:04:02,639
Um, now, with the expansion of 5G and the advances of technology, it seems something that a level that has been reached.

37
00:04:02,640 --> 00:04:06,220
But ten years ago it was still exploratory research.

38
00:04:06,250 --> 00:04:15,629
You and I and I did my PhD even in, in autonomous vehicles or wireless communications for autonomous vehicles.

39
00:04:15,630 --> 00:04:21,000
But I wanted to once I finished, I felt I wanted to apply this in a different area.

40
00:04:23,190 --> 00:04:33,300
And I also saw that the autonomous vehicles field wasn't probably going to be applied on a larger scale for many more years to come.

41
00:04:33,870 --> 00:04:46,439
So I started looking for alternatives in the digital health space, and there was a lecturer at Brooks that was doing a project on diabetes,

42
00:04:46,440 --> 00:04:49,469
a European project on diabetes, and they were also looking for people.

43
00:04:49,470 --> 00:04:54,390
And that was an excellent opportunity, I think, to move into the health care space.

44
00:04:54,480 --> 00:04:59,820
So what how would you apply digital technologies in in diabetes in.

45
00:04:59,880 --> 00:05:14,700
Diabetes. So the the link between the two is that in that diabetes project, there was a Bluetooth, uh, insulin pump that was connected to a phone.

46
00:05:14,940 --> 00:05:22,290
Now to the same phone there was a insulin, um, blood glucose reader.

47
00:05:23,010 --> 00:05:28,290
And that blue blood glucose reader would send readings automatically every 6 minutes to the phone.

48
00:05:28,650 --> 00:05:38,790
The phone would compute what insulin level was needed, and then it would send the command this pump to to provide the insulin level automatically.

49
00:05:38,850 --> 00:05:44,459
All right. So the patient didn't need to do any decision making. They just left it to the they just left it to the device.

50
00:05:44,460 --> 00:05:52,680
So there was a lot of wireless communication involved, but there was a company doing doing mainly all this integration.

51
00:05:52,680 --> 00:05:58,770
So I joined more on the what they called the work pay, the management work package.

52
00:05:59,340 --> 00:06:04,590
Um, and I oversaw this project, how it, um, how it progressed.

53
00:06:04,590 --> 00:06:11,709
I was an assistant. I wasn't doing, I was assisting the pi in over managing this project.

54
00:06:11,710 --> 00:06:18,470
So, but I did get to learn how machine learning is being integrated in the systems, um,

55
00:06:19,050 --> 00:06:24,780
and how to apply also these wireless technologies into a product that is used by the end users.

56
00:06:25,110 --> 00:06:32,819
What I had done for my PhD was, was theoretical also applied, but no one used it in a way.

57
00:06:32,820 --> 00:06:39,000
I did the simulations, I applied them into hardware. I compare the results between hardware and simulation.

58
00:06:39,360 --> 00:06:44,220
I published the paper and then my algorithm stayed there while in this project.

59
00:06:45,030 --> 00:06:49,200
Um, maybe the complexity of the algorithms, the communication ones,

60
00:06:49,980 --> 00:06:57,690
was one that wasn't as high as what I did, but at least I could see that it works and people can, can use it.

61
00:06:57,690 --> 00:07:00,990
It has some utility in the, um.

62
00:07:02,940 --> 00:07:06,660
In the general well-being of humankind.

63
00:07:06,900 --> 00:07:12,910
Well-Being of humankind. Yes, that's right. And I got tested and I said, this is actually what I want to do.

64
00:07:12,930 --> 00:07:20,550
I want to apply. I want to do applied. I want to apply my skills into something that people use.

65
00:07:21,000 --> 00:07:25,040
And what was digital health looking like at that point globally?

66
00:07:25,050 --> 00:07:28,590
I mean, it was it's a growing field, presumably. It is.

67
00:07:30,330 --> 00:07:40,889
I think at the time it was still very exploratory in the sense that there were just pockets of research being done here and there.

68
00:07:40,890 --> 00:07:47,600
But there wasn't a large companies hadn't got interested in this yet, you know.

69
00:07:48,070 --> 00:07:56,280
So you could I would say maybe we would have that Start-Up feeling that you're building something that at some point might become.

70
00:07:57,060 --> 00:08:01,469
Although, I must admit, I did not have the vision of saying, Oh, this will become big.

71
00:08:01,470 --> 00:08:06,630
And at some point I just enjoyed the tech side and I just enjoyed the project on itself.

72
00:08:07,140 --> 00:08:18,390
I realised the difficulty that it was with diabetes because this insulin pump had to be extremely failsafe.

73
00:08:19,050 --> 00:08:28,680
If you gave too much insulin it would be quite dangerous if you didn't provide enough would have the same level of the interest.

74
00:08:28,800 --> 00:08:38,520
So not only I notice that communication is important, but everything needs to work together.

75
00:08:38,520 --> 00:08:47,190
Otherwise, taking the individual system, the individual parts won't bring any any benefit overall.

76
00:08:48,390 --> 00:09:01,360
And. I think since there has been more interest in in well, all this area, digital health has gained traction, but not a lot of the time.

77
00:09:01,750 --> 00:09:07,500
At the time, I only know about universities or research projects, right?

78
00:09:08,050 --> 00:09:11,170
Yes. Yes. So. So what next?

79
00:09:12,370 --> 00:09:20,720
What next? So after while doing this diabetes project, I,

80
00:09:21,040 --> 00:09:29,710
I wasn't doing very much technical work because I was in this management package stream and I, I wanted to apply a much more of this.

81
00:09:31,450 --> 00:09:37,720
I wanted to apply much more of the skills. And this opportunity came to, to join Lionel's group.

82
00:09:38,260 --> 00:09:49,540
There was an opening, uh, and I knew one of the, one of the post-docs that was working in Linus group at the time before me,

83
00:09:49,540 --> 00:09:53,469
the Jobling, and said, Why don't you have, have a go see what happens.

84
00:09:53,470 --> 00:10:00,040
And I applied and I was very happy to well want to be invited for an interview and

85
00:10:00,040 --> 00:10:05,799
then to after the interview to be told that I'm accepted to join to join the group.

86
00:10:05,800 --> 00:10:09,100
And that was in 2018, right? Yeah, in February.

87
00:10:09,100 --> 00:10:13,690
And I started in the beginning of April. And just tell me a bit more about the group as a whole.

88
00:10:13,690 --> 00:10:18,790
What does Lionel's group do? So Lionel's group is the biomedical signal processing and machine learning group.

89
00:10:19,490 --> 00:10:23,860
They we are experts in signal processing.

90
00:10:24,250 --> 00:10:26,650
How do we take raw signal,

91
00:10:28,150 --> 00:10:40,030
apply different algorithms and filters on it to make it more suitable for machine learning algorithms or for interpretation in clinical settings?

92
00:10:41,200 --> 00:10:47,620
So the reason why we applied such as all these filters before we put data into machine learning algorithms,

93
00:10:47,620 --> 00:10:52,230
is that we want to have clean data in a way.

94
00:10:52,860 --> 00:10:56,169
You might have heard this already. Garbage in, garbage out.

95
00:10:56,170 --> 00:10:57,250
That comes right now.

96
00:10:57,490 --> 00:11:04,389
We want to make sure we put as the garbage in as possible and then we have a good outcome that's in the machine learning context.

97
00:11:04,390 --> 00:11:17,000
But you can do a lot of useful things just by applying noise filtering in signals that you get in in medical devices.

98
00:11:17,020 --> 00:11:23,860
So what sort of signals are you looking for? What's the raw data coming from? For instance, EKGs signals and signals from the heart.

99
00:11:23,860 --> 00:11:27,610
If it's if the patient is moving or the signal is very noisy,

100
00:11:28,540 --> 00:11:35,769
you would either discard those that signal from appearing or you would try to filter it so that it has a

101
00:11:35,770 --> 00:11:42,670
smoother curve so that the clinician can have better interpretation of lots of what's what's going on.

102
00:11:43,090 --> 00:11:50,320
That's probably one example that everyone would be familiar with and which project we use specifically working on when you joined the group?

103
00:11:50,950 --> 00:12:02,890
So when I joined the group, I was working on a remote monitoring project with blood pressure monitoring in patients that I that had had a stroke,

104
00:12:03,460 --> 00:12:07,300
and that is with Professor Richard McManus in Primary Care.

105
00:12:08,400 --> 00:12:23,980
Um, the idea was to have these patients monitor their blood pressure at least three times over the well, three days over the period of one month,

106
00:12:24,550 --> 00:12:36,130
so that they can get better follow up from their clinician and reduce the risk of having a subsequent, um, tia or transmit.

107
00:12:36,220 --> 00:12:41,020
I think that, I think I'm missing a word there. Transient ischaemic attack.

108
00:12:41,020 --> 00:12:47,320
Transient ischaemic attack or stroke? Yes. So that was the that was the idea of the project.

109
00:12:47,830 --> 00:12:54,250
The objectives are also to reduce hopefully their systolic and diastolic blood pressure after the the intervention.

110
00:12:54,520 --> 00:13:02,380
So I mean, patient I mean, I've got one patients you can just buy a cheap blood pressure monitor from boots and do a reading and paste it in.

111
00:13:02,920 --> 00:13:06,430
How did your remote system improve on that?

112
00:13:07,750 --> 00:13:15,790
So that's a very good question. You have self-monitoring and then you have tele monitoring.

113
00:13:16,390 --> 00:13:18,700
Now, I think what self-monitoring is,

114
00:13:18,700 --> 00:13:24,730
is the scenario you just mentioned where someone takes their blood pressure reading right there on a piece of paper,

115
00:13:25,030 --> 00:13:29,469
and then whenever they go to the GP practice, they hand in the piece of paper.

116
00:13:29,470 --> 00:13:38,380
So these are my readings. The difference between self-monitoring and monitoring is that rather than writing these values on a piece of paper,

117
00:13:38,680 --> 00:13:47,850
you input them on a mobile phone to simplify, and then these readings get sent automatically to the GP practice.

118
00:13:47,860 --> 00:13:51,650
There's also some automatic algorithms in the back.

119
00:13:51,750 --> 00:14:00,840
On that, do some safety checks. So if the blood pressure reading is too high, they would have an automated message that you can take another reading.

120
00:14:00,840 --> 00:14:03,300
And then if the second reading is high,

121
00:14:03,750 --> 00:14:12,660
then an alert will be triggered both at the GP and also on the patient's participant side to seek assistance as soon as possible.

122
00:14:14,190 --> 00:14:24,690
So you can shorten the time that you get to see a clinician when the when the.

123
00:14:26,090 --> 00:14:29,209
Health situation deteriorate concern.

124
00:14:29,210 --> 00:14:44,210
Yes. If there's concern and having this integrated loop of of communication between participant and GP's also makes the participant a little bit more,

125
00:14:44,360 --> 00:14:46,159
more self aware about their condition.

126
00:14:46,160 --> 00:14:53,990
And I think it improves their medication adherence and also the way they um they approach their, their condition.

127
00:14:54,440 --> 00:15:05,120
So I think there are multiple benefits on this though the main one is probably that the participant knows that there's

128
00:15:05,120 --> 00:15:12,770
someone looking at their readings and they're under supervision and it probably provides us with some reassurance in a way.

129
00:15:13,370 --> 00:15:21,049
The I mentioned earlier that there was only you would input the readings on a mobile app.

130
00:15:21,050 --> 00:15:31,520
But in this specific scenario with people that had a stroke, they might have some consequences of that stroke, either reduced mobility, aphasia.

131
00:15:31,910 --> 00:15:35,299
So there were different options that had to be developed.

132
00:15:35,300 --> 00:15:42,740
And not just using a mobile phone, which you'd have to install an app, you had to open the app, you have to click here in the readings.

133
00:15:43,280 --> 00:15:52,940
And the other options were either sending an SMS or going onto a computer with on a website where they could also input the readings where everything,

134
00:15:53,150 --> 00:15:56,720
everything would be slightly more aphasia friendly.

135
00:15:57,680 --> 00:16:03,650
So providing these options, I think is key to these digital interventions.

136
00:16:03,830 --> 00:16:13,159
There is that you didn't get as far as the um the blood pressure monitor itself automatically sending the readings to the GP not,

137
00:16:13,160 --> 00:16:16,850
not in that project that we've had it done in the project further down the line.

138
00:16:16,850 --> 00:16:19,740
Like we'll talk about that when we get. Yes. Okay. Yeah.

139
00:16:19,800 --> 00:16:27,050
So for that project it was the main objective was to prove that this digital intervention can help

140
00:16:27,430 --> 00:16:33,710
these people lower their blood pressure and also lower the risk of having another another stroke.

141
00:16:33,740 --> 00:16:39,020
And did you has it been going on long enough for you to have collected that data or is that still in progress?

142
00:16:40,010 --> 00:16:52,700
Unfortunately, because of the pandemic funding was, um, funding was withdrawn by the British Heart Foundation, um, I think in May 2020.

143
00:16:52,820 --> 00:16:55,850
Right. Okay. So I was arrested halfway through.

144
00:16:55,850 --> 00:16:58,850
It was okay. So we've got, we've arrived at the pandemic.

145
00:16:58,850 --> 00:17:09,620
So, so can, can you remember when you first heard that there was something serious happening in Lucan and,

146
00:17:09,620 --> 00:17:15,320
and, and how long was it before the group realised that there was work they could do that would be relevant.

147
00:17:16,550 --> 00:17:22,220
So the first time I heard about this was when my father called me one evening,

148
00:17:23,300 --> 00:17:28,100
I think it was in February, and he said, You heard about the virus in China.

149
00:17:28,100 --> 00:17:39,379
This is quite serious. Now my dad follows the news very much and when I heard first thought about it, I said, Oh,

150
00:17:39,380 --> 00:17:47,090
it's probably going to be something that a week as the bird flu in India, I think 28.

151
00:17:47,510 --> 00:17:51,140
29,000. Yes, the bird flu. And there was silence before that.

152
00:17:51,440 --> 00:17:56,209
Exactly. So I thought, oh, well, it's probably going to stay in that part of the world.

153
00:17:56,210 --> 00:18:00,560
We will have similar to what happens with the bird flu and then that's that.

154
00:18:01,760 --> 00:18:11,180
But I think ten days or two weeks later I was in the office and the phone rang at around 4 p.m.

155
00:18:11,420 --> 00:18:20,750
The phone usually never rings and I picked up and it was Leila's assistant and she said,

156
00:18:21,470 --> 00:18:24,740
Christian, there's Lionel for you on the phone if she wants to speak with you.

157
00:18:25,640 --> 00:18:26,930
And I said, Yes, of course.

158
00:18:28,400 --> 00:18:34,970
No, I didn't know what to expect, as I think I've never spoken with Lionel on the phone, although I hadn't been in the group for two years,

159
00:18:35,600 --> 00:18:44,030
and then Lionel was put through the line and said, Hi, Christian, I don't know if you've heard about the virus.

160
00:18:44,030 --> 00:18:51,439
You need to drop everything you're doing absolutely everything and focus on the voodoo you, meaning the virtual high dependency unit.

161
00:18:51,440 --> 00:18:58,790
And this is this project about monitoring vital signs with, um, with wearable devices.

162
00:18:59,300 --> 00:19:03,440
And I remember saying, oh, I know, but I have this deadline for this project.

163
00:19:03,620 --> 00:19:13,040
Drop everything, he said. Also speak with Mauro and Marco if they're around, which are my colleagues and you need to work on this.

164
00:19:13,040 --> 00:19:17,540
I think we'll be meeting at the hospital tomorrow at 5 p.m. to discuss about this.

165
00:19:19,310 --> 00:19:25,370
And that was almost as if.

166
00:19:25,440 --> 00:19:27,659
In the general sense.

167
00:19:27,660 --> 00:19:36,870
And I close the phone and then I just remember having my colleague Mark on the side of and I told him, we have to drop everything.

168
00:19:36,870 --> 00:19:40,470
We're working on this. We do now full time.

169
00:19:42,540 --> 00:19:46,019
And what what once that was what it was in February.

170
00:19:46,020 --> 00:19:53,120
And I think it was a Wednesday because we met at the hospital on Thursday at 5 p.m. and I don't remember quite a lot.

171
00:19:53,130 --> 00:19:58,800
I also had to contact some of the former colleagues of of the group because I needed the Bluetooth

172
00:19:58,800 --> 00:20:05,730
protocols for some of the wearable devices which were integrated into the into the view at the time,

173
00:20:05,730 --> 00:20:14,160
but only as prototypes. So we were also looking at integrating, for example, Bluetooth temperature problem, which wasn't part of the project.

174
00:20:14,370 --> 00:20:18,089
So the virtual high dependency unit was a project that was already up and running.

175
00:20:18,090 --> 00:20:25,620
Yes, thought started but but presumably just as something that was one of many projects in the in the group and without a particular sense

176
00:20:25,620 --> 00:20:34,769
of urgency until one not so there wasn't any particular sense of urgency at the time the project was already going since 2017.

177
00:20:34,770 --> 00:20:40,049
It's a project which took a lot of took a lot of time to build up.

178
00:20:40,050 --> 00:20:49,500
I think I know and research colleagues from the clinical research group led by Professor Peter Watkinson,

179
00:20:49,830 --> 00:20:56,639
were working on this I think for the past ten years on trying to build such as a virtual high dependency unit.

180
00:20:56,640 --> 00:21:06,600
And in 2017 I think with the new BRC funding they they managed to do a um to set up a

181
00:21:06,600 --> 00:21:10,830
project that will end in a randomised controlled trial to see if such interventions work.

182
00:21:11,190 --> 00:21:19,620
Now when this phone call occurred in 2020, out of the five phases that the project had, we had already gone through three phases.

183
00:21:20,460 --> 00:21:26,700
The first phase was the Wearability phase where we would see the acceptability of such wearables.

184
00:21:27,180 --> 00:21:32,010
The second phase was testing the accuracy of these devices to be used in the clinical setting.

185
00:21:32,400 --> 00:21:44,040
And then the third phase was seeing if this can be used on the ward and any data losses and communication issues that we have in the overall system,

186
00:21:44,040 --> 00:21:50,309
which links to what I spoke to at the beginning about how to integrate these systems in

187
00:21:50,310 --> 00:21:54,810
a useful manner in the hospital and to cook the wireless communication between them.

188
00:21:54,810 --> 00:21:57,600
So at this point in in time,

189
00:21:57,600 --> 00:22:06,930
we had already been through these three steps which were quite important to give us the confidence that this system does work,

190
00:22:07,380 --> 00:22:17,220
that the wearables are accurate. The wear ability has passed our test and we had already started testing it in the hospital,

191
00:22:18,570 --> 00:22:26,729
the phases four and five that were still meant to be done and therefore there was no urgency at the time was phase four was to test it on the ward,

192
00:22:26,730 --> 00:22:32,129
on patients with the clinical staff being blinded and then phase five having it in a full

193
00:22:32,130 --> 00:22:35,850
randomised controlled trial where some patients would be blinded and otherwise wouldn't,

194
00:22:35,880 --> 00:22:46,050
others wouldn't be blinded. But there's, there is a big step from having a system under constant supervision from research staff,

195
00:22:46,050 --> 00:22:58,920
either be engineers or research nurses and clinicians interested in research and having it placed in a hospital for use in Bailey,

196
00:22:59,640 --> 00:23:05,160
for use in daily care and in an emergency on top of that.

197
00:23:05,160 --> 00:23:13,230
In an emergency on top of that. And the next day that we did go to the hospital for that meeting at five,

198
00:23:13,410 --> 00:23:20,399
we went to the infectious disease walk and they there were a few senior people from the hospital there.

199
00:23:20,400 --> 00:23:24,810
We had the system and I was wearing some of the wearables.

200
00:23:27,720 --> 00:23:32,280
And then at some point I remember Lionel saying, Christine, can you please give a demo?

201
00:23:33,150 --> 00:23:37,290
And at that point, I think my heart rate skyrocketed to about 190.

202
00:23:38,430 --> 00:23:49,770
And as the my heart rate wouldn't go down below 190, I remember one of the clinicians came next to me and took took my hand to see for now,

203
00:23:50,070 --> 00:23:55,530
I don't know if they were concerned about my health, but they wanted to see that be 190 matches.

204
00:23:55,530 --> 00:24:02,759
Actually, it was actually that the the the the conclusion was that they agreed for the

205
00:24:02,760 --> 00:24:09,479
system to be implemented on the on the infectious disease ward and weren't.

206
00:24:09,480 --> 00:24:17,280
And we had approximately three weeks from then to to get it to to get it fixed.

207
00:24:17,280 --> 00:24:25,230
Now I say three weeks now in hindsight, but we didn't really know when the first COVID patients was going to come to the UK and.

208
00:24:25,260 --> 00:24:32,510
We didn't know how fast it was going to spread. Then we we have to remember that no one knew what was going to happen at the time.

209
00:24:32,520 --> 00:24:45,120
There was concerns about staff safety. There were concerns about the dangerous ity of this virus.

210
00:24:45,240 --> 00:24:56,790
And probably it was seen that this system could help with the exposure that the staff might have in monitoring these patients,

211
00:24:56,790 --> 00:25:01,500
but also having a closer and keeping a close eye on these on these.

212
00:25:02,660 --> 00:25:06,440
On these patients that got infected at the beginning.

213
00:25:07,460 --> 00:25:10,100
So we should but we just talk a little bit more about the wearables themselves.

214
00:25:10,250 --> 00:25:13,820
So what the what what was the the full package of wearables looked like?

215
00:25:13,940 --> 00:25:22,170
What did they do? So the full package of wearables is a it's composed of two of two components.

216
00:25:22,190 --> 00:25:35,240
The first one is a patch, um, manufactured by, um, vital connect in the United States and it's placed on the left side of the chest.

217
00:25:36,380 --> 00:25:42,620
And with this patch you get the EKG and you get heart rate, you get respiratory rate.

218
00:25:43,580 --> 00:25:48,409
And you also at the time got skin temperature, which we didn't use.

219
00:25:48,410 --> 00:25:59,150
And then the second the second wearable is the is the pulse oximeter that you attach to the rest of the the main component is on the wrist.

220
00:25:59,540 --> 00:26:06,590
And then it is linked to a very soft plasticky finger probe that you place on the finger.

221
00:26:06,920 --> 00:26:13,160
And with this device, you managed to get the heart rate and the blood oxygen saturation.

222
00:26:13,180 --> 00:26:18,200
So the level of oxygen that you have in your blood with COVID, that seemed to be extremely relevant.

223
00:26:18,710 --> 00:26:23,390
You needed the you had these patients that were doing fine.

224
00:26:23,390 --> 00:26:28,460
And then all of a sudden they deteriorated to levels which were below accepted.

225
00:26:29,390 --> 00:26:35,600
And then the the pulse oximeter was was probably the key in this in this scenario.

226
00:26:36,350 --> 00:26:42,829
And those both those two things transmit wirelessly to a central monitor somewhere else, a monitor.

227
00:26:42,830 --> 00:26:54,680
So these both these devices transmits to a tablet which is placed in the same in the in the vicinity of the of the patient.

228
00:26:55,130 --> 00:27:00,110
And then this tablet transmits all the data to a dashboard which is at the central nurse's station.

229
00:27:00,590 --> 00:27:10,100
So you would have multiple tablets with different patients, which would then all transmit this information to the central to the central station.

230
00:27:10,310 --> 00:27:17,450
So essentially the nurses can see how the patient is doing without actually having to come in and make those recordings themselves.

231
00:27:17,690 --> 00:27:25,940
Exactly. So they would be able to stay at the at the nurse's station and then assess if the patient is doing well,

232
00:27:25,940 --> 00:27:33,740
if they're improving, deteriorating now because this device this system hasn't been ce mark on the time,

233
00:27:34,130 --> 00:27:37,100
we wouldn't we weren't able to have audible alerts,

234
00:27:37,310 --> 00:27:43,969
meaning whenever there was a duration detected automatically by the system that would alert the nurses,

235
00:27:43,970 --> 00:27:50,300
but they would still be able to see the trend that the patient was having over time.

236
00:27:51,080 --> 00:28:05,540
And initially, every time that the nurse had to assess the patient, they had to use the full, um, the full clinical gear.

237
00:28:06,620 --> 00:28:18,170
And this system probably allowed them to go to therapy, especially to go in a bit less often into the isolation room than, uh, than an otherwise.

238
00:28:18,710 --> 00:28:24,000
But it is important to note that this wasn't replacing their usual care.

239
00:28:24,000 --> 00:28:30,409
I mean, if they had to go in to take their observations, they would still go in and take the observations,

240
00:28:30,410 --> 00:28:38,750
but it would allow a better time management in the ward and also allowed them to keep an eye if something

241
00:28:39,170 --> 00:28:46,639
went wrong rather than going and having a look through the window constantly going around the world.

242
00:28:46,640 --> 00:28:53,810
Now, these patients that the system was placed on weren't extremely sick patients, extremely sick patients.

243
00:28:53,810 --> 00:29:00,830
They were either in intensive care or the people, the patients that were doing very, very well.

244
00:29:02,570 --> 00:29:07,550
They were in this infectious disease ward at the beginning, but maybe not placed on the monitor.

245
00:29:09,830 --> 00:29:15,020
As we've all learned afterwards. We were all told to stay at home if we got we got sick.

246
00:29:15,020 --> 00:29:16,150
So the patient,

247
00:29:16,160 --> 00:29:23,690
the system was probably placed on patients that were at risk of deteriorating or that that the clinical staff wanted to keep a closer eye.

248
00:29:24,290 --> 00:29:33,530
And what the system allows you on top of all the things we've said is rather than having snapshots of how the patient is every four or five,

249
00:29:34,400 --> 00:29:38,630
6 hours from the observations that nurses take,

250
00:29:38,900 --> 00:29:47,299
you would be able to get a data every 5 minutes and you would be able to get at a trend of what

251
00:29:47,300 --> 00:29:53,530
happens even in between those observations that are taken on track and trace systems in the hospital.

252
00:29:55,460 --> 00:30:01,970
So, um, one funny thing was funny is that we.

253
00:30:03,420 --> 00:30:12,600
Interesting is that some of the younger patients would take their oxygen mask off at some point and this would immediately be picked

254
00:30:12,600 --> 00:30:19,049
up by the system because their CO2 would drop and the nurses would know that this happened just by looking at the dashboard.

255
00:30:19,050 --> 00:30:23,820
So they would immediately go and probably say, put the mask back on and then they would put it on.

256
00:30:24,180 --> 00:30:27,419
But if you didn't have such a system,

257
00:30:27,420 --> 00:30:37,530
then you'll be more difficult to pick up the instances where patients to deteriorate in between between observations.

258
00:30:38,100 --> 00:30:41,970
So I slightly interrupted you to talk about that. So you talked about this.

259
00:30:42,060 --> 00:30:46,130
You had this three week window to. Yes. To get the whole thing set up.

260
00:30:46,140 --> 00:30:51,090
What what kind of a challenge was that? It was a it was a big challenge, I think.

261
00:30:53,460 --> 00:31:00,690
I don't think I've worked so intensely ever before or even after.

262
00:31:01,170 --> 00:31:05,430
So I think from from that frontline phone call up until June,

263
00:31:06,720 --> 00:31:19,140
they were probably the most intense working period of my life that we have to build a system that was 24 seven reliable.

264
00:31:20,520 --> 00:31:22,710
And that was the difficulty of it.

265
00:31:22,860 --> 00:31:30,120
They there couldn't be a nurse that went and pressed the button if something went wrong, because that was the whole point of the system.

266
00:31:32,010 --> 00:31:38,219
The point was systems that you would have that tablet inside the isolation room with the patient and then they would constantly monitor.

267
00:31:38,220 --> 00:31:44,640
You can expect that if the app or if our system crashes, that they would go to the patient and say,

268
00:31:44,640 --> 00:31:47,100
Can you please press that button and follow this procedure?

269
00:31:48,150 --> 00:31:56,990
You can also expect a nurse to ask a nurse to go in to the isolation room and say, Oh, can you press those buttons on the tablet?

270
00:31:57,000 --> 00:32:00,210
Because the whole point of the system was to expose, to reduce exposure.

271
00:32:02,340 --> 00:32:10,980
So both Mauro, Marco and myself worked very, very hard in trying to create this reliability of the system,

272
00:32:11,310 --> 00:32:17,280
but also provide a user friendly interface for the nurse, for the clinical staff to use.

273
00:32:19,020 --> 00:32:28,590
Up until then, I think it was engineers that were mainly looking at the data and we had a dashboard designed for engineers.

274
00:32:29,640 --> 00:32:33,690
There is a you can have the best system ever if you don't provide the good

275
00:32:33,690 --> 00:32:38,010
user interface for people from other domains to use and understand the data.

276
00:32:38,550 --> 00:32:49,830
It won't work. And Marco focussed specifically on creating this user friendly dashboard and he did an amazing job.

277
00:32:50,220 --> 00:32:58,560
We split the tasks amongst us with Marco working specifically on the user interface model,

278
00:32:58,770 --> 00:33:10,620
working on all the back end of of getting the data, making sure it is in a suitable form to be presented on the dashboard.

279
00:33:10,980 --> 00:33:19,910
And then I focussed on the app on the tablet side of getting this data from these wearable from these wearable devices reliably with,

280
00:33:19,970 --> 00:33:21,930
with Bluetooth, which was already the case.

281
00:33:21,930 --> 00:33:28,710
But we all needed to make sure that none of our sections break at any point and if it breaks, they automatically recover.

282
00:33:29,550 --> 00:33:37,050
Um, and there was very little overlap between our work which focussed very hard.

283
00:33:37,050 --> 00:33:39,450
We spoke, there was very intense communication.

284
00:33:40,170 --> 00:33:52,409
Um, but we each worked individually and then we put the pieces together and I think there were without exaggeration,

285
00:33:52,410 --> 00:34:02,190
between 18 and 20 work hours per day working at least until mid of April, by which time there were COVID patients in the hospital.

286
00:34:02,370 --> 00:34:06,240
The first COVID patient, I think was on Monday, 23rd of March.

287
00:34:06,870 --> 00:34:10,259
Um, which ones?

288
00:34:10,260 --> 00:34:16,260
That first patient got registered. We got an extra level of adrenaline.

289
00:34:16,470 --> 00:34:22,860
At least they got and we worked even on it because then we could see that the system

290
00:34:22,860 --> 00:34:28,860
is used and then we had to fix everything that that wasn't entirely working.

291
00:34:30,030 --> 00:34:33,960
It was almost compulsive at some point of checking to see how the system works there.

292
00:34:34,140 --> 00:34:41,520
I don't think there was one hour where we wouldn't check either one of the three of us and we would call This isn't working.

293
00:34:41,520 --> 00:34:47,460
This the button needs to be so. Yeah, this is how it happened.

294
00:34:47,500 --> 00:34:52,380
Ooh. And did you have to interact a lot with the nurses as well in this phase?

295
00:34:52,410 --> 00:35:02,090
Absolutely. Yes. Yes. So the nurses were key in this because the the link between the staff at the hospital and the engineering team were.

296
00:35:02,200 --> 00:35:10,990
Nurses from the Critical Care Research Group, and they were the ones that one provided us with the context of what was happening and gave us.

297
00:35:12,330 --> 00:35:18,810
The mean. We even translated the needs to something that we would understand a bit better.

298
00:35:19,350 --> 00:35:23,370
Yeah. The nurses. This wouldn't have been possible without the nurses I've got.

299
00:35:23,530 --> 00:35:25,979
I focussed on the engineering side to start with,

300
00:35:25,980 --> 00:35:33,300
but this was a full collaboration with the people at the Critical Care Research Group and we worked very, very closely with them as well.

301
00:35:34,020 --> 00:35:41,970
Um, I remember the weekends having to call or we would message an email, the nurses as well.

302
00:35:42,120 --> 00:35:45,899
So yeah, we would, we would call and message the nurses.

303
00:35:45,900 --> 00:35:54,590
And then if, if we thought it was something worth discussing with, then they would go to the ward and also discuss with the staff there.

304
00:35:54,600 --> 00:35:57,660
So there was always this flow of.

305
00:35:59,970 --> 00:36:03,260
Communication between between us. Yeah, we would.

306
00:36:03,270 --> 00:36:08,010
For example, we wouldn't go as engineers just directly onto the world on our own.

307
00:36:08,130 --> 00:36:13,940
But we went a few times, but it was just an extreme and urgent situations.

308
00:36:13,950 --> 00:36:18,210
Otherwise we would always go with, uh, with the research nurses and research stuff.

309
00:36:20,130 --> 00:36:23,220
And so were once COVID patients started coming in,

310
00:36:23,220 --> 00:36:30,660
we're all those who weren't actually in the ICU being fitted with the wearable technology, or was it were you trying to do a comparison?

311
00:36:30,960 --> 00:36:38,250
So initially, I think they were placing all these patients that came in with COVID on the system.

312
00:36:38,250 --> 00:36:44,160
But as things evolved, I think they started selecting more of which patients they placed the system.

313
00:36:44,910 --> 00:36:55,860
Now the patch is very well accepted by patients, but the finger probe is a bit less because the finger gets a little bit moist.

314
00:36:56,250 --> 00:37:02,700
Now, depending on the frailty of the patient, they might accept it or might not accept it, depending on their mobility.

315
00:37:03,000 --> 00:37:10,440
The one advantage of this system is that the patients can remain mobile and in all respiratory diseases.

316
00:37:10,440 --> 00:37:21,749
It is important for people to remain mobile. But if a patient is bedbound and is also a bit frail and doesn't, um, doesn't want to wear it,

317
00:37:21,750 --> 00:37:28,889
they probably made a decision that they wouldn't wear if they wanted to because this patient is a bit frail and this is no non-mobile.

318
00:37:28,890 --> 00:37:33,480
Maybe they wanted to keep an extra eye. So this was a decision left for the clinical staff.

319
00:37:33,480 --> 00:37:39,360
But we did see an evolution from being placed on everyone in the beginning to selecting on who it is placed on.

320
00:37:40,440 --> 00:37:49,530
And this definitely evolved over time. And and the I mean, there have been a couple of publications now.

321
00:37:49,530 --> 00:37:57,780
So so how did the evaluation look once you had the system in place throughout, certainly the first wave.

322
00:37:58,000 --> 00:38:06,149
So, um, so in terms of, um, I think there's been this publication on the engineering side,

323
00:38:06,150 --> 00:38:13,650
there's further publications coming out on uh, the qualitative aspect of what the nurses did, what the nurses thought of the system.

324
00:38:14,130 --> 00:38:20,810
Um, there were a lot of interviews took place with the research nurses, uh,

325
00:38:21,570 --> 00:38:26,879
to try to understand if they found that useful in the who did those interviews.

326
00:38:26,880 --> 00:38:30,030
So, um, the team led by Sarah Wallem.

327
00:38:30,180 --> 00:38:39,600
All right, Peter Watkiss is the research group. Um, and I think hopefully that that would be they'll be released soon.

328
00:38:40,050 --> 00:38:46,440
I mean, has been submitted for review. And I think the system was useful.

329
00:38:46,890 --> 00:38:54,540
The, the main, the main things I would take is that it allowed them to have better time management.

330
00:38:56,250 --> 00:38:58,950
It allowed them to have some reassurance.

331
00:38:58,950 --> 00:39:08,250
Whenever they were a bit short staffed, they would place these devices on the patients and then they would be able to look, um,

332
00:39:08,580 --> 00:39:15,659
as they were passing at the nurses station, they would be able to look at the patient quickly and assess if they need further attention or not.

333
00:39:15,660 --> 00:39:22,200
And if they not, they could go to see the patients that actually needed attention, then come back 10 minutes later, see if the situation is the same.

334
00:39:23,040 --> 00:39:32,790
They would go see another patient. And the, the, the third aspect is the, um.

335
00:39:36,230 --> 00:39:40,490
This. Benefit of having a little bit more time.

336
00:39:40,670 --> 00:39:43,820
So I think I would I would say these are the main ones.

337
00:39:47,680 --> 00:39:55,240
Yeah. The I wouldn't be able to to say more without further analysis.

338
00:39:56,590 --> 00:40:00,820
I don't know. And I'll be continuing to use the system in the hospital at the moment.

339
00:40:01,690 --> 00:40:10,270
So, no, the system moved on to the next phases that we've discussed in the system was stopped in April 2021,

340
00:40:11,500 --> 00:40:19,209
and the use of the system could be matched on the waves of the of the pandemic

341
00:40:19,210 --> 00:40:24,700
that were happening in the UK with each wave in which each peak of admissions.

342
00:40:24,700 --> 00:40:31,540
We could also see a peak of people registered on the video system, which shows that the system was used by staff in an appreciated.

343
00:40:33,700 --> 00:40:39,220
I cannot say I think by everyone, but I think by the majority of of the.

344
00:40:39,250 --> 00:40:43,710
So it was entirely up to them whether they chose to use it. It was entirely up to them if they chose to use it.

345
00:40:43,720 --> 00:40:56,710
Yes. The in April 2021, the system stopped being used on the infectious disease ward and we continued with our work on the phase four and phase five.

346
00:40:57,400 --> 00:41:00,309
And now when we were speaking,

347
00:41:00,310 --> 00:41:09,460
we are now in the randomised controlled trial phase of the project and the system is deployed on the post-surgery ward at the hospital.

348
00:41:10,030 --> 00:41:17,200
So these are patients that have had a surgery and the system is placed on it and then to detect early deterioration that occurs,

349
00:41:17,200 --> 00:41:22,929
which is the main focus of the of the system and that as you say, that's a randomised controlled trial.

350
00:41:22,930 --> 00:41:27,250
So even some some patients are on the system and some are not.

351
00:41:28,040 --> 00:41:31,980
Yes. Yes. So and you also you mentioned blinding AI.

352
00:41:33,130 --> 00:41:36,730
It would be difficult to blind when you've got something obvious like a patch.

353
00:41:36,730 --> 00:41:43,660
So it's open. Open, blind, meaning they they, they have the patch and the pulse oximetry.

354
00:41:43,660 --> 00:41:50,830
But the values of those patients in the control group aren't being shown the right that are being shown at the nurses station.

355
00:41:50,830 --> 00:41:57,399
Right. Yeah. So you have some patients which have the patch and the pulse oximeter and also a blood pressure

356
00:41:57,400 --> 00:42:02,500
device which we've integrated and that have all these vitals shown at the nurses station.

357
00:42:02,710 --> 00:42:09,640
And whenever the system detects a deterioration, it has an audible alert for the nurse to say patient is deteriorating.

358
00:42:10,330 --> 00:42:18,640
And then there's another arm which has all these devices, but there is no audible alerts and values aren't being shown at the nurse station.

359
00:42:18,970 --> 00:42:26,990
And at the end, we'll be able to compare the two groups and see the outcome of does the wearable technology actually improve patient outcome?

360
00:42:27,260 --> 00:42:37,299
Oh yeah, that's very good. So I read that in addition to this, you did find an application for your community monitoring,

361
00:42:37,300 --> 00:42:42,550
blood pressure monitoring, and that was relevant to what was going on during COVID.

362
00:42:42,640 --> 00:42:52,180
Yes. So we discussed about them monitoring or remote monitoring of patients that had stroke.

363
00:42:53,560 --> 00:43:01,060
But there's also a large part of interest in our group about remote monitoring of blood pressure and gestational.

364
00:43:01,510 --> 00:43:08,410
Um, so in, um, in pregnancy, postpartum,

365
00:43:08,650 --> 00:43:24,580
but also during pregnancy and this work that happened at the same time with both Professor Richard Mamas and Professor Lucy McKillop, um,

366
00:43:26,140 --> 00:43:35,799
gained traction with having these, having pregnant women being monitored for their blood pressure at home because the

367
00:43:35,800 --> 00:43:42,010
number of appointments or face to face appointments at the hospital suddenly dropped.

368
00:43:42,160 --> 00:43:48,340
And they still wanted to know that these these women are being done,

369
00:43:49,240 --> 00:43:53,170
are being monitored and don't have high blood pressure because of the risk of pre-eclampsia.

370
00:43:53,230 --> 00:43:55,090
Because of the risk of pre-eclampsia. Exactly.

371
00:43:56,860 --> 00:44:06,040
And the work that the group has been has been doing a lot of work on this in trials which were called bump and pop HD.

372
00:44:07,360 --> 00:44:19,960
And when the pandemic hit, the work from bump went on to being spun out into a product that was acquired by Send Sign Health,

373
00:44:20,980 --> 00:44:28,780
and that went on to being deployed in in several trusts across the country and also at the UAH here in Oxford.

374
00:44:30,220 --> 00:44:34,420
And the principle was again quite simple,

375
00:44:34,720 --> 00:44:46,330
similar to the monitoring in in in stroke patients were pregnant women would take their blood pressure readings one.

376
00:44:47,550 --> 00:44:51,780
Once a day or once a week, depending on the schedule set by their doctor.

377
00:44:52,170 --> 00:44:58,140
And these readings would be sent automatically to a dashboard that can be reviewed by the clinician.

378
00:44:58,560 --> 00:45:05,730
But if their blood pressure was very high, then they would be told to seek immediate medical advice.

379
00:45:06,870 --> 00:45:13,679
And I think from the feedback that we've had, there was at least one instance where this proved to be extremely,

380
00:45:13,680 --> 00:45:18,270
extremely useful because maybe had very, very high blood pressure that she didn't know.

381
00:45:18,270 --> 00:45:26,400
And if it wasn't for such a system, she would have probably would have stayed the blood pressure high from multiple days before being picked up.

382
00:45:26,520 --> 00:45:31,260
Bye bye bye by an appointment.

383
00:45:33,150 --> 00:45:39,389
So those that also added an extra workload that Maura and myself had to work

384
00:45:39,390 --> 00:45:44,190
in the same period of March to June when we worked very closely with people.

385
00:45:44,190 --> 00:45:53,610
That sense in order to have this knowledge transfer of bump to what they called was BPM health was their commercial side.

386
00:45:53,610 --> 00:45:58,860
And then we did also a bump plus, which was the research project that we made.

387
00:45:59,160 --> 00:46:04,860
So what was so you had your your research team working with the commercial company.

388
00:46:04,980 --> 00:46:11,580
What were the particular tasks that you and your colleagues had to had to deal with for them?

389
00:46:12,630 --> 00:46:22,890
So it was mainly on the on the technical side of porting this code from being into a from being a research project,

390
00:46:23,460 --> 00:46:26,850
which is done up to a certain level of standard,

391
00:46:26,860 --> 00:46:35,759
but it is an industry standard up to a the product which would go through CE marking and would go

392
00:46:35,760 --> 00:46:44,790
through all the regulatory approvals to be classed as a to be classed as a medical use software,

393
00:46:47,880 --> 00:46:58,800
not only approved from a standardised point, but also to be able to scale to multiple thousands of patients or to go live nationwide.

394
00:46:59,250 --> 00:47:03,690
As much as we want to, we we cannot with with the system that we had built,

395
00:47:03,690 --> 00:47:09,990
we would only probably go without completely slowing down everything, maybe to a thousand or 2000 patients.

396
00:47:10,380 --> 00:47:19,530
While with the solution that Samsung and the engineers, then you could go to hundreds of thousands of patients and scale very easily.

397
00:47:19,590 --> 00:47:31,620
So this transition was was extremely, extremely useful even for us to see how how things are done in industry.

398
00:47:32,310 --> 00:47:39,600
And it was I think we've learned a lot. So maybe the the engineers from Genzyme learned a little bit about the product,

399
00:47:39,870 --> 00:47:52,300
but we learned how how things are done on an on a large scale, how to do, how to deploy products on a.

400
00:47:53,900 --> 00:47:58,070
And is that envisaged as being useful even when there isn't a pandemic going on?

401
00:47:58,940 --> 00:48:04,340
Absolutely. I think that's continued to be useful even now because of the pressure on GP's

402
00:48:04,550 --> 00:48:08,810
presumably at the moment and hospitals because the pressure groups of hospitals,

403
00:48:09,860 --> 00:48:18,620
it also provides reassurance to patients. It it, it allows probably a reduction in appointments and less needed to.

404
00:48:19,060 --> 00:48:26,150
Now there's this high pressure on GP's that even I think the patient experience going to

405
00:48:26,150 --> 00:48:30,049
the hospital at Oxford is sometimes not ideal because the parking issues that take time,

406
00:48:30,050 --> 00:48:38,660
so people mention all sorts of reasons of why they don't want to maybe go away from their home or from their local area.

407
00:48:39,410 --> 00:48:42,260
For those people that do want to go, there should always be an option.

408
00:48:42,260 --> 00:48:47,090
But for those that still prefer to be home and have this remote monitoring solution,

409
00:48:47,180 --> 00:48:53,700
I think it's great that now the hospital does have an option of monitoring these patients remotely.

410
00:48:53,990 --> 00:49:03,470
And even even if the dashboard is not the dashboard where all these blood pressure readings are monitored constantly by GP's,

411
00:49:03,500 --> 00:49:07,610
by having these automated algorithms that provide instructions on the next step to do.

412
00:49:09,730 --> 00:49:14,820
E You can have this failsafe mechanism where people are told,

413
00:49:15,060 --> 00:49:19,200
please seek urgent assessment within 4 hours of your maternity unit or please

414
00:49:19,200 --> 00:49:26,040
go and see your GP called GP now because your blood pressure is very high. So simple steps can have big impact.

415
00:49:27,810 --> 00:49:34,860
And I think this is what works the best, especially in an area where this is being deployed on a larger scale, maybe in 30 years,

416
00:49:34,860 --> 00:49:44,639
or will speak about having automated passive algorithms and patients waking up in the

417
00:49:44,640 --> 00:49:48,390
morning and taking all their vitals by being in the bathroom sitting on the scale.

418
00:49:48,390 --> 00:49:55,380
But we need to start with something simple and probably this is something simple is what has been implemented up until now.

419
00:49:56,160 --> 00:49:59,970
And this is I mean, how satisfied do you feel that you've met, you know,

420
00:50:00,060 --> 00:50:04,020
what you said after your do so you really wanted to do something where you could see an impact.

421
00:50:04,380 --> 00:50:08,450
Do you feel you've achieved that? Mm hmm.

422
00:50:10,940 --> 00:50:14,519
Um, I. I don't know if I. I wouldn't say I've achieved it.

423
00:50:14,520 --> 00:50:23,099
I'm still working on it. I'm very I'm very happy and feel one overly privileged of the position I'm in, the group.

424
00:50:23,100 --> 00:50:27,960
I'm working over the privilege also for the fact that I had the opportunity to help during the pandemic.

425
00:50:28,590 --> 00:50:35,220
But, um, if I think about achieving it, it almost means that I'm closing the chapter and I'm still working on it.

426
00:50:35,430 --> 00:50:39,839
I still want to continue my efforts. I still want to go I want to go further.

427
00:50:39,840 --> 00:50:44,130
I want to go further in applying this technology and pushing it forward.

428
00:50:44,140 --> 00:50:50,310
Uh, you know, I don't want to think that it's finished yet because then I think it will become lazy.

429
00:50:51,180 --> 00:50:57,710
So I always have to be on the, um, my toes.

430
00:50:57,870 --> 00:51:01,290
They know I want more. I want to push more, I want to integrate more.

431
00:51:01,800 --> 00:51:09,720
Um, I want to deploy these systems more. I know this isn't about, um, just engineering work.

432
00:51:09,720 --> 00:51:12,990
I think we need interdisciplinary collaborations with clinicians.

433
00:51:13,530 --> 00:51:19,620
As much as I try to understand the I won't be able to get out the soul, the subtleties of what's going on,

434
00:51:21,120 --> 00:51:29,190
and I want to continue collaborating with people to say this is the solutions that we have can be useful in your domain and keep,

435
00:51:30,060 --> 00:51:35,640
keep, keep doing that. But I am I am happy. I am very happy that I managed to apply some of it.

436
00:51:36,330 --> 00:51:39,450
Uh, what project are you working on now?

437
00:51:39,960 --> 00:51:51,450
So, um, I'm working mainly on this phase five of the randomised controlled trial and then we're continuing the work with primary care on,

438
00:51:52,830 --> 00:51:55,920
on remote monitoring of blood pressure in pregnancy.

439
00:51:56,490 --> 00:52:01,170
So that is, that is the it's Wednesday.

440
00:52:03,210 --> 00:52:12,840
So the, the primary care project's on, uh, remote blood pressure monitoring in this time, also including medication.

441
00:52:12,840 --> 00:52:16,320
So automatic obliteration and down duration of medication.

442
00:52:16,920 --> 00:52:21,300
Then these are my two main areas of, of work.

443
00:52:22,080 --> 00:52:32,200
And on the video side, interesting, we, we've been applying this on the challenge trials at the university where um,

444
00:52:32,760 --> 00:52:38,640
we've been placing these wearable devices actually only one the patch,

445
00:52:39,450 --> 00:52:40,130
um,

446
00:52:41,790 --> 00:52:52,079
the patch with the another wristband that monitors blood pressure on these patients that are exposed to the virus in a very controlled environment.

447
00:52:52,080 --> 00:52:58,049
And that is also to get a better understanding of their of the vitals evolution while exposed to

448
00:52:58,050 --> 00:53:05,940
the virus with hopefully some interesting results on predicting on when one might be infectious.

449
00:53:05,940 --> 00:53:10,890
So we are very eager to see. Um, great.

450
00:53:11,520 --> 00:53:18,030
Can we just go back to, um, you know, how the, how working through the pandemic impacted on you personally?

451
00:53:18,030 --> 00:53:24,180
I mean, to what extent did you feel threatened by the virus itself? Did you did you feel you might become ill yourself?

452
00:53:26,310 --> 00:53:38,250
I didn't feel threatened by it until one episode, which was quite revealing of how scared I was up to that point.

453
00:53:38,250 --> 00:53:50,490
But we during this period of implementation, we would go to the ward and present the system to the to the to the clinical staff.

454
00:53:51,180 --> 00:53:59,430
Patients had already been registered. And I went to the ward with my colleagues from the clinical side with the siren cross,

455
00:54:00,480 --> 00:54:04,200
and there were around ten, 15 clinicians in front of the dashboard.

456
00:54:05,010 --> 00:54:13,320
And I was there giving a very short presentation, usually less than 5 minutes because they were busy.

457
00:54:14,430 --> 00:54:17,499
Um, of course they were obviously overly stretched.

458
00:54:17,500 --> 00:54:22,049
One of the things they had to do so the shortest, the better.

459
00:54:22,050 --> 00:54:30,360
And during that five minute that there was a port they're passing by behind with a, um, fortunately with someone that had just died,

460
00:54:30,600 --> 00:54:39,060
wrapped up in the, um, in a black bag and oh, I stopped talking and everyone well looked then.

461
00:54:39,510 --> 00:54:43,139
But while the portal passed, no one said a word.

462
00:54:43,140 --> 00:54:46,049
And then once the water passed, we just continue the conversation.

463
00:54:46,050 --> 00:54:58,070
But as soon as I got back to the office, then I realised actually this is very serious and yeah, and we will masked in that meeting no less.

464
00:54:59,580 --> 00:55:07,710
So at the time the, the masks there were, there was this lack of PPE even for clinicians I think maybe.

465
00:55:07,790 --> 00:55:12,379
Some clinicians were, most especially the ones that were constantly going in and out.

466
00:55:12,380 --> 00:55:21,800
But as an engineer, I wasn't most of the time. I think the masks came in much later, but I didn't feel the fear of the mask.

467
00:55:21,800 --> 00:55:26,660
The mask wasn't something natural at the time and the windows were open.

468
00:55:26,960 --> 00:55:33,320
The bottom line is, I didn't catch COVID until much later, a year and a half later, and I caught it four months ago.

469
00:55:34,880 --> 00:55:39,260
But even after this episode, I still went to the ward without mask.

470
00:55:39,800 --> 00:55:47,990
I would never go into the patient isolation. It was only in the main corridor to the nurse's station for five or 10 minutes, and then I'd come out.

471
00:55:49,190 --> 00:55:55,639
And that's what the other clinicians said, that as soon as they started wearing a mask,

472
00:55:55,640 --> 00:56:00,070
my research colleagues from the research nurses, I also started wearing one.

473
00:56:00,080 --> 00:56:09,710
But at the time it wasn't the mask policy. Yeah, I got scared I think for that evening.

474
00:56:11,090 --> 00:56:13,940
But then there were so many things that had to be done that I forgot about it.

475
00:56:14,240 --> 00:56:25,460
And I just went on and I constantly thought, I'm young and fit and I mean, I should, you know, I should be.

476
00:56:27,170 --> 00:56:31,819
I'm happy that I could help. Um, and what about your family members?

477
00:56:31,820 --> 00:56:34,969
I mean, I don't know how things were in Romania. Sorry.

478
00:56:34,970 --> 00:56:42,560
In in France. In France, yes. But my, my, my parents were, um, staying at home mainly.

479
00:56:42,830 --> 00:56:52,159
I think my dad went to work a few times because he, he worked for the fire men for the fire department as a computer scientist.

480
00:56:52,160 --> 00:56:57,770
But I think he had to go a few times. But my mom, my mother stayed at home, which I was quite reassured that they weren't travelling anymore.

481
00:56:58,670 --> 00:57:04,970
Um, and they were very supportive. Um, I, I cannot, yeah.

482
00:57:05,360 --> 00:57:13,879
Thinking that, you know, knowing that they're there, seeing their only child is working in this hospital with a virus that no one knows.

483
00:57:13,880 --> 00:57:20,000
I think they, they reacted really well. They never showed their, um, emotions or anger.

484
00:57:20,030 --> 00:57:24,200
Oh, no, not anger. Anger is the wrong word. There is an alarm.

485
00:57:24,200 --> 00:57:29,540
Yes, alarm. Yeah, they that they they called me, I remember only once the evening,

486
00:57:29,540 --> 00:57:34,399
but this was a bit I think probably they looked at on TV and they got a bit panicked and I spoke with them.

487
00:57:34,400 --> 00:57:38,030
Everything's fine, you know, the way and that was it. But none of they were very supportive.

488
00:57:38,780 --> 00:57:44,900
So, um, yeah, yeah. And you mentioned that worked very, very intensely between March and June.

489
00:57:45,680 --> 00:57:58,100
How did that impact on your wellbeing and did you have to take steps to try and retain your sanity and not during the time?

490
00:57:58,220 --> 00:58:04,190
No, no. So I worked very, very closely with, um, with Moreau at the time.

491
00:58:04,190 --> 00:58:06,409
We also, we were close to Mark Colvin. Marco.

492
00:58:06,410 --> 00:58:12,800
We could do ourselves some of the things remotely, but with Mauro I worked very closely and I would meet him every day.

493
00:58:13,190 --> 00:58:15,500
I think even though he was coming into the lab every day,

494
00:58:15,680 --> 00:58:22,670
we were coming into the lab up to a certain point where it wasn't, um, where we couldn't come any more.

495
00:58:24,590 --> 00:58:28,670
Although we always, we always had the opportunity because we got the letters of being.

496
00:58:28,760 --> 00:58:32,030
You were, you had the key work. We had the key work letter very, very early.

497
00:58:32,420 --> 00:58:39,680
Um, but I think, um, it was much more difficult to come here.

498
00:58:39,680 --> 00:58:44,749
So we started meeting at my place with Mauro, also Moros nursery class.

499
00:58:44,750 --> 00:58:47,660
He had a, a young child that at the time.

500
00:58:48,650 --> 00:58:54,620
And then with all these key workers being repositioned with the nursery, I think he got a nursery closer to where I live.

501
00:58:54,620 --> 00:59:04,519
So, um, we started working at my, at my house for, for a very long period of time, a long period, 2 to 3 months.

502
00:59:04,520 --> 00:59:06,170
But we would also come to the hospital.

503
00:59:06,950 --> 00:59:17,149
So I think how it worked is Monday to Friday we would work in my place and we would meet Saturday and Sunday at the hospital on our work days,

504
00:59:17,150 --> 00:59:25,130
would finish at midnight or one or two and morrow would drive me back and then he would go home and then the back at eight would be meeting again.

505
00:59:25,760 --> 00:59:33,890
Um, so I think that was our schedule for up until mid-May, early June.

506
00:59:34,730 --> 00:59:38,959
Um, you mentioned about impact, but I didn't see it as such.

507
00:59:38,960 --> 00:59:43,070
I enjoyed it during the time I was in the adrenaline rush.

508
00:59:43,070 --> 00:59:45,560
I'm, I'm a competitive person.

509
00:59:45,890 --> 00:59:59,330
Um, and I, well, I was much more competitive when I was in my early twenties and, but um, I only realised the impact it had after once it slowed down.

510
01:00:00,200 --> 01:00:05,929
So the first wave was very intense. We continued over the summer to implement these features.

511
01:00:05,930 --> 01:00:09,830
The second one came Wave. And I kept going to the hospital.

512
01:00:10,160 --> 01:00:18,320
But then I think in 2021, in January, I didn't have to go to the hospital that much or work as hard.

513
01:00:19,820 --> 01:00:27,530
And when it slowed down, then I had noticed how how tired I am.

514
01:00:28,220 --> 01:00:35,330
And then with the next lockdown that followed, where I didn't have to go to the office or to the hospital,

515
01:00:35,330 --> 01:00:42,140
and I said, well, given I don't have to go, I should maybe just stay at home and follow the rules like everyone.

516
01:00:42,290 --> 01:00:48,350
Then it got tough. It actually got tougher when I had just had to stay at home and work on my things but not be as active.

517
01:00:49,760 --> 01:00:55,819
And yeah, so this is not a it's the I know.

518
01:00:55,820 --> 01:01:00,290
This is the I think that's how I would summarise it.

519
01:01:00,290 --> 01:01:05,060
It was tougher after not doing during. I was very happy to be able to tell.

520
01:01:05,180 --> 01:01:11,659
Yeah. Yeah, no it was again that's I mean, I think you've answered the question that I've been asking everyone whether the

521
01:01:11,660 --> 01:01:16,520
fact that you were doing something that was contributing to the the the fighting,

522
01:01:16,520 --> 01:01:20,150
the virus helped to support your sense of well-being?

523
01:01:20,660 --> 01:01:24,680
Definitely. Yeah. Yeah. Felt useful.

524
01:01:26,780 --> 01:01:33,410
Yeah, it might be. We all look for this sense of utility.

525
01:01:34,130 --> 01:01:42,440
I had that sense of utility. I cannot compare my utility with the with the clinicians that were exposed to this every day.

526
01:01:42,440 --> 01:01:48,250
And they but it may be it was a false sense of utility that I enjoyed it.

527
01:01:48,380 --> 01:01:59,990
By the way, the impact of the system, maybe it's minimal in the large scale of things in my day had may well looking at the scale

528
01:01:59,990 --> 01:02:05,750
that we've had only 160 patients it might have had if it had an impact on some of those.

529
01:02:06,350 --> 01:02:11,750
It's already it was all worth the effort, I think. Oh, great.

530
01:02:12,350 --> 01:02:16,520
Um, I think we've just about got to the end so that the yeah,

531
01:02:16,520 --> 01:02:25,490
my final question is whether your involvement in this work over the pandemic has changed your thinking about what you might want to do in the future,

532
01:02:26,300 --> 01:02:31,250
and is there anything that you'd like to see change as a result of that experience?

533
01:02:36,660 --> 01:02:45,360
So for the future, it probably just strengthens my belief that I want to continue continue doing applied digital health or applied engineering.

534
01:02:47,340 --> 01:02:50,940
I was always keen on finding the next.

535
01:02:53,210 --> 01:02:57,620
Algorithm. Algorithm. The next thing that will can be published.

536
01:02:59,990 --> 01:03:04,850
Or that would have. I would go into a high impact journal.

537
01:03:05,420 --> 01:03:10,670
I'm less keen on that now and I'm working on getting applied things, small things that work, but getting applied.

538
01:03:11,690 --> 01:03:17,090
I just have a high interest of building up these models and simulating testing.

539
01:03:17,450 --> 01:03:21,500
But always in my mind now the objective is, can this be applied?

540
01:03:21,860 --> 01:03:30,620
And I'm trying to build things which are easy to to change so that they can easily be integrated into a product that can be used.

541
01:03:31,250 --> 01:03:38,270
So I think this is the one thing I reduce a little bit, the exploratory research for the applicability one.

542
01:03:39,380 --> 01:03:48,290
And number two is. I think it helped with, um, acceptance from all.

543
01:03:51,080 --> 01:03:56,150
For data from the entire spectrum of society of how these remote monitoring solutions are seen.

544
01:03:59,000 --> 01:04:08,030
If before people wouldn't probably trust them more be see the utility of them.

545
01:04:08,030 --> 01:04:11,540
Now they can probably see that. The can can.

546
01:04:12,520 --> 01:04:22,480
Save money and it can improve the health of the whole system by probably by hopefully saving lives and improving patients outcomes.

547
01:04:22,690 --> 01:04:25,120
So there is a I mean, yes, there is a fear.

548
01:04:25,480 --> 01:04:32,890
I mean, it's something you read in the media that, you know, robots will take over and there'll be a loss of human contact in health care.

549
01:04:33,880 --> 01:04:41,560
But that that's you feel that this kind of work is is managing to dispel that kind of thinking?

550
01:04:42,130 --> 01:04:46,240
I think that fear is probably coming from the unknown.

551
01:04:46,570 --> 01:04:49,510
Once you start to know the robots, you realise that there is.

552
01:04:49,510 --> 01:04:58,660
And the human contact is a very good point and I don't think it should replace the human contact.

553
01:04:58,690 --> 01:05:06,880
It should be used complementary to what's already there and help clinicians in their workload, which keeps increasing.

554
01:05:09,760 --> 01:05:12,220
It's not about replacing the coalition whatsoever.

555
01:05:12,370 --> 01:05:19,090
It is about giving them more filtered information so that they can make better informed decisions about it.

556
01:05:20,720 --> 01:05:31,370
And um, there are, there's one thing of, of looking at it only from a national perspective,

557
01:05:31,370 --> 01:05:38,089
and there's also one thing of looking at scaling this to other to other environments where they aren't as developed as

558
01:05:38,090 --> 01:05:45,440
in the health care system and where these types of monitoring can help detect conditions that they might not have.

559
01:05:47,690 --> 01:05:55,879
And if we look at surveys across the world, in other countries where where people don't have access to health care as well,

560
01:05:55,880 --> 01:06:00,500
their level of agreement of tech and what people called robots is much higher

561
01:06:00,500 --> 01:06:05,389
because they see the benefit of it might not be the same as seeing a person,

562
01:06:05,390 --> 01:06:11,660
but at least if there is something really bad with me, this machine will tell me that I have to see someone.

563
01:06:12,170 --> 01:06:18,889
And this is, I think, what the flow will always be. The the human will never be replaced in this.

564
01:06:18,890 --> 01:06:24,290
It will be the machine will tell me I have something really bad or something worth looking at.

565
01:06:24,290 --> 01:06:29,449
The human will look at, it will confirm, and then it will go further. It won't be the machine detected something.

566
01:06:29,450 --> 01:06:35,940
I have to go on the surgery table immediately. So. Yeah.

567
01:06:36,760 --> 01:06:41,330
No, I don't think robots, you know, I don't think it'll happen.

568
01:06:41,670 --> 01:06:50,360
The analogy that I keep in my mind is similar to agriculture. I think the same fears were happening when the optimisation of agriculture occurred, um,

569
01:06:51,240 --> 01:06:55,229
mainly in the year after the war and people were worried were going to get out of jobs.

570
01:06:55,230 --> 01:07:06,800
And, and in the end now we managed to have better food variety and better health because of this, um, intense.

571
01:07:06,810 --> 01:07:15,240
Well, that's what, which also has a lot of back falls, but that it allows us to have the standard of living that we have now.

572
01:07:15,660 --> 01:07:22,590
Um, and the society adapts. I think it will be be the same with this, um, with digital health.

573
01:07:22,920 --> 01:07:26,190
It will hopefully be part of people's life, but something passive.

574
01:07:26,850 --> 01:07:30,569
And we never know when the world is going to close again. Hopefully not soon.

575
01:07:30,570 --> 01:07:41,550
But uh, yeah, it is one other tool that we have now in our bag to say that we're ready for what happened, what occurs if we need to.

576
01:07:42,570 --> 01:07:46,110
I think that's how I would I would remember it.

577
01:07:48,420 --> 01:07:48,930
Remember it?