1
00:00:03,210 --> 00:00:09,780
Okay, So can you just start by saying your name and what your current position is and your association with Oxford University?

2
00:00:11,090 --> 00:00:13,350
Thanks, Georgina. So I am Sheetal Prakash.

3
00:00:14,180 --> 00:00:21,920
I am an associate professor in the Department of Statistical Sciences at the University of Cape Town in Cape Town, South Africa.

4
00:00:22,280 --> 00:00:27,110
And I'm also the director of the Modelling and Simulation Hub Africa,

5
00:00:27,290 --> 00:00:33,110
which is a disease modelling institute based in the Department of Statistical Sciences at City.

6
00:00:33,440 --> 00:00:39,680
I've had a long running a relationship with Oxford University for the last ten years or so,

7
00:00:40,580 --> 00:00:45,350
with many collaborations in disease modelling over the years,

8
00:00:45,950 --> 00:00:54,470
working with colleagues from Oxford University and in projects that are led by Oxford University around the world supporting malaria modelling.

9
00:00:55,100 --> 00:00:57,290
But more recently from 2017,

10
00:00:57,440 --> 00:01:03,470
I've been appointed as an honorary Visiting Research Fellow in tropical Disease Modelling in the Nuffield Department of Medicine.

11
00:01:03,830 --> 00:01:10,459
That's great. That's very comprehensive. Thank you. So without telling me your whole life story, but going back to the very beginning,

12
00:01:10,460 --> 00:01:15,890
how did you first get interested in science becoming a becoming a research scientist?

13
00:01:17,300 --> 00:01:24,410
So I was I have quite a varied background why I have always been good and good at maths and drawn to,

14
00:01:24,410 --> 00:01:29,899
to to the subject of mathematics and the application of mathematics And one of the most well advertised

15
00:01:29,900 --> 00:01:36,590
routes to apply your mathematics when you're in school is to follow a path of in the finance world.

16
00:01:37,100 --> 00:01:43,969
And that's what I did. I my, my undergraduate degrees were in actuarial science and in quantitative finance and statistics,

17
00:01:43,970 --> 00:01:48,709
but I was very bored in the field, not with the mathematics, but with the application of it.

18
00:01:48,710 --> 00:01:57,830
I didn't find it attractive at all. And when I moved across to these statistics department to do a master's degree in operations research,

19
00:01:58,100 --> 00:02:02,600
I came across a variety of other ways in which one could use maths to do problem solving

20
00:02:02,930 --> 00:02:07,430
and which has which institution where you in at the time at University of Cape Town.

21
00:02:08,750 --> 00:02:16,190
Yes. Yes. And we and at that point I came across a few publications on using disease,

22
00:02:16,760 --> 00:02:21,200
using mathematical methods to manage disease, and that was very interesting.

23
00:02:21,200 --> 00:02:26,809
And I was drawn into public health immediately. I also at the same time was whilst doing my masters,

24
00:02:26,810 --> 00:02:34,340
was working at a health economics unit based at the School of Public Health and Family Medicine at University of Cape Town.

25
00:02:34,490 --> 00:02:41,930
And that gave me my first foray into looking at the into working with public health data and having your work make an impact.

26
00:02:41,930 --> 00:02:50,569
Very early on some of these statistics. The work I did in maternal health was referenced by the W.H.O. in forming guidelines and so on.

27
00:02:50,570 --> 00:02:58,040
And so I got to got to understand that that even quantitative applications in health can make can make a difference.

28
00:02:58,190 --> 00:03:04,730
And that's how I got drawn into data based evidence of evidence making in health.

29
00:03:05,090 --> 00:03:09,050
Mm hmm. So can you give me an example? So, I mean, what what have you.

30
00:03:09,530 --> 00:03:12,770
So malaria has been an area, a big area of interest of yours.

31
00:03:13,130 --> 00:03:18,230
How is data science applied in trying to understand and grapple with the issue?

32
00:03:18,350 --> 00:03:26,749
No. Let's start with another question first. How big a public health issue is malaria at the moment globally still, as a malaria is,

33
00:03:26,750 --> 00:03:33,230
would you believe that malaria is curable and preventable for the last 4000 years or so?

34
00:03:33,440 --> 00:03:38,330
But in the last year, we had 247 million cases worldwide.

35
00:03:38,660 --> 00:03:47,959
Malaria is also not just a problem with morbidity that largely children die from it, but it's a mortality rather.

36
00:03:47,960 --> 00:03:53,690
But it's also a great problem of morbidity in that it's a disease as linked to poverty and it makes one ill.

37
00:03:53,690 --> 00:03:57,229
But you can't go to work and if you can't go to work, you can't earn a wage.

38
00:03:57,230 --> 00:04:00,620
And if you can't earn a wage, there's a vicious cycle of poverty continues.

39
00:04:00,860 --> 00:04:04,010
And so you find malaria is a massive global health problem.

40
00:04:04,280 --> 00:04:12,649
And in the last ten or 12 years, or perhaps actually a bit longer now almost a year, a bit longer, 15 years or so,

41
00:04:12,650 --> 00:04:21,469
we've had a global push of political will towards eliminating malaria from the world because it is preventable and treatable.

42
00:04:21,470 --> 00:04:24,890
One feels like it should be an easy problem to solve. It most certainly isn't.

43
00:04:25,520 --> 00:04:30,979
But because of all of this political will, countries are doing their best to try eliminate malaria.

44
00:04:30,980 --> 00:04:39,860
And so all scientific hands are on board. And in order to support decision making to to achieve this goal.

45
00:04:40,850 --> 00:04:46,460
And yes, I go back to my previous question how does data science and statistics contribute to that goal?

46
00:04:47,240 --> 00:04:54,740
So it so it's not obvious actually. The you know, we have you have data that emerges from the management of malaria.

47
00:04:54,740 --> 00:04:59,490
So malaria in many cases is in many countries is a what you call a notifiable disease.

48
00:04:59,490 --> 00:05:03,559
So anyone who presents with a malaria case has to notify the government and records

49
00:05:03,560 --> 00:05:07,459
are kept probing questions are asked the demographics of the person in question,

50
00:05:07,460 --> 00:05:10,580
they travel patterns and so on are all are all recorded.

51
00:05:11,140 --> 00:05:20,350
But then if we have to make decisions on the many tools that can be used to combat malaria and decrease malaria incidents in country,

52
00:05:20,560 --> 00:05:28,990
it comes down to a question of how best can we spend the lot of money that we have or how, you know,

53
00:05:29,610 --> 00:05:36,219
the resources are one thing, but but from a health perspective, what is the most impacting impacting interventions?

54
00:05:36,220 --> 00:05:39,290
Should we be using bednets? Should we be trying out a certain drug?

55
00:05:39,310 --> 00:05:44,440
Should we be increasing access to health care or improving improving surveillance systems?

56
00:05:44,710 --> 00:05:47,770
These are all questions that we can use mathematics to solve,

57
00:05:47,950 --> 00:05:55,270
and we do that by creating a simulation or a virtual reality of the malaria situation on the ground,

58
00:05:55,420 --> 00:06:01,180
accounting for not just how malaria or malaria transmits from mosquito to human back to mosquito and so on.

59
00:06:01,420 --> 00:06:10,629
But also the health system challenges that a country is faced with and the population behaviour and demographics and so on in the inner country.

60
00:06:10,630 --> 00:06:17,830
So we basically build this whole world, this whole health system and world, all the malaria and on a computer.

61
00:06:17,980 --> 00:06:26,500
And then we try out various interventions. We can ask ourselves if we want to increase bed net coverage, what would the impact on the cost of that be?

62
00:06:26,500 --> 00:06:31,330
And we can do so in a fraction of the time it would take to implement a full clinical trial.

63
00:06:31,510 --> 00:06:37,150
And so there are many savings that one can make by building simulations using mathematics.

64
00:06:37,810 --> 00:06:42,280
And so through through the systems of of equations and computer code,

65
00:06:42,430 --> 00:06:50,620
we are therefore able to generate many cost savings and support better scientific understanding of how malaria transmits.

66
00:06:50,620 --> 00:06:55,150
And so that that enables us to be able to combat it from a more informed perspective.

67
00:06:56,170 --> 00:07:06,280
So you say you've been saying we you yourself are based in South Africa, but there are I understand you're part of a global collaboration.

68
00:07:06,280 --> 00:07:13,240
What what other parts of the world have you collaborated with and have you have you actually been based yourself elsewhere?

69
00:07:14,200 --> 00:07:22,959
I've been based in South Africa for mostly patriotic reasons, but I've worked in the end.

70
00:07:22,960 --> 00:07:26,170
The beauty of collaboration is that you do get to work all around the world.

71
00:07:26,170 --> 00:07:35,680
So with my colleagues at also university and at other universities, we have done malaria investment cases for the 22 countries of the Asia Pacific,

72
00:07:35,860 --> 00:07:40,959
for example, I have led and I've led investment cases in the Guyana Shield.

73
00:07:40,960 --> 00:07:49,870
So the countries at the top of South America, for many countries in Africa, for many other groupings in Melanesia and so on.

74
00:07:49,870 --> 00:07:58,000
So we've actually done a range of malaria investment cases and other disease and disease based work all all around, all around the world.

75
00:07:58,660 --> 00:08:07,030
So in terms of a global consortium, when I say we, I was referring to models in in general, they are groups of malaria models.

76
00:08:07,030 --> 00:08:14,229
They are individual malaria models, a network of Oxford modellers and myself and economists.

77
00:08:14,230 --> 00:08:18,850
And modelling is very much a multidisciplinary field as and as you can understand it.

78
00:08:18,850 --> 00:08:24,370
So we we often band together to perform these investment cases all around the world.

79
00:08:26,110 --> 00:08:31,200
So I think we're moving towards COVID now.

80
00:08:32,050 --> 00:08:40,270
So can you remember where you were when you first heard that there was something going on in China and how

81
00:08:40,270 --> 00:08:46,210
soon you became aware that this was something that your skills were going to be needed to to address?

82
00:08:47,050 --> 00:08:52,330
And it was almost immediately, oh, didn't take more than a week or two before the W.H.O. called me,

83
00:08:53,440 --> 00:08:58,180
and they were assembling a task force of modellers around the world.

84
00:08:58,990 --> 00:09:09,460
And I had previously met the head of the Global Emergency Response at the World Economic Forum in 2019, earlier that year in July.

85
00:09:09,700 --> 00:09:22,480
And we became well acquainted. And so a late in December of 2019, when COVID was still SARS-CoV-2 to a so-called source ncov or something like that.

86
00:09:22,900 --> 00:09:26,950
I should even remember the old name. It had it had an old name,

87
00:09:26,950 --> 00:09:33,249
and it was still very much a zoonotic transmission from from from the wet markets and so

88
00:09:33,250 --> 00:09:38,559
on that I was called up to be one of two Africans who were on the on that task force.

89
00:09:38,560 --> 00:09:40,840
And it was a group of modernism around the world,

90
00:09:40,840 --> 00:09:47,830
all this sort of learning coming together and just understanding what data was being collected on the ground, trying to make sense of it.

91
00:09:48,010 --> 00:09:54,510
And it was still very much a European and an Asian problem at the time.

92
00:09:54,520 --> 00:10:01,959
There was perhaps only one or two cases yet on the African continent did we imagine

93
00:10:01,960 --> 00:10:06,540
that we would be a global pandemic to the extent that we have actually seen?

94
00:10:06,550 --> 00:10:14,110
No, I think no at the time we do that, though, that was the time when you still thought it was by understanding the chain of transmission that,

95
00:10:14,110 --> 00:10:20,680
yes, it would spread, but it wouldn't be quite so catastrophic with so many variants and so many iterations.

96
00:10:21,610 --> 00:10:28,150
But that that was it was very soon that I was involved in Cobra back in December of 2019.

97
00:10:28,390 --> 00:10:36,040
I was in South Africa at the at the time and in the first case arrived in South Africa.

98
00:10:36,040 --> 00:10:40,299
And I say arrived because it was via international travel that the first

99
00:10:40,300 --> 00:10:45,580
detected case was a was was picked up in South Africa on the 5th of March 2020.

100
00:10:45,850 --> 00:10:52,720
At that point, I was also already alerted by my own government to be requested to develop models.

101
00:10:53,290 --> 00:11:01,569
So modellers were among the first people to be to be called because this was exactly the time for when her skillset was was needed.

102
00:11:01,570 --> 00:11:08,950
And my previous work in malaria actually with this African government was the reason that I was I was called on.

103
00:11:09,340 --> 00:11:14,440
I didn't have to approach and showcase my skills to government.

104
00:11:14,440 --> 00:11:19,930
Rather, it was a case of government calling me to say, Can you build us a COVID model?

105
00:11:20,680 --> 00:11:24,760
And and you know, what is it going to take?

106
00:11:25,180 --> 00:11:29,139
And so we started very early in in March with the first few cases.

107
00:11:29,140 --> 00:11:35,780
Still only international no local trials for local transmission was not that yet, but we were already involved.

108
00:11:36,280 --> 00:11:43,000
So what were you putting into your model If you had so few cases, what other factors were you able to build into your model?

109
00:11:43,720 --> 00:11:51,070
So the scientific evolution of the model building and COVID, it's it really did change over time and it changed quite rapidly.

110
00:11:51,220 --> 00:11:56,920
So stage one or stage zero, perhaps we can call it right at the very beginning, we have no local data,

111
00:11:57,130 --> 00:12:03,459
so the best approach of modelling in the setting then is to see what data is available around the world and

112
00:12:03,460 --> 00:12:10,690
then to call on local experts to translate that data as to what should be appropriate for your models.

113
00:12:10,810 --> 00:12:15,310
So very soon in South Africa we established in SAM in March of 2020,

114
00:12:15,310 --> 00:12:22,030
we established what we know as the South African COVID 19 Modelling Consortium, that is the CMC.

115
00:12:22,150 --> 00:12:25,820
It was a group of modellers in-country and local scientific experts,

116
00:12:25,820 --> 00:12:31,330
so local clinicians, public health specialists, virologists, intensivists and so on.

117
00:12:31,690 --> 00:12:34,270
And so we would amass data like for example,

118
00:12:34,390 --> 00:12:42,100
people with severe illness in the Netherlands were spending up to 23 days on average in hospital in an ICU bed.

119
00:12:42,580 --> 00:12:49,450
Now we could put that information into our models and that would take up all our ICU resources very quickly in a

120
00:12:49,450 --> 00:12:54,670
model to consume more resources very quickly for the simple reason that admissions all stay for long periods.

121
00:12:55,060 --> 00:13:02,500
But in speaking to intensivists who are working on the ground in hospitals around the country, they would say that it is a practice.

122
00:13:02,500 --> 00:13:03,190
In South Africa.

123
00:13:03,190 --> 00:13:13,240
We have a much higher threshold of entry into into an ICU setting, and nobody will spend more than eight days in ICU because nobody can afford it.

124
00:13:13,480 --> 00:13:15,160
Even medical aids will not afford it.

125
00:13:15,160 --> 00:13:22,570
So therefore, we can't use 23 days as a reasonable estimate for South Africa, though it is in practice overseas, we will use 7 to 8.

126
00:13:22,690 --> 00:13:26,719
And so we had this process of translation. Of the profiles.

127
00:13:26,720 --> 00:13:35,960
And at that time, there was some some evidence coming from the W.H.O. commission in China to say there's no such thing as asymptomatic cases of COVID.

128
00:13:36,230 --> 00:13:43,370
And then at the same time, there was conflicting evidence from a bunch of other sources to say that 80% of the Diamond Princess,

129
00:13:43,370 --> 00:13:50,569
you might recall that that ship, 80% of the Diamond Princess cases were asymptomatic.

130
00:13:50,570 --> 00:13:55,790
And also sitting basically with a piece of data that says how long is a piece of string?

131
00:13:56,720 --> 00:14:00,620
And you've got to nail it. That. And as a as a model of Lee,

132
00:14:01,490 --> 00:14:05,690
I made it very clear that that is not my responsibility to translate that I

133
00:14:05,690 --> 00:14:09,589
need to I'm not the expert clinician who will know what that best practice is.

134
00:14:09,590 --> 00:14:11,000
I will put it into the models.

135
00:14:11,300 --> 00:14:19,100
So that is where the value of having a multidisciplinary team, like a full consortium with many scientific partners, became so useful.

136
00:14:19,460 --> 00:14:29,210
So that was at the very beginning. Look at what data was emerging quite literally on a daily basis and and, and adapt your models.

137
00:14:29,360 --> 00:14:33,130
And so it became important, number one, to be an experienced modeller.

138
00:14:33,860 --> 00:14:38,150
You couldn't ask a novel, a novel model, a novice model, rather.

139
00:14:38,330 --> 00:14:44,600
You couldn't ask a novice modeller to develop a model for your country with your population demographics on the go.

140
00:14:44,810 --> 00:14:50,360
My colleagues at CMI de in in the United Kingdom had a model, you know,

141
00:14:50,360 --> 00:14:58,400
a respiratory flu model that was in development for 20 years, and they could just, you know, adapt to COVID almost immediately.

142
00:14:58,880 --> 00:15:05,930
In our case, in South Africa being an LMC, we don't have readymade models like that for the simple reason that we don't have

143
00:15:05,930 --> 00:15:10,339
data for the population for 20 years having gone through apartheid and so on.

144
00:15:10,340 --> 00:15:14,180
There isn't reliable data to enable such a model to be developed.

145
00:15:14,180 --> 00:15:21,979
And so I, from my malaria days had a provincial, a model that as a subnational level already quite well developed,

146
00:15:21,980 --> 00:15:31,070
that I could adapt very quickly within a within a few days to then be presenting and incorporating all of this new new information.

147
00:15:31,340 --> 00:15:38,389
And so it became a process of adapting your models every few days in order to bring take into account new papers that

148
00:15:38,390 --> 00:15:44,719
were being that were being generated and then adapting when those papers got retracted because they were errors,

149
00:15:44,720 --> 00:15:49,100
because new information had come out and checking was done.

150
00:15:49,100 --> 00:15:54,350
You know, you you learn quickly about how you that you had to rely on preprints, but you had to be aware,

151
00:15:55,190 --> 00:16:01,400
you know, that that this was that this was grey literature and had not been through a peer review process,

152
00:16:01,430 --> 00:16:10,220
but there was no time to wait for eight months for a peer review process to happen because, because the government needed answers immediately.

153
00:16:11,360 --> 00:16:17,600
And were the government making policy decisions on the basis of your model, even as it was evolving?

154
00:16:18,020 --> 00:16:26,719
Oh, absolutely. And that was that was one of the the the the key takeaways and one of the one of them one of the best parts, I think,

155
00:16:26,720 --> 00:16:33,380
of being part of this modelling consortium, that, number one, it was commissioned by government and coordinated by government.

156
00:16:33,740 --> 00:16:39,410
And so when government takes the first step, there is already that setting up that pathway to decision making,

157
00:16:39,410 --> 00:16:42,950
almost that pipeline of scientific evidence to decision making.

158
00:16:43,310 --> 00:16:46,340
And whilst a you may say, you know,

159
00:16:47,180 --> 00:16:53,920
some governments may be angry that your your your model output keeps changing and your projections and your output will keep changing.

160
00:16:53,930 --> 00:17:00,240
They themselves were aware of the changing situation and appreciated that we would update every single day.

161
00:17:00,260 --> 00:17:04,309
We would update our models every single week and that things would in fact change.

162
00:17:04,310 --> 00:17:07,010
I think they would have been more suspicious if it didn't.

163
00:17:07,280 --> 00:17:13,790
And then the fact that it that it did at the same time, that we would we would put forward our caveats and our disclaimers,

164
00:17:13,790 --> 00:17:17,930
because we also have to protect ourselves professionally as a scientist,

165
00:17:18,140 --> 00:17:24,230
that we would have to say that if we we understand it is necessary at the end to

166
00:17:24,260 --> 00:17:30,920
make a six month projection as to how the entire pandemic would would pan out.

167
00:17:31,460 --> 00:17:37,550
And though you had, you know, only two weeks or three weeks of data in country, we understand that is not enough.

168
00:17:37,550 --> 00:17:42,950
But a decision needs to be made because a tender, a procurement policy,

169
00:17:42,950 --> 00:17:47,929
a a procurement set and a request for proposal needs to go out to purchase ventilators,

170
00:17:47,930 --> 00:17:52,400
to purchase sufficient oxygen to commission hospital resource beds,

171
00:17:52,610 --> 00:17:57,440
to be to be made field hospitals to be commissioned if they need to be to be built.

172
00:17:57,590 --> 00:18:01,040
And these are all things that can be done on a on a two week basis.

173
00:18:01,160 --> 00:18:05,990
They actually need to be long term decisions. And so we have to make those projections.

174
00:18:06,110 --> 00:18:11,089
But put in all the disclaimers to say models are biased simplifications of reality.

175
00:18:11,090 --> 00:18:16,250
They are based on the data that are that are used to inform them and the data are changing.

176
00:18:16,850 --> 00:18:24,620
But we had but that's also where there's constant communication and in-person communication, or at least a virtual communication,

177
00:18:24,620 --> 00:18:30,260
not just submitting a report but having meetings to do the model translation became vitally important.

178
00:18:31,520 --> 00:18:40,010
And how did the the epidemic in within South Africa evolve, and was it different from what we saw here in the UK?

179
00:18:40,790 --> 00:18:49,159
It was it was different. I think the UK and South Africa were among the first two places to detect almost simultaneously variants.

180
00:18:49,160 --> 00:18:57,410
The first time we ever detected variants the UK had in December I think was 14 December or so, had announced the alpha variant.

181
00:18:57,410 --> 00:19:01,130
By the 26th of December we had announced the beta variant, so a different one.

182
00:19:01,550 --> 00:19:06,040
This is 2021, we're talking about the end of 2020 and 2020.

183
00:19:06,470 --> 00:19:12,220
So that's what I mean. Yes, sorry, end of 2020. And so it's been in pandemic state for about a year now.

184
00:19:12,560 --> 00:19:19,040
It's all it's all melded together. But yes, that was we that was the first, but different variants.

185
00:19:19,670 --> 00:19:23,660
And then we had multiple waves it with different timing to the to the United Kingdom.

186
00:19:23,840 --> 00:19:29,450
Our vaccines arrived quite a bit later than the United Kingdom as well.

187
00:19:29,990 --> 00:19:37,040
For the again, I don't think it is any fault of a of of policy and so on.

188
00:19:37,040 --> 00:19:43,099
Rather, we had received a shipment from the from the UK manufacturers of the AstraZeneca

189
00:19:43,100 --> 00:19:48,080
vaccine in December ready to vaccinate our healthcare workers December 2020.

190
00:19:48,410 --> 00:19:51,680
But the AstraZeneca vaccine at that very time,

191
00:19:52,070 --> 00:19:56,360
the results came out of the clinical trial to show that it was simply not effective

192
00:19:56,360 --> 00:20:00,049
against the beta variant whilst being effective against the alpha variant.

193
00:20:00,050 --> 00:20:07,790
So we had a brand new shipment of vaccine that was literally not effective at all against the variant in South Africa

194
00:20:07,790 --> 00:20:15,379
that had taken over and all infection the way the variants manifested because they were each time we got a new variant,

195
00:20:15,380 --> 00:20:16,070
it was more,

196
00:20:16,730 --> 00:20:28,250
it was more infectious, it transmitted with higher efficiency than the that the previous circulating variant or the wild type ancestral COVID,

197
00:20:28,250 --> 00:20:34,489
as we as we knew it in wave one, wave one, that meant that there was a very quick takeover of the variant.

198
00:20:34,490 --> 00:20:42,260
It dominated transmission very quickly. And so there was literally no point in rolling out any AstraZeneca vaccine.

199
00:20:42,260 --> 00:20:51,350
It did a lot of harm to to to to public to public belief in in government and harm towards vaccine hesitancy as well.

200
00:20:52,340 --> 00:20:59,299
But it was we had to wait for forever and another set of vaccines that would actually be effective.

201
00:20:59,300 --> 00:21:04,850
And that came a few months later. So vaccine rollout was different between the United Kingdom and South Africa.

202
00:21:05,060 --> 00:21:10,960
So the Africa was also because of our strength in genomic surveillance having.

203
00:21:11,030 --> 00:21:13,999
Among the best capacity in the world also for genomic surveillance.

204
00:21:14,000 --> 00:21:24,920
We were detecting variants a quite a lot, and so we detected the electron variant in December of 2020.

205
00:21:25,160 --> 00:21:28,400
A Is it about December?

206
00:21:28,670 --> 00:21:34,610
I think December of 2021 would have been a al-Muqrin being being detected,

207
00:21:34,700 --> 00:21:39,109
and that was it was a first detected in South Africa, like Beta was first detected in South Africa.

208
00:21:39,110 --> 00:21:42,439
And so we had in South Africa what was different to the UK,

209
00:21:42,440 --> 00:21:49,460
I would say primarily even apart from vaccination, is that our waves were large, the Delta variant,

210
00:21:49,770 --> 00:21:58,790
and in fact infected a very large number of the population that by the time Omicron had had passed through the country,

211
00:21:59,780 --> 00:22:07,429
up to 80 to 90% of the country had been infected by COVID 19.

212
00:22:07,430 --> 00:22:16,310
So would have had immunity or would it be zero positive for COVID 19 from infection rather than vaccination?

213
00:22:16,460 --> 00:22:22,910
If it took vaccination into account, you would have had a seropositivity of around 97% of the country.

214
00:22:22,910 --> 00:22:26,600
And that came from as on blood bank studies and a set of other studies.

215
00:22:26,600 --> 00:22:34,520
So we were among the first countries in the world with a with almost a fully immunised population from natural infection.

216
00:22:36,250 --> 00:22:41,150
And that. And how did that I mean, were your hospitals overwhelmed?

217
00:22:41,170 --> 00:22:49,180
What was the rate of morbidity and mortality as high, relatively speaking, as it as it had been in the UK?

218
00:22:49,840 --> 00:22:52,810
So two factors. What it would be would need to be taken into account.

219
00:22:52,840 --> 00:22:58,149
One is we have a quite a bit younger population than the United Kingdom and severity

220
00:22:58,150 --> 00:23:03,630
across all variants and ancestral COVID was always highest in the older population.

221
00:23:03,640 --> 00:23:15,220
So from that perspective, the the the, the we had fewer age of, you know, age related, a lower age related mortality.

222
00:23:15,230 --> 00:23:21,160
But but the health system and access to health care is also very different.

223
00:23:21,340 --> 00:23:28,899
So from that perspective, when one looks at excess deaths compared to hospital reported deaths or excess deaths,

224
00:23:28,900 --> 00:23:32,950
taking into account deaths that would have been attributable to COVID.

225
00:23:33,100 --> 00:23:38,120
So a portion of all excess deaths would have been attributable to COVID and, you know,

226
00:23:38,140 --> 00:23:44,440
taken place either at home and not yet recorded as being due to COVID from a reporting system.

227
00:23:45,250 --> 00:23:51,220
The number of reported the number of deaths is estimated to be three times the level of reported deaths.

228
00:23:51,430 --> 00:23:56,710
So so that's the had among the highest excess deaths estimated in the world.

229
00:23:57,640 --> 00:24:03,460
So we suffered considerably from COVID 19, and that happened primarily during the Delta where the Delta wave,

230
00:24:03,730 --> 00:24:12,760
the Delta variant was a more severe variant and therefore having higher mortality than both the beta variant and ancestral COVID.

231
00:24:13,180 --> 00:24:18,190
So by the end of the delta wave, when there was say, around about 70,

232
00:24:18,670 --> 00:24:25,930
70% of the population now being protected or zero positive due to natural infection.

233
00:24:26,080 --> 00:24:33,640
When Omicron hit South Africa, also as a large wave, we saw much reduced severity.

234
00:24:34,210 --> 00:24:39,940
And so our hospitals were not overwhelmed. In fact, they were hardly registering severe cases at all.

235
00:24:39,940 --> 00:24:42,790
Of course, severe cases did happen, and that is always unfortunate.

236
00:24:43,360 --> 00:24:54,010
But on the ground as a whole, at the population level did not manifest anywhere as severely as as the delta and and previous previous waves.

237
00:24:54,670 --> 00:25:02,830
And so we had where we experienced hospital resources being overwhelmed was during wave one and during the beta wave and of the delta wave,

238
00:25:03,220 --> 00:25:08,770
there were scattered reportings all manifesting differently at different geographic levels

239
00:25:08,770 --> 00:25:15,070
because some areas of some parts of the country were affected worse during a certain waves.

240
00:25:15,850 --> 00:25:24,100
It wasn't all at the same time. And were the models able to anticipate some of these changes as they came along?

241
00:25:24,100 --> 00:25:33,400
And what I mean, was it possible to put in place non-pharmaceutical interventions that could at least mitigate the the severity of these waves?

242
00:25:34,240 --> 00:25:37,780
So we the inverse the answer to that is actually yes.

243
00:25:37,990 --> 00:25:45,160
The models were able to anticipate quite a quite a bit of this in the in the sense that and this is where I'll speak a

244
00:25:45,160 --> 00:25:51,010
little bit on modelling methodology and how we changed our thinking between the waves because decision making changed.

245
00:25:51,550 --> 00:25:53,920
Our primary purpose was to support decision making.

246
00:25:54,100 --> 00:26:00,690
During the first wave of COVID, the decision was primarily on, you know, can we flatten that curve,

247
00:26:00,700 --> 00:26:05,650
Can we achieve zero policy of zero COVID, you know, and have a policy towards that?

248
00:26:05,770 --> 00:26:10,299
Will we have enough hospital resources and what will that prediction be?

249
00:26:10,300 --> 00:26:18,430
So we spent a lot of time making projections on on what the wave would look like for for for wave one.

250
00:26:18,790 --> 00:26:22,270
But when wave two happened in wave three and so on,

251
00:26:22,420 --> 00:26:28,450
we changed our modelling methodology to no longer be about trying to forecast or to predict what was going to happen,

252
00:26:28,630 --> 00:26:35,680
but rather to do a what if scenario type analysis to say we cannot predict what the next variant is going to be.

253
00:26:35,680 --> 00:26:40,870
We cannot also predict when it will emerge and what is characteristics are going to be.

254
00:26:41,080 --> 00:26:47,229
So in the face of all that uncertainty, what we can do is to to generate a number of what if scenarios,

255
00:26:47,230 --> 00:26:55,389
to say something like if the next variant arrives in the next few months and it happens to be a have a higher

256
00:26:55,390 --> 00:27:03,430
transmission than prior previous variants and it happens to to have some immune loss potential and or not,

257
00:27:03,430 --> 00:27:10,750
and look at various iterations of those scenarios to make a range of projections as to whether hospital resources would be overwhelmed or not.

258
00:27:10,960 --> 00:27:15,730
That was the key planning information. We were able to then provide, provide government.

259
00:27:16,390 --> 00:27:24,280
And so we did this in advance of the Delta where we did a we released model a report on a hypothetical variant with increased transmissibility.

260
00:27:24,280 --> 00:27:26,560
What would its impact be?

261
00:27:26,740 --> 00:27:35,260
In advance of the al-Mukhtar Way, we released another report on a hypothetical bird with some immune loss, as well as increased transmissibility.

262
00:27:35,260 --> 00:27:43,840
And so. In that way, we were able to actually among those scenario sets, we we actually pre-empted what was what ended up emerging.

263
00:27:44,410 --> 00:27:46,960
And that did help did help decision making.

264
00:27:47,410 --> 00:27:56,150
Coming to you to your point of the last part of your question on pharmaceutical intervention, non-pharmaceutical goods,

265
00:27:56,500 --> 00:28:05,500
non-pharmaceutical interventions, as we know now, public health and social measures to be to be implemented in unlike the UK,

266
00:28:05,710 --> 00:28:13,270
in South Africa, we did not have data that would have allowed us to say the proportion of the population using face masks

267
00:28:13,270 --> 00:28:19,780
is so much and the contribution of social distancing is so much and handwashing and hygiene practices.

268
00:28:19,930 --> 00:28:23,919
And so we were unable to disentangle any of the public health,

269
00:28:23,920 --> 00:28:29,440
the social measure in terms of the utilisation and adoption, but also in terms of the impact.

270
00:28:29,680 --> 00:28:40,330
And so we chose another different methodological approach to address that, to be looking at the relative, the relative compliance, if you will,

271
00:28:40,480 --> 00:28:51,210
or, or it was really a two part process of looking at the relative impact of what the government, the restriction level was at the time, right.

272
00:28:51,260 --> 00:28:55,899
To the lockdown restriction level to say you cannot, you know,

273
00:28:55,900 --> 00:29:02,560
face masks must be worn inside and at outdoor activities and restrictions on capacity in venues and so on,

274
00:29:02,740 --> 00:29:09,610
but then also take into account the population's adoption or adherence to these policies.

275
00:29:09,760 --> 00:29:14,229
So we looked at it all as a single measure. And when we made our scenario projections,

276
00:29:14,230 --> 00:29:19,450
we said we looked at a variety of scenarios as to whether the population would

277
00:29:19,570 --> 00:29:26,590
maintain the same level of adoption of these policies as per previous wave.

278
00:29:27,070 --> 00:29:32,620
And whether that would have been what if it were 20% less or 50% less and so on,

279
00:29:32,770 --> 00:29:39,429
because it gave a sense of relativity and it took into account what we with what we know as NPI fatigue,

280
00:29:39,430 --> 00:29:45,850
that the population were not being as strictly adherent to social distancing and mask wearing

281
00:29:45,850 --> 00:29:50,950
for two years straight because we were all humans and fatigue would set in quite naturally.

282
00:29:50,950 --> 00:29:53,650
And so we adopted that relative approach rather.

283
00:29:54,710 --> 00:30:02,270
Because that was also a lot more understandable to the population and it would have been what we would call spurious accuracy

284
00:30:02,480 --> 00:30:10,520
to try to say there's a 20% contribution from handwashing and a 10% contribution from from face mask wearing and so on.

285
00:30:12,060 --> 00:30:20,210
Mm hmm. And yes.

286
00:30:20,220 --> 00:30:31,820
And to what extent was the community in the public health community in South Africa united in its approach to dealing with the the outbreak?

287
00:30:32,870 --> 00:30:40,819
So there was quite a lot I think there's a quite quite a lot in terms of the sort of the country coming coming together,

288
00:30:40,820 --> 00:30:45,799
I think especially in wave one with regular meetings from the from the presidency,

289
00:30:45,800 --> 00:30:49,610
there was no belief that we could all, you know, work through this together.

290
00:30:49,760 --> 00:30:53,720
Over time. Of course, there will always be dissenting elements.

291
00:30:54,440 --> 00:30:59,330
And, you know, with vaccine hesitancy and, you know, inefficiencies of government,

292
00:30:59,960 --> 00:31:03,830
we had a few corruption scandals in between changes of ministers of health.

293
00:31:04,010 --> 00:31:13,250
And all of that does not speak well to to enable public, you know, public goodwill and so on.

294
00:31:13,580 --> 00:31:19,940
But the way we worked as as models is that we had a very well-established decision making pipeline

295
00:31:20,180 --> 00:31:25,069
in the sense that as even though I was part of the modelling consortium and my group Maisha,

296
00:31:25,070 --> 00:31:29,270
we were leading the model development for the models that I've just described to you.

297
00:31:29,310 --> 00:31:38,530
Now we also I'm, I also sit on what we call the Ministerial Advisory Committee that's akin to to SAGE in the UK.

298
00:31:38,960 --> 00:31:44,300
So this ministerial advisory committee has a direct line to the Minister of Health and we also,

299
00:31:44,300 --> 00:31:47,870
as the modelling consultant, had a direct line to the Minister of Health and the Presidency.

300
00:31:48,110 --> 00:31:52,640
And so we would, we would send through this information on the evolving pandemic.

301
00:31:52,850 --> 00:31:59,930
And, and, you know, and our recommendations as part of the Ministerial Advisory Committee as to what the restrictions should be.

302
00:32:00,230 --> 00:32:02,629
And so the when the presidency would communicate,

303
00:32:02,630 --> 00:32:11,390
there was always that scientific basis and that background that could be spoken to that that could be shared with the with the public.

304
00:32:11,570 --> 00:32:15,049
Was it ideal? No. Could it have been done better also?

305
00:32:15,050 --> 00:32:18,500
No. But were there at least attempts made? Yes.

306
00:32:18,710 --> 00:32:24,470
Was it better than our previous experience with the HIV pandemic ten, 15 years ago?

307
00:32:24,680 --> 00:32:30,259
Absolutely, yes. I think, you know, overall, it was a trying time for everyone,

308
00:32:30,260 --> 00:32:35,600
everyone in government and me being an outsider as a scientist from a from a

309
00:32:35,600 --> 00:32:40,940
public university working with public servants in government on a daily basis,

310
00:32:41,180 --> 00:32:45,319
I will be able to say that they are good people who are doing their best.

311
00:32:45,320 --> 00:32:50,080
And everybody was was was working really hard and, you know,

312
00:32:50,120 --> 00:33:00,169
trying to work through red tape and an impossible system in an impossible situation to to work towards the public, to work towards the public good.

313
00:33:00,170 --> 00:33:09,560
You will not win all battles. And the the the the the sad part of modelling and I think of some of the scientific evidence that was generated is that

314
00:33:09,770 --> 00:33:17,050
if you if you had a model that showed doom and gloom and it is used for advocacy to generate better public health,

315
00:33:17,060 --> 00:33:22,580
a social measure, adoption and policy creation and so on, the doom and gloom may not result.

316
00:33:22,580 --> 00:33:24,860
And then you criticise for having a wrong.

317
00:33:26,450 --> 00:33:36,680
But you know that's a of stuff that that that was almost the point achieved and and that happened multiple multiple times.

318
00:33:36,920 --> 00:33:43,550
It was not it was not always a rosy place to be as a modeller in the, in the public,

319
00:33:43,760 --> 00:33:47,510
you know, where it was publicly known that you were that you are modelling.

320
00:33:47,510 --> 00:33:54,040
We were involved in in court cases by sort of public action groups who are upset.

321
00:33:54,620 --> 00:33:59,120
You know, we were involved in you know, there were many articles in the newspaper,

322
00:33:59,120 --> 00:34:05,239
There were personal attacks on people, not myself personally, but some of my colleagues had articles written about them.

323
00:34:05,240 --> 00:34:11,899
They were often articles in the newspaper attacking us for doing a you know, by even other scientists who would say,

324
00:34:11,900 --> 00:34:18,130
you're doing all the wrong things without actually knowing what on earth you were you were actually doing in the in the first place.

325
00:34:18,140 --> 00:34:21,460
So it took a long a lot of effort to, one,

326
00:34:21,470 --> 00:34:28,490
be scouring the media all the time and then having to write your own reports that responded

327
00:34:28,490 --> 00:34:33,380
to these to these articles and responded to to participate in court cases and so on.

328
00:34:34,010 --> 00:34:40,219
But overall, I think that it was a it was a very positive it was a very positive experience globally for the world,

329
00:34:40,220 --> 00:34:44,780
of course, has been a negative experience, but positive When I say positive here,

330
00:34:44,780 --> 00:34:53,179
I mean from the from a scientific perspective of working with with government and establishing these decision making pipelines,

331
00:34:53,180 --> 00:34:57,080
that that modelling isn't just beneficial to close, it is beneficial to health.

332
00:34:57,380 --> 00:35:05,330
And it is one of these one of these toolkits that doesn't need to just be called upon in an emergency but can be used to support health.

333
00:35:05,330 --> 00:35:13,340
And so what COVID did for health management was to cement this relationship between all the the,

334
00:35:13,340 --> 00:35:18,860
the, the scientific expertise that lies at university around a country and.

335
00:35:19,110 --> 00:35:24,120
And within government, there are many positive externalities that have that have resulted from that,

336
00:35:24,300 --> 00:35:30,420
including a data Science for Health master's internship program that that we've set up where government

337
00:35:30,420 --> 00:35:35,010
data sets are being analysed by master's students and their dissertation for their dissertation projects.

338
00:35:35,710 --> 00:35:38,700
And these data sets wouldn't ordinarily be analysed.

339
00:35:38,700 --> 00:35:44,309
But here we've got to understand where we are and it benefits the students to do something meaningful.

340
00:35:44,310 --> 00:35:49,590
It benefits government because, you know, the data is being analysed and better decisions can be made.

341
00:35:49,590 --> 00:35:53,489
And so there's been a lot of good that has has, has come out of it.

342
00:35:53,490 --> 00:36:02,430
And I think as we as we look back even over, you know, ten, 20, 30 years from now, we will see benefits coming,

343
00:36:02,430 --> 00:36:07,170
not just arising, not just during the pandemic, but actually in terms of years to come.

344
00:36:07,410 --> 00:36:15,270
We are already seeing the formation of pandemic centres around the world in order to support pandemic preparedness,

345
00:36:15,840 --> 00:36:23,610
in order to support the routine health challenges of interfacing with respect to TB and HIV and just general health systems.

346
00:36:24,510 --> 00:36:35,490
I myself am sitting on a on an advisory group for vaccine immunisation with the W.H.O. over Implementation Vaccine Research Advisory Committee,

347
00:36:35,730 --> 00:36:44,160
and we are already overseeing global global work that has that has emerged and accelerated because of the of the pandemic,

348
00:36:44,160 --> 00:36:48,660
but to support other health functions that are not not COVID related.

349
00:36:48,870 --> 00:36:54,240
And I think the wonders that have emerged from vaccine manufacturing at a you know,

350
00:36:54,630 --> 00:37:00,900
at an accelerated pace compared to what we were what we were used to, we will see those benefits for many years to come.

351
00:37:02,040 --> 00:37:03,869
And how's the work that you've done?

352
00:37:03,870 --> 00:37:09,870
Have you been able to share the work that you've done with other low and middle income countries, or is it very specific to South Africa?

353
00:37:10,410 --> 00:37:15,840
And so during the pandemic, I did actually support other other countries as well.

354
00:37:15,840 --> 00:37:19,379
Through the Como Consortium based at Oxford University.

355
00:37:19,380 --> 00:37:24,810
I supported Mozambique during wave one through my own unit.

356
00:37:24,960 --> 00:37:29,160
M.A. supported neighbouring Namibia as well.

357
00:37:29,830 --> 00:37:38,129
And so the the lessons learned have all have been shared, not perhaps in the traditional way of publication those are under underway,

358
00:37:38,130 --> 00:37:45,600
but in more non-traditional settings of networking between between groups of shared seminal shared platforms.

359
00:37:45,600 --> 00:37:51,629
Funders have contributed a lot to the sharing of our of our of our work in South Africa,

360
00:37:51,630 --> 00:37:57,090
which has been seen as quite exemplary globally for the impact that we've that we've had.

361
00:37:57,420 --> 00:38:02,460
And so we've share that with many, many NYC settings around around the world.

362
00:38:03,550 --> 00:38:11,430
Oh, very good. So you have you now gone back to has your focus shifted now back to malaria again?

363
00:38:12,210 --> 00:38:20,760
And my focus, Andrew. And I think it's because of COVID 19 as well, has expanded into a range of range of diseases we have.

364
00:38:20,820 --> 00:38:24,480
And I'm very happy to be back in malaria as well.

365
00:38:24,490 --> 00:38:30,780
But I mean, I work in a number of diseases at the moment, quite a few vaccine preventable diseases.

366
00:38:30,780 --> 00:38:39,629
So we just completed a project in Hepatitis A working on as a project, a global project on alum disease for VAX,

367
00:38:39,630 --> 00:38:47,520
for DTP booster vaccination, DTP being part of the early immunisation programme or schedule.

368
00:38:48,180 --> 00:38:51,570
Diphtheria, tetanus, pertussis. Is that right?

369
00:38:52,120 --> 00:38:55,450
H.A. That's anaesthetists, Yes, Yes.

370
00:38:55,470 --> 00:39:03,090
Whooping cough. So that a booster vaccination and and of course throughout too much of the capacity

371
00:39:03,090 --> 00:39:09,419
building work that I do in in terms of trying to develop sustainable capacity in Olympics,

372
00:39:09,420 --> 00:39:16,380
which is one of my passions is training modellers from their own countries to be the experts in their in their own countries.

373
00:39:16,530 --> 00:39:21,030
And we do a lot of a lot of application in in a range of of countries.

374
00:39:22,110 --> 00:39:27,030
And and in terms of developing modelling capacity, but then also across a range of diseases.

375
00:39:27,030 --> 00:39:30,900
So we saw doing some COVID work, we, you know, completing a bronze.

376
00:39:31,470 --> 00:39:36,510
There is still some analysis to be done in terms of future rollout of vaccines and the need for

377
00:39:36,510 --> 00:39:42,569
booster doses in a in a setting with a high SEROPREVALENCE How long might that immunity last,

378
00:39:42,570 --> 00:39:49,320
if not challenged? What will this new family of own occurrence of variants lead to in the future

379
00:39:49,330 --> 00:39:53,340
Certainly does not seem to be the case that it's going to be severe infection.

380
00:39:53,550 --> 00:39:58,740
But what about long-covid? What about future waves?

381
00:39:58,890 --> 00:40:03,570
Might immunity wane to a point where even protection against severe infection might win?

382
00:40:03,750 --> 00:40:06,410
These are still questions that I'll be asked at the moment.

383
00:40:06,420 --> 00:40:15,690
The the the impetus, though, is much, much reduced and has allowed us to shift focus into into other diseases.

384
00:40:15,720 --> 00:40:20,910
So right now, it's a very healthy place for for modelling, working across a range of diseases,

385
00:40:21,030 --> 00:40:27,360
supporting a number of decision makers across across the world at the at the moment as well.

386
00:40:28,190 --> 00:40:32,250
And you mentioned that you'd been in Oxford just recently doing some teaching.

387
00:40:32,640 --> 00:40:34,050
What course was that?

388
00:40:34,860 --> 00:40:42,750
So I support the modelling for Global Health Master's program in the Nuffield Department of Medicine based in the Big Data Institute,

389
00:40:42,930 --> 00:40:49,140
and I am the module lead on a course called Malaria Modelling for Strategy Design.

390
00:40:49,380 --> 00:40:50,540
So the purpose of the course,

391
00:40:50,560 --> 00:40:58,110
the purpose of the degree is to teach mathematical modelling and the application of such my modules takes the theoretical and

392
00:40:58,110 --> 00:41:06,450
foundational underpinnings of the students would have learnt until that point and provides an application to a particular disease area,

393
00:41:06,450 --> 00:41:15,719
but with one focus only in that we modelling for policy. Now to separate why I say it's a separate focus on modelling in general is that you can have

394
00:41:15,720 --> 00:41:20,760
scientific modelling where you use your modelling to answer scientific biological questions,

395
00:41:20,970 --> 00:41:25,980
or you could be using our modelling to be supporting policy decisions.

396
00:41:25,980 --> 00:41:31,410
And when the type of modelling is a bit different, not just methodologically, but in how we approach the modelling.

397
00:41:31,680 --> 00:41:37,379
When you're modelling the policy, it involves a lot of engagement with policymakers, with reviewing of national documents,

398
00:41:37,380 --> 00:41:44,100
trying to understand what has happened on the ground compared to what is on a piece of paper and and so on.

399
00:41:44,280 --> 00:41:51,000
And so that is the course that I teach, teach the students and how to engage with, with policy documentation,

400
00:41:51,270 --> 00:41:55,169
what it takes to build a malaria model, not just from the biological perspective,

401
00:41:55,170 --> 00:42:01,530
but from the entire systemic or a systems thinking perspective to take into account population behaviour,

402
00:42:01,530 --> 00:42:07,890
geographic setting, environmental consideration and policy setting and health system access.

403
00:42:08,490 --> 00:42:12,910
And that includes economic factors as well as economic factors as well.

404
00:42:13,120 --> 00:42:20,399
Yes. So this is a two week module that was taught for the second year running at at Oxford.

405
00:42:20,400 --> 00:42:26,430
And we had a, we had a wonderful cohort of students from them from around the world who deeply appreciated the course.

406
00:42:27,150 --> 00:42:34,170
MM Wonderful. I'm going to switch now a little bit to focus more on your personal experience of the course of the pandemic.

407
00:42:34,440 --> 00:42:39,750
So first of all, how threatened did you feel personally by the risk of infection?

408
00:42:40,820 --> 00:42:47,600
Um, personally, I am. So I am actually a high risk individual, as in that I'm a diabetic.

409
00:42:48,200 --> 00:42:55,040
And so right up front it was known that, you know, that we would have a weakened immune systems and would be at higher risk.

410
00:42:55,340 --> 00:43:04,040
Um, I guess it was a good thing that I was quite so busy that I actually never left the house for a very long time because we were working from home,

411
00:43:04,040 --> 00:43:07,279
even though it was not a strict lockdown for more than five weeks.

412
00:43:07,280 --> 00:43:13,160
So we, you know, one was able to, to go back into work and and so on and and move around.

413
00:43:13,310 --> 00:43:21,110
I actually I, I protected myself by being at home for for simply being a high risk, high risk individual.

414
00:43:21,290 --> 00:43:23,329
Did I feel threatened necessarily?

415
00:43:23,330 --> 00:43:29,360
No, I would not say Well, not say that I knew, For example, what would allay my fears is that I knew that masks work.

416
00:43:29,630 --> 00:43:33,410
I knew the social distancing works. You know, I had to being being a modeller.

417
00:43:33,410 --> 00:43:43,100
I had no I at no point did I have any disbelief towards public health and social measures in place, nor the that the policies nor the nor vaccination.

418
00:43:43,670 --> 00:43:49,340
And so I was among among the first to be vaccinated being a high risk individual as well.

419
00:43:49,520 --> 00:43:55,219
So in terms of fear and and being threatened, no, I won't say I felt that,

420
00:43:55,220 --> 00:44:00,470
but I was very much aware and and took the to the desired and took the appropriate the

421
00:44:00,650 --> 00:44:06,559
risk appropriate behaviour for my for my personal health condition and what my friends,

422
00:44:06,560 --> 00:44:10,430
colleagues, family members. Were you anxious for their wellbeing?

423
00:44:11,060 --> 00:44:18,530
I yes, I was. I mean I had a it's a I was indeed anxious for, for, for their wellbeing.

424
00:44:18,530 --> 00:44:25,940
Not everyone is a scientist, so not everybody understands the, the benefits of public health and social and social measures.

425
00:44:26,090 --> 00:44:31,550
All of the you know, many would misinterpret what was being communicated for a simple reason.

426
00:44:31,610 --> 00:44:35,090
Not everybody was equipped to be a health, you know, a health specialist.

427
00:44:35,630 --> 00:44:44,390
And so often I was a point of verification to fight all the fake news that was abounding among amongst all the WhatsApp groups.

428
00:44:44,960 --> 00:44:50,630
And I was the. I was I was at point of contact.

429
00:44:50,840 --> 00:44:54,980
I was the point of verification.

430
00:44:55,190 --> 00:45:05,959
And it did become quite, I guess you could say, scary or rather saddening at a point where during the beta epidemic, as I mentioned earlier,

431
00:45:05,960 --> 00:45:14,450
various parts of the country were affected worst during some waves simply because they had not been affected badly in a previous wave,

432
00:45:14,450 --> 00:45:19,669
meaning you had a greater proportion of of the population being susceptible in that space.

433
00:45:19,670 --> 00:45:32,300
So during the beta wave, the on the east coast of the country, in the Guinea or the KwaZulu-Natal province of the country, which is where I am from,

434
00:45:32,690 --> 00:45:39,200
that province is particularly badly hit at the height of December summer by the beta,

435
00:45:39,380 --> 00:45:43,460
by the beta variant, which was transmitting despite everybody being outdoors.

436
00:45:44,350 --> 00:45:46,520
And so the having not having much of a seasonal effect.

437
00:45:46,520 --> 00:45:55,460
And during that time it was a it was a period of where every every day or every one and a half days, somebody I knew passed away.

438
00:45:56,770 --> 00:46:06,020
And being Indian by, by heritage, our familial relationships extend quite, quite a lot further than one's immediate family.

439
00:46:06,020 --> 00:46:10,580
So somebody could be related to you being twice, thrice removed, one might say,

440
00:46:10,640 --> 00:46:17,000
or along with seven or eight familial linkages and still be well known to you.

441
00:46:17,300 --> 00:46:24,320
So by virtue of having a larger community and knowing therefore more people in a lineage,

442
00:46:24,350 --> 00:46:30,890
was it happened that every day or every one and a half days somebody I knew passed away and that was a devastating time.

443
00:46:31,190 --> 00:46:34,759
It was a very bad in your psyche.

444
00:46:34,760 --> 00:46:43,130
At the same time, we were protected in the sense of being so busy that because the whole country's in a panic and death is abounding.

445
00:46:43,340 --> 00:46:49,100
But you know what? Your responsibilities as a modeller, you've got to get at that point of the beach where we were.

446
00:46:49,310 --> 00:46:56,120
My my group, Maisha, we set up the dashboard for the government that would allow the government

447
00:46:56,120 --> 00:47:01,219
planners and the public for the first time to see cases that were the cases,

448
00:47:01,220 --> 00:47:04,270
deaths and hospitalisations around the country.

449
00:47:04,280 --> 00:47:10,640
And up until that point of either been ten months into COVID and everything was at a national and a provincial level,

450
00:47:10,760 --> 00:47:14,960
not at a city sort of level or suburb level.

451
00:47:15,080 --> 00:47:20,299
And the government had no capacity to create such a dashboard. The data were there, but they had no capacity.

452
00:47:20,300 --> 00:47:26,930
And so I approached them to say, Would you like us to do it? And in two weeks we created this national the dashboard for the country.

453
00:47:27,920 --> 00:47:31,729
And so we knew the work and the analysis that had to be done.

454
00:47:31,730 --> 00:47:34,459
We were updating that dashboard three times a week,

455
00:47:34,460 --> 00:47:41,210
so all of this pressure was on us to perform this work so the rest of the country would have knowledge

456
00:47:41,210 --> 00:47:46,940
and be able to plan the day they did their travel and they did their day to day activity better.

457
00:47:47,510 --> 00:47:53,780
And that helps to protect us from some of the grief we might have experienced had we been

458
00:47:53,780 --> 00:47:59,390
sitting at home and waiting for information and not being as involved as we as we were.

459
00:47:59,810 --> 00:48:04,420
And that and I will say that that was a that was actually quite protective for us.

460
00:48:04,430 --> 00:48:10,819
It enabled us to to experience some of that grief just a couple of months after it happened.

461
00:48:10,820 --> 00:48:18,799
I did lose very close family members immediately, sort of shortly after that, within the space of two weeks,

462
00:48:18,800 --> 00:48:27,380
I lost my grandmother and and my cousin, my cousin from a very young cousin at the age of 40 from a COVID related post incident.

463
00:48:27,800 --> 00:48:32,629
My grandmother, actually at the age of 91, had COVID 19, was asymptomatic.

464
00:48:32,630 --> 00:48:36,950
She beat it, and then ten months or so later passed away of old age.

465
00:48:37,760 --> 00:48:44,960
So that was not necessarily a sad, sad generally, but not a COVID related event.

466
00:48:45,740 --> 00:48:52,070
So that was I think of it from a personal perspective that I think was the most trying time of the of the pandemic.

467
00:48:52,940 --> 00:48:58,310
So you said you were you were so busy. I mean, you were working. I mean, you sound to me like somebody who works all the time anyway.

468
00:48:58,490 --> 00:49:03,440
But were you working longer hours than you normally do? Um, yes.

469
00:49:04,100 --> 00:49:10,129
So, so granted, the work that we do is, you know, does mean that we, we do, we, we, we work, you know, all the, all the time.

470
00:49:10,130 --> 00:49:19,100
But I think that that phrase of working all the time, it's it was different pre-pandemic during the pandemic I was working what you would

471
00:49:19,100 --> 00:49:23,660
colloquially call an all nighter and quite literally working throughout the night,

472
00:49:24,440 --> 00:49:27,919
at least twice and twice a week, especially during wave one,

473
00:49:27,920 --> 00:49:33,860
when the Minister of Health or the people directly under the Minister of Health are texting you every morning.

474
00:49:33,860 --> 00:49:40,849
We are the role models. We need it now. Testing you personally not to the coordinating mechanism of your colleagues who are for this consortium,

475
00:49:40,850 --> 00:49:44,299
but you personally having to generate, you know,

476
00:49:44,300 --> 00:49:46,330
the does this work knowing the weight of the country,

477
00:49:46,580 --> 00:49:53,750
of the weight of decision making that's now being placed on almost on your on your head and then the pressure from the media that was on us.

478
00:49:54,560 --> 00:49:59,120
I was working throughout the night several times a week for a period of two years.

479
00:49:59,930 --> 00:50:04,190
It was it was extremely challenging from that perspective.

480
00:50:04,400 --> 00:50:07,040
It was challenging from a psychological perspective.

481
00:50:07,040 --> 00:50:14,929
So in wave one, for the for the first couple of weeks, my my team were assisting me in developing the models.

482
00:50:14,930 --> 00:50:18,950
I have a very young team, so they were assisting me in procuring information.

483
00:50:18,950 --> 00:50:22,070
So reading through papers and having and getting the understanding,

484
00:50:22,250 --> 00:50:27,739
working through the literature in developing on the computing side, developing pipelines and outputs and so on.

485
00:50:27,740 --> 00:50:34,490
But the actual motor development I was doing by my by myself also, it wasn't the kind of weight and the present.

486
00:50:34,520 --> 00:50:38,840
The presentation of such is not the kind of weight I could place on very young individuals.

487
00:50:38,990 --> 00:50:44,420
When you're up against, you know, stakeholders, the likes of the presidency and and so on.

488
00:50:44,420 --> 00:50:54,469
So the idea, however, a few weeks into the pandemic, you could start to see the psychological impact that it was having on my team when we had to.

489
00:50:54,470 --> 00:50:59,150
We had to support some some functions. I guess we never thought as much as we would be doing.

490
00:50:59,150 --> 00:51:04,709
For example, helping people in government plan burial sites during the during the waves

491
00:51:04,710 --> 00:51:09,500
of is supporting the procurement of mortuary containers at hospitals begins.

492
00:51:09,500 --> 00:51:14,420
Representations of people were showing you pictures of bodies in mortuaries that were not

493
00:51:14,420 --> 00:51:20,180
overflowing or just being stacked at the back because there weren't enough enough space and so on.

494
00:51:20,840 --> 00:51:25,970
And it was one morning where that just became too much for my for my team.

495
00:51:26,120 --> 00:51:32,270
And I had a received WhatsApp text saying, I want to be a warrior, I want to be like you, but I can't wake up.

496
00:51:32,540 --> 00:51:39,530
I can't get up. And at that point I said, All right, guys, two weeks you'll all of just just take take a time off.

497
00:51:39,740 --> 00:51:46,130
You will relax. I could not afford myself that that break because I still needed to to pick up the helm.

498
00:51:46,400 --> 00:51:50,180
I then organised and with the I'm very grateful to my university for supporting me

499
00:51:50,180 --> 00:51:55,309
in this that they organised group therapy and individual therapy sessions for my,

500
00:51:55,310 --> 00:51:58,459
for my team to take us through that, to take us through that time.

501
00:51:58,460 --> 00:52:08,210
It was absolutely beneficial. I saw pushed, pushed through and at that from that point on most though I also I held back with my with my team.

502
00:52:08,390 --> 00:52:17,780
I reduced the their involvement in terms of their sort of the day to day horror show that was covered at the at the time.

503
00:52:17,780 --> 00:52:25,310
I took the role myself and just helped them to just adjust psychologically from from from that.

504
00:52:26,000 --> 00:52:31,120
Mm hmm. Did you did you take advantage of any therapy yourself?

505
00:52:31,750 --> 00:52:36,640
I did. When the university provided that. Provided that of service, I did.

506
00:52:36,880 --> 00:52:42,610
I wanted to take advantage of it for longer and to adjust to other times.

507
00:52:42,970 --> 00:52:49,270
I could not I could not find the find the time. But I had my own personal coping coping mechanisms as well.

508
00:52:49,280 --> 00:52:55,810
I think the the fact that we were doing this for a purpose is itself very much a coping mechanism,

509
00:52:56,050 --> 00:52:59,440
that it's for a greater it's something altruistic, it's for a greater good.

510
00:52:59,440 --> 00:53:02,650
We were not ever paid to do any of this covert work.

511
00:53:02,980 --> 00:53:10,510
In fact, when the Chief Directorate of Health of the National Treasury approached me first in early March 2020 to ask me,

512
00:53:10,510 --> 00:53:15,190
will you develop a COVID model? Well, his second question to me was, well, you know, how much will you charge?

513
00:53:15,190 --> 00:53:18,849
What can we pay you? And I said, Oh, no, don't worry, this is a national service.

514
00:53:18,850 --> 00:53:25,060
I'll do this for the for the country. You won't have to pay me taxpayers money to do something like this as a service of the country.

515
00:53:25,210 --> 00:53:33,100
Also not thinking it would be to. Yes, as well. But that being said, no, I would never have taken money.

516
00:53:33,100 --> 00:53:39,579
It was also important for us to be to not be funded by government when you're providing evidence and support to government

517
00:53:39,580 --> 00:53:47,290
because it adds to the objectivity of the scientific process and the and the function that you are that you are serving.

518
00:53:48,370 --> 00:53:51,759
So I wasn't able to take advantage of that coping mechanisms.

519
00:53:51,760 --> 00:53:57,190
I wish I was able to do more exercise. I did not. But I had my my yoga.

520
00:53:57,190 --> 00:54:03,070
I did a lot of a Hindu scientific breathing exercises called Pranayama.

521
00:54:04,000 --> 00:54:07,990
And honestly, I think that's what kept me kept me alive. I'm a devout Hindu.

522
00:54:08,380 --> 00:54:15,460
And so having the the support structure of my faith is really what what kept me going.

523
00:54:15,940 --> 00:54:26,260
I was always I was always cognisant of my health knowing that I had to at the end of the malaria investment case in 2018 that we did in South Africa,

524
00:54:26,530 --> 00:54:30,610
I needed to. This is amusing. I guess now maybe not so much of the time.

525
00:54:30,820 --> 00:54:37,810
I needed to also spend an all nighters in generating it to generate the evidence for by the required time frames.

526
00:54:38,080 --> 00:54:44,170
But my diabetic medication was putting me to sleep at night and preventing me from staying up all night.

527
00:54:44,170 --> 00:54:48,070
So I was keeping my medication to stay up and to perform the work.

528
00:54:48,340 --> 00:54:52,030
And skipping your medication leads to your immune system being compromised.

529
00:54:52,210 --> 00:54:58,240
And I got really sick and I ended up in hospital at the end of 2018 and I learned my lesson.

530
00:54:58,450 --> 00:55:05,500
So all throughout the COVID 19 pandemic, I took my medication judiciously and I still managed to do the all nighters.

531
00:55:05,620 --> 00:55:14,350
And so that's what protected me. In fact, even when my husband at home was asthmatic, when he got COVID 19 during the Delta wave,

532
00:55:15,130 --> 00:55:19,270
I tested it with PCR is highly sensitive test twice and was not.

533
00:55:19,780 --> 00:55:23,080
I had only one Pfizer vaccine at the time and I did not contracted.

534
00:55:23,560 --> 00:55:27,610
So my medication paired with one vaccine was a was working for me.

535
00:55:27,610 --> 00:55:34,509
So I was I was very glad for that. I knew I knew that the sacrifice that was being made that helped me cope,

536
00:55:34,510 --> 00:55:41,650
but it also helped me prior experience with my medication, you know, stunning of my health, helped me to cope responsibly.

537
00:55:43,270 --> 00:55:45,160
That's very good. Very good story.

538
00:55:45,970 --> 00:55:55,980
So I think we've worked through everything unless unless there are any particular anecdotes or stories that you remember from that time that you know,

539
00:55:56,080 --> 00:55:59,930
you'll tell your grandchildren about. Oh, gosh.

540
00:56:00,000 --> 00:56:03,960
I think this is so much, too. There is so much to to reflect.

541
00:56:04,260 --> 00:56:08,730
To reflect on. And one of I think I'll go through a series of reflections.

542
00:56:08,980 --> 00:56:16,650
So one is it's an it was an honour that I will I will always say it was an honour to be able to serve my country and the countries I did.

543
00:56:17,160 --> 00:56:25,470
The other countries that I supported in this way to be able to to make a difference, to be able to be part of this scientific communication.

544
00:56:25,650 --> 00:56:28,680
The goal of modelling is not to have your decision.

545
00:56:28,920 --> 00:56:32,970
A Listen to what your models say and in that though they did it multiple times.

546
00:56:33,270 --> 00:56:40,739
I think the band, the DVD goal of modelling is to have your evidence heard before decision makers and for them to weigh

547
00:56:40,740 --> 00:56:45,960
up your evidence along with all the other sources of evidence that your model is not accounting for.

548
00:56:46,050 --> 00:56:48,090
And then to make an overall decision.

549
00:56:48,270 --> 00:56:55,410
So the fact that we were able to establish that pipeline and be part of that decision making process several times over.

550
00:56:56,160 --> 00:57:01,350
I think is a has has been one of the biggest highlights of my a lot of my career.

551
00:57:01,350 --> 00:57:06,120
And what I think remain remains a hopefully in my lifetime and my working life.

552
00:57:06,120 --> 00:57:11,430
And there isn't another pandemic that I have deal with that would surpass the impact that we've had in this one.

553
00:57:11,670 --> 00:57:15,540
So I will say that that will always be a highlight of my career.

554
00:57:16,140 --> 00:57:26,879
It was additionally a highlight to present to a different set of stakeholders, and one is used to usually when one is doing policy facing modelling,

555
00:57:26,880 --> 00:57:35,610
there is a group of decision makers who are part of government or part of a nai tag, an immunisation advisory group or so on.

556
00:57:36,060 --> 00:57:39,810
In this case you are presenting to a presidency a cabinet of ministers.

557
00:57:39,990 --> 00:57:46,020
I did one very interesting presentation before a panel of High court judges or chief justices,

558
00:57:46,230 --> 00:57:50,700
which was which was an incredible experience because I've never been I've never had

559
00:57:50,700 --> 00:57:55,170
to present in a court setting like that where you're interrogated by a chief justice.

560
00:57:55,380 --> 00:58:03,780
This was with respect to the question of whether we should be hosting elections or not and the ended and

561
00:58:03,900 --> 00:58:11,070
know the impact of rallying and having political campaigns prior to that and the effect of of transmission,

562
00:58:11,070 --> 00:58:14,220
what it would do to transmission and so on.

563
00:58:14,700 --> 00:58:21,720
It was a wonderful experience because I got to be interrogated by a non-scientist in a most thorough way.

564
00:58:21,930 --> 00:58:27,900
And I and it was a wonderful learning experience for me, but also a very enjoyable, enjoyable one.

565
00:58:28,980 --> 00:58:35,700
I think of it on a personal note, a person who come to the personal notes a little bit.

566
00:58:35,820 --> 00:58:42,500
One more scientific note that I think will that will be one of the key lessons that emerged from from COVID

567
00:58:42,750 --> 00:58:48,930
and from this pandemic experience is as a model that one needs to know when is the right time to model?

568
00:58:49,710 --> 00:58:55,770
During the beta, when the beta variant first emerged in December of 2020, the.

569
00:58:57,450 --> 00:59:02,219
We didn't know what it was going to how it was going to manifest, what its characteristics were going to be.

570
00:59:02,220 --> 00:59:09,330
We just knew that it was there. And so government immediately asked us, you know, what will the size of the end of this next wave be?

571
00:59:09,330 --> 00:59:10,440
What can we expect?

572
00:59:10,680 --> 00:59:19,350
And our questions were, we do not think it is responsible to model because we do not know what those features of the variant are going to be.

573
00:59:19,360 --> 00:59:29,370
The modelling that that I do is mechanistic. So it takes cause it is not statistical modelling where one can look at associations and trends in data.

574
00:59:29,370 --> 00:59:37,980
Rather it is mechanistic and causal that a variant that is more transmissible will lead to more to more infections and so on.

575
00:59:38,070 --> 00:59:42,300
And so if you do not know if it is in fact more transmissible and you put in an incorrect assumption,

576
00:59:42,480 --> 00:59:45,390
then you are going to be your model will produce nonsense.

577
00:59:45,840 --> 00:59:52,560
And so we made a decision to change methodology instead of actually using standard mechanistic models to develop a different kind,

578
00:59:52,770 --> 00:59:56,790
set a model that would be more statistical and not reliant on cause.

579
00:59:57,330 --> 01:00:03,690
And that proved very much to our benefit because even prior to the to the detection of the beta variant,

580
01:00:03,840 --> 01:00:08,160
we had to be making developing these, these models and so on.

581
01:00:08,400 --> 01:00:18,720
So knowing when it's a very brave decision to be saying this is the chief tool in my tool kit and I'm electing not to use it.

582
01:00:19,080 --> 01:00:22,390
It's, I think it was Maslow who had said, you know,

583
01:00:22,410 --> 01:00:30,250
that it is it is tempting to that if you have a hammer to treat everything as if it were a nail and without,

584
01:00:30,480 --> 01:00:37,320
we've got to be careful as modellers to not to not employ the same methodologies all the time, regardless of the situation.

585
01:00:37,410 --> 01:00:44,309
It's a lesson that I think will be relevant even in 20 or 30 years time, that every country is different, every situation is different.

586
01:00:44,310 --> 01:00:53,090
And so you have to evaluate each situation individually to know what is the appropriate scientific method to to apply them.

587
01:00:53,130 --> 01:00:57,750
So that definitely, I think is something to to reflect, to reflect on.

588
01:00:58,560 --> 01:01:09,750
I think what's another another highlight is the importance of communication that that modelling and I think it speaks to science in general,

589
01:01:09,750 --> 01:01:12,330
but I'll speak the modelling aspect of the modelling as a package.

590
01:01:12,840 --> 01:01:17,159
Sometimes it's least about the model itself and developing the model, it's about the interaction,

591
01:01:17,160 --> 01:01:22,200
it's about the how you package your, your, your, your, your, your model results.

592
01:01:22,200 --> 01:01:26,130
Gone are the days where someone in government has a time to read a 50 page report,

593
01:01:26,460 --> 01:01:31,350
and if you're generating a presentation, it needs to have your conclusions on each slide.

594
01:01:31,530 --> 01:01:38,099
It can't be a standalone picture where the interpretation is left up to the reader and you know that

595
01:01:38,100 --> 01:01:43,830
we've got to we have to adapt and adjust very quickly with the different and new modes of communication.

596
01:01:44,010 --> 01:01:49,799
We've got to be the whole package. If you want to be a model during this time in this pandemic,

597
01:01:49,800 --> 01:02:00,300
I've done any number of radio interviews and television interviews and the thing that benefited me the most there in being confident to run to do

598
01:02:00,300 --> 01:02:09,330
these press releases and press conferences was in fact my school experience of being a national debater as a national school debating champion.

599
01:02:09,540 --> 01:02:14,220
It served me really well in my experience as a lecturer at university meant that, you know,

600
01:02:14,310 --> 01:02:21,840
being able to to communicate clearly and explain a quite high level science in a in a way

601
01:02:21,840 --> 01:02:28,140
that is digestible to the public and to and to decision makers as well as non-experts,

602
01:02:28,320 --> 01:02:35,490
I think was a was a skill that needs to be developed and inculcated in in the new generation of scientists.

603
01:02:35,490 --> 01:02:38,310
And what is the one that can't cannot be understated.

604
01:02:39,000 --> 01:02:46,020
And so that I think is is definitely something that will still be of relevance in in the next few few decades.

605
01:02:46,770 --> 01:02:54,419
On a personal note, I think looking back on the we will be highly amused to be recollecting what

606
01:02:54,420 --> 01:03:00,540
a wonderful time the pandemic was for all pets at home and my brood at home.

607
01:03:00,540 --> 01:03:07,680
I don't have human children, but I do have seven cats and four dogs, two ducks and three chickens.

608
01:03:08,310 --> 01:03:16,860
And then I brood at home. My cats and my dogs could not have been happier to have had me at home for for two years.

609
01:03:16,860 --> 01:03:21,519
Ordinarily, I'm travelling for at least three months a year in different pockets throughout the years.

610
01:03:21,520 --> 01:03:33,360
So these children were thrilled. And I think we will see the we'll be looking back fondly at at this time for for that for that experience.

611
01:03:34,630 --> 01:03:40,420
I think it's a lovely place to stop. Thank you very much indeed. Thank you, Jodie and it's been wonderful to.