Embed Size (px)
Citation preview
An intervention model for the COVID-19 epidemic in South
AfricaProf Alex van den Heever
20 May 2020
“Model” features and objectives
• Follows the SEIR approaches
• Top-down
• Matches interventions to disease trajectory
• Purpose: examine optimal strategic intervention scenarios – and is not used for forecasting – focus is on informing decisions
• Estimates produced till 31 December 2020
• Calibrates “model” using intervention success in other countries and assumes we should get similar results in South Africa (which we haven’t)
Generating the model parameters
1.9
8
1.9
8
1.9
8
1.9
8
1.9
8
1.9
8
1.9
8
1.9
8
1.9
8
1.9
8
1.9
8
1.9
8
1.9
8
1.9
8
1.9
8
1.9
8
1.9
8
1.9
8
1.9
8
1.5
8
1.5
8
1.5
8
1.5
8
1.5
8
1.5
8
1.5
8
1.5
8
1.5
8
0.9
6
0.9
6
0.9
6
0.9
6
0.9
6
0.9
6
0.9
6
0.9
6
0.9
6
0.9
6
0.9
6
0.9
6
0.9
6
0.9
6
0.9
6
0.9
6
0
1
1
2
2
3
0
1 000
2 000
3 000
4 000
5 000
6 000
7 000
8 000
1 3 5 7 9
11
13
15
17
19
21
23
25
27
29
31
33
35
37
39
41
43
Re
pro
du
ctio
n R
ate
®
Ne
w in
fect
ion
s
Day
New infections (est) New infections (actual) Reproduction Rate (R)
Italy
What is the R associated with a lockdown?
0
0
0
1
1
1
1
1
2
2
0
100
200
300
400
500
600
700
800
900
1 000
1 4 7
10
13
16
19
22
25
28
31
34
37
40
43
46
49
52
55
58
61
64
67
70
73
76
79
82
Re
pro
du
ctio
n R
ate
®
Ne
w in
fect
ion
s
Date
New infections (est) New infections (actual) Reproduction Rate (R)
R = 1.6
R = 0.7What is the R associated with mass testing and contact tracing?
South Korea
How do we get to scenarios for 2020?
Individual interventions with associated reductions in R
Response packages involving combinations of interventions with an aggregate R
2020 scenarios involving the application of different packages to specific periods
Interventions Days R
Normal behaviour - cold climate 4 2.43
Normal behaviour - hot climate 4 -0.24
Border screening and advice 4 -0.05
Close borders 4 -0.20
Social distancing 4 -0.30
Employer protocols 4 -0.30
Lock-down - targeted 4 -0.20
Lock-down - general 4 -0.55
Targeted testing and contact tracing 4 -0.05
Mass testing and contact tracing 4 -0.50
Level 4 Intervention
Intervention scenario Days R Intervention
Normal behaviour - cold climate 4 2.43
Normal behaviour - hot climate 4 0.00 No
Border screening and advice 4 -0.05 Yes
Close borders 4 -0.20 Yes
Social distancing 4 -0.30 Yes
Lock-down - targeted 4 0.00 No
Lock-down - general 4 -0.55 Yes
Targeted testing and contact tracing 4 -0.55 Yes
Mass testing and contact tracing 4 0.00 No
R outcome 4 0.79
L4: General lock-down with targeted screening - cold weatherMake choices about the
configuration of interventions
(packages)
March December
SCENARIOS TO DECEMBER 2020 – each package has an aggregate R based on the intervention configurations
Inte
rve
nti
on
s an
d t
he
as
sum
ed
re
du
ctio
n in
RIn
terv
en
tio
n p
acka
ges
Scenarios are generated by day of the year
Level 2 Intervention
Intervention scenario Days R Intervention
Normal behaviour - cold climate 4 2.43
Normal behaviour - hot climate 4 -0.24 Yes
Border screening and advice 4 -0.05 Yes
Close borders 4 -0.10 Yes
Social distancing 4 -0.65 Yes
Lock-down - targeted 4 0.00 No
Lock-down - general 4 0.00 No
Targeted testing and contact tracing 4 -0.05 Yes
Mass testing and contact tracing 4 0.00 No
R outcome 4 1.34
L2: Intermediate - warm weatherLevel 3 Intervention
Intervention scenario Days R Intervention
Normal behaviour - cold climate 4 2.43
Normal behaviour - hot climate 4 -0.24 Yes
Border screening and advice 4 -0.05 Yes
Close borders 4 -0.10 Yes
Social distancing 4 -0.65 Yes
Lock-down - targeted 4 -0.20 Yes
Lock-down - general 4 0.00 No
Targeted testing and contact tracing 4 0.00 No
Mass testing and contact tracing 4 -0.40 Yes
R outcome 4 0.79
L3: Mass screening with selective lock-downs - warm weatherLevel 4 Intervention
Intervention scenario Days R Intervention
Normal behaviour - cold climate 4 2.43
Normal behaviour - hot climate 4 -0.24 Yes
Border screening and advice 4 -0.05 Yes
Close borders 4 -0.10 Yes
Social distancing 4 -0.65 Yes
Lock-down - targeted 4 0.00 No
Lock-down - general 4 -0.62 Yes
Targeted testing and contact tracing 4 -0.05 Yes
Mass testing and contact tracing 4 0.00 No
R outcome 4 0.72
L4: General lock-downs with targeted screening - warm weather
L0: No intervention - warm weather 2.19
L0: No intervention - cold weather 2.43
L1: Limited - warm weather 2.09
L1: Limited - cold weather 1.98
L2: Intermediate - warm weather 1.34
L2: Intermediate - cold weather 1.58
L3: Mass screening with selective lock-downs - warm weather 0.79
L3: Mass screening with selective lock-downs - cold weather 1.03
L4: General lock-downs with targeted screening - warm weather 0.72
L4: General lock-down with targeted screening - cold weather 0.96
Do the parameters predict the South African epidemic –accounting for interventions?
0
1
2
3
4
5
6
7
8
0
50
100
150
200
250
300
05
Mar
ch 2
02
0
07
Mar
ch 2
02
0
09
Mar
ch 2
02
0
11
Mar
ch 2
02
0
13
Mar
ch 2
02
0
15
Mar
ch 2
02
0
17
Mar
ch 2
02
0
19
Mar
ch 2
02
0
21
Mar
ch 2
02
0
23
Mar
ch 2
02
0
25
Mar
ch 2
02
0
27
Mar
ch 2
02
0
29
Mar
ch 2
02
0
31
Mar
ch 2
02
0
02
Ap
ril 2
02
0
04
Ap
ril 2
02
0
06
Ap
ril 2
02
0
08
Ap
ril 2
02
0
10
Ap
ril 2
02
0
12
Ap
ril 2
02
0
14
Ap
ril 2
02
0
16
Ap
ril 2
02
0
Mo
rtal
ity
Ne
w in
fect
ion
s
Date
New infections (est) New infections (actual) Mortality (est) Mortality (actual)
Limited intervention
Social distancing
Lockdown
0
100
200
300
400
500
600
700
800
05
Mar
ch 2
02
0
12
Mar
ch 2
02
0
19
Mar
ch 2
02
0
26
Mar
ch 2
02
0
02
Ap
ril 2
02
0
09
Ap
ril 2
02
0
16
Ap
ril 2
02
0
23
Ap
ril 2
02
0
30
Ap
ril 2
02
0
07
May
20
20
Ne
w in
fect
ion
s
Date
Series2 Series1
Social distancing
LockdownLimited intervention
0
500
1 000
1 500
2 000
2 500
3 000
3 500
05
Mar
ch 2
02
0
12
Mar
ch 2
02
0
19
Mar
ch 2
02
0
26
Mar
ch 2
02
0
02
Ap
ril 2
02
0
09
Ap
ril 2
02
0
16
Ap
ril 2
02
0
23
Ap
ril 2
02
0
30
Ap
ril 2
02
0
07
May
20
20
14
May
20
20
21
May
20
20
28
May
20
20
Ne
w in
fect
ion
s
Date
New infections (est) New infections (actual) 5 per. Mov. Avg. (New infections (actual))
Lockdown-dependent strategy
0 1 2 3 4 5 6 7 8 9 10
0
100
200
300
400
500
600
05 M
arch
202
0
05 A
pril
2020
05 M
ay 2
020
05 Ju
ne 2
020
05 Ju
ly 2
020
05 A
ugus
t 20
20
05 S
epte
mbe
r 20
20
05 O
ctob
er 2
020
05 N
ovem
ber
2020
05 D
ecem
ber
2020
Mor
talit
y
New
infe
ctio
ns
Date
New infections Mortality (est)
Lockdown
Lockdown
Lockdown
Intermediate
L0: No intervention - warm weather 2.19
L0: No intervention - cold weather 2.43
L1: Limited - warm weather 2.09
L1: Limited - cold weather 1.98
L2: Intermediate - warm weather 1.34
L2: Intermediate - cold weather 1.58
L3: Mass screening with selective lock-downs - warm weather 0.37
L3: Mass screening with selective lock-downs - cold weather 0.61
L4: General lock-downs with targeted screening - warm weather 0.72
L4: General lock-down with targeted screening - cold weather 0.96
Intervention level options
Choose intervention scenario
Lockdown-dependent
These reflect the number of people each
infected person would infect associated
with each intervention level. Above you
can selected the overall scenario to 31
December 2020
0
1
2
3
4
5
6
7
8
9
10
0
100
200
300
400
500
600
05
Mar
ch 2
02
0
05
Ap
ril 2
02
0
05
May
20
20
05
Ju
ne
202
0
05
Ju
ly 2
02
0
05
Au
gust
20
20
05
Se
pte
mb
er
202
0
05
Oct
ob
er
202
0
05
No
vem
ber
202
0
05
Dec
em
ber
20
20
Mo
rtal
ity
Ne
w in
fect
ion
s
Date
New infections Mortality (est)
35 days 142 days 16 days
L2:
Inte
rme
dia
te
L2:
Inte
rme
dia
te
L2:
Inte
rme
dia
teL4: Lockdown L4: Lockdown
L4: Lockdown
Risk-adjusted strategy
Lockdown
Mass screening
Intermediate
L0: No intervention - warm weather 2.19
L0: No intervention - cold weather 2.43
L1: Limited - warm weather 2.09
L1: Limited - cold weather 1.98
L2: Intermediate - warm weather 1.34
L2: Intermediate - cold weather 1.58
L3: Mass screening with selective lock-downs - warm weather 0.37
L3: Mass screening with selective lock-downs - cold weather 0.61
L4: General lock-downs with targeted screening - warm weather 0.72
L4: General lock-down with targeted screening - cold weather 0.96
Choose intervention scenario
Risk-adjusted
These reflect the number of people each
infected person would infect associated
with each intervention level. Above you
can selected the overall scenario to 31
December 2020
L3: Mass testing and contact-tracing
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
0
50
100
150
200
250
300
05
Mar
ch 2
02
0
05
Ap
ril 2
02
0
05
May
20
20
05
Ju
ne
202
0
05
Ju
ly 2
02
0
05
Au
gust
20
20
05
Se
pte
mb
er
202
0
05
Oct
ob
er
202
0
05
No
vem
ber
202
0
05
Dec
em
ber
20
20
Mo
rtal
ity
Ne
w in
fect
ion
s
Date
New infections Mortality (est)
35 days184 days
61 days
L2: Intermediate
L4: Lockdown
![The reproductive index from SEIR model of Covid-19 ...Apr 24, 2020 · [2] Faculty of Business Administration, Huachiew Chalermprakiet University, Samutprakan province, 10540, Thailand](https://img.pdfslide.us/doc/110x75/60bac6e9b6a82a5a7b30b6cb/the-reproductive-index-from-seir-model-of-covid-19-apr-24-2020-2-faculty.jpg)
![Complete maximum likelihood estimation for SEIR epidemic ... · arXiv:1907.10679v1 [q-bio.PE] 24 Jul 2019 Complete maximum likelihood estimation for SEIR epidemic models: theoretical](https://img.pdfslide.us/doc/110x75/5fb37461f92b52058f5c53bd/complete-maximum-likelihood-estimation-for-seir-epidemic-arxiv190710679v1.jpg)
