-
Notifications
You must be signed in to change notification settings - Fork 3
/
Copy pathspm_COVID_priors.m
268 lines (238 loc) · 11.8 KB
/
spm_COVID_priors.m
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
function [P,C,str,rfx] = spm_COVID_priors
% Generate prior expectation and covariance log parameters
% FORMAT [pE,pC,str,rfx] = spm_COVID_priors
%
% pE - prior expectation (structure)
% pC - prior covariances (structure)
% str.factor - latent or hidden factors
% str.factors - levels of each factor
% str.outcome - outcome names (see spm_COVID_gen)
% str.names - parameter names
% str.field - field names of random effects
% rfx - indices of random effects
%
% This routine assembles the (Gaussian) and priors over the parameters of a
% generative model for COVID-19. This generative model is based upon a mean
% field approximation to ensemble of population dynamics, in which four
% marginal distributions are coupled through probability transition
% matrices. The marginal distributions correspond to 4 factors; namely,
% location, infection, symptom and testing (LIST) states. The parameters of
% this model determine the initial (probability) states and the transitions
% among the states that show certain conditional independences.
%
% These parameters can either be interpreted in terms of the probability of
% moving from one state to another of a given factor, conditioned on
% another. Alternatively, in some instances (specifically, staying in the
% same state),the parameters can be thought of as log transformed rate
% constants or inverse time constants.
%
% All the parameters of this generative model are log scale parameters. In
% other words, the parameters are non-negative but are encoded in terms of
% their logarithms. This means that priors over parameters can be specified
% in terms of a prior expectation and covariance and Gaussian assumptions
% (i.e., lognormal priors over scale parameters).
%__________________________________________________________________________
% Copyright (C) 2020 Wellcome Centre for Human Neuroimaging
% Karl Friston
% $Id: spm_COVID_priors.m 7929 2020-08-16 13:43:49Z karl $
# SPDX-License-Identifier: GPL-2.0
% sources and background
%--------------------------------------------------------------------------
% https://jamanetwork.com/journals/jama/fullarticle/2761044
% https://www.statista.com/chart/21105/number-of-critical-care-beds-per-100000-inhabitants/
%--------------------------------------------------------------------------
% The NHS also maintains critical care beds for patients who are seriously
% ill and require constant support. Unlike most other categories of
% hospital bed, the total number of critical care beds has increased in
% recent years. In 2011/12 there were around 5,400 critical care beds, by
% 2019/20 this had risen to 5,900 (NHS England 2019b) (Figure 5). Of these,
% around 70 per cent are for use by adults and the remainder for children
% and infants.
%--------------------------------------------------------------------------
% https://www.gov.uk/guidance/coronavirus-covid-19-information-for-the-public
%
% As of 9am on 28 March 2020, a total of 120,776 people have been tested,
% of which 103,687 were confirmed negative and 17,089 were confirmed
% positive.
%--------------------------------------------------------------------------
% As of 5pm on 27 March 2020, 1,019 patients in the UK who tested positive
% for coronavirus (COVID-19) have died.
%--------------------------------------------------------------------------
% parameter names (where %** denotes fixed effects)
%==========================================================================
names{1} = 'initial cases'; %**
names{2} = 'population size';
names{3} = 'initial proportion';
names{4} = 'P(work | home)';
names{5} = 'social distancing';
names{6} = 'bed availability';
names{7} = 'contacts: home';
names{8} = 'contacts: work';
names{9} = 'transmission strength';
names{10} = 'infected period';
names{11} = 'contagious period';
names{12} = 'incubation period';
names{13} = 'P(ARDS | symptoms)';
names{14} = 'symptomatic period';
names{15} = 'time in CCU';
names{16} = 'P(fatality | CCU)';
names{17} = 'P(survival | home)';
names{18} = 'track and trace'; %**
names{19} = 'testing latency'; %**
names{20} = 'test delay'; %**
names{21} = 'test selectivity'; %**
names{22} = 'sustained testing'; %**
names{23} = 'baseline testing'; %**
names{24} = 'immune period'; %**
names{25} = 'exempt period'; %**
names{26} = 'resistance'; %**
names{27} = 'innate immunity'; %**
% random effects (i.e., effects that are common in countries)
%--------------------------------------------------------------------------
rfx = 2:17;
% latent or hidden factors
%--------------------------------------------------------------------------
factors = {'Location','Infection','Symptoms','Testing'};
factor{1} = {'home','work','hospital','removed','isolated'};
factor{2} = {'susceptible','infected','infectious','Ab +ve','Ab -ve'};
factor{3} = {'none','symptoms','ARDS','deceased'};
factor{4} = {'untested','waiting','PCR +ve','PCR -ve'};
% labels or strings for plotting
%--------------------------------------------------------------------------
str.outcome = {'Death rate',...
'New tests',...
'CCU occupancy',...
'ERR',...
'Herd immunity',...
'Tests',...
'Contagion risk (%)',...
'Prevalence {%}',...
'FTTI target',...
'New cases'};
str.factors = factors;
str.factor = factor;
str.names = names;
% cut and paste to see the effects of changing different prior expectations
%xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
if false
pE = spm_COVID_priors; M.T = 365;
[Y,X] = spm_COVID_gen(pE,M,1:3);
u = exp(pE.cap + pE.N)*1e6;
spm_COVID_plot(Y,X,[],u)
end
%xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
% https://annals.org/aim/fullarticle/2762808/incubation-period-coronavirus-disease-2019-covid-19-from-publicly-reported
% Expectations (either heuristic or taken from the above sources)
%==========================================================================
P.n = 4; % number of initial cases
P.N = 8; % population size (in millions)
P.m = 1/4; % initial proportion
% location parameters
%--------------------------------------------------------------------------
P.out = 1/3; % P(work | home)
P.sde = 1/32; % social distancing threshold
P.cap = 16/100000; % bed availability threshold (per capita)
% infection parameters
%--------------------------------------------------------------------------
P.Rin = 4; % effective number of contacts: home
P.Rou = 48; % effective number of contacts: work
P.trn = 1/3; % P(contagion | contact)
P.Tin = 4; % infected (pre-contagious) period
P.Tcn = 4; % infectious (contagious) period
% clinical parameters
%--------------------------------------------------------------------------
P.Tic = 16; % time till symptoms (days)
P.sev = 1/32; % P(severe symptoms | symptomatic)
P.Tsy = 8; % symptomatic period
P.Trd = 10; % CCU period
P.fat = 1/2; % P(fatality | severe, CCU)
P.sur = 1/8; % P(survival | severe, home)
% testing parameters
%--------------------------------------------------------------------------
P.ttt = 1/10000; % test, track and trace
P.ont = 2; % testing latency (months)
P.del = 4; % test delay (days)
P.tes = 8; % test selectivity (for infection)
P.sus = 4/10000; % sustained testing
P.bas = 4/10000; % baseline testing
% immunity
%--------------------------------------------------------------------------
P.Tim = 16; % period of immunity (months)
P.Tex = 2; % period of exemption (days)
P.r = 1/2; % proportion resistant cases
P.res = 1/2; % proportion with innate immunity
P.tts = 16; % testing buildup
% total mortality rate (for susceptible population)
%--------------------------------------------------------------------------
% IFR (hospital): P.sev*P.fat*100
% IFR (carehome): P.sev*(1 - P.sur)*100
% IFR (hospital): exp(Ep.sev)*exp(Ep.fat)*100
% IFR (carehome): exp(Ep.sev)*(1 - exp(Ep.sur))*100
% We describe what we believe is the first instance of complete COVID-19
% testing of all passengers and crew on an isolated cruise ship during the
% current COVID-19 pandemic. Of the 217 passengers and crew on board, 128
% tested positive for COVID-19 on reverse transcription-PCR (59 pc). Of the
% COVID- 19-positive patients, 19 pc (24) were symptomatic; 6.2 pc (8)
% required medical evacuation; 3.1 pc (4) were intubated and ventilated;
% and the mortality was 0.8 pc (1). The majority of COVID-19-positive
% patients were asymptomatic (81 pc, 104 patients).
% probability of developing symptoms if infected (19%)
%--------------------------------------------------------------------------
% 1 - exp(-1/P.Tic)^(P.Tin + P.Tcn)
% probability of ARDS if symptomatic (4/24)
%--------------------------------------------------------------------------
% P.sev
% probability of dying if ARDS (1/4)
%--------------------------------------------------------------------------
% P.fat
% Variances (mildly informative priors, apart from initial cases and size)
%==========================================================================
U = 1; % flat priors
V = 1/16; % informative priors
W = 1/256; % precise priors
C.n = U; % number of initial cases
C.N = U; % size of population with mixing
C.m = W; % initial proportion
% location parameters
%--------------------------------------------------------------------------
C.out = W; % P(going home | work)
C.sde = W; % social distancing threshold
C.cap = W; % bed availability threshold (per capita)
% infection parameters
%--------------------------------------------------------------------------
C.Rin = V; % effective number of contacts: home
C.Rou = V; % effective number of contacts: work
C.trn = V; % P(transmission | infectious)
C.Tin = W; % infected (pre-contagious) period
C.Tcn = W; % contagious period
% clinical parameters
%--------------------------------------------------------------------------
C.Tic = W; % time until symptoms
C.sev = W; % P(severe symptoms | symptomatic)
C.Tsy = W; % symptomatic period
C.Trd = W; % period in CCU
C.fat = W; % P(fatality | CCU)
C.sur = W; % P(fatality | home)
% testing parameters
%--------------------------------------------------------------------------
C.ttt = U; % test, track and trace
C.ont = U; % testing latency (months)
C.del = W; % test delay (days)
C.tes = V; % test selectivity (for infection)
C.sus = U; % sustained testing
C.bas = V; % baseline testing
% immunity
%--------------------------------------------------------------------------
C.Tim = W; % period of immunity
C.Tex = W; % period of exemption
C.r = W; % proportion of people not susceptible
C.res = W; % proportion with innate immunity
C.tts = U; % testing buildup
% log transform
%==========================================================================
P = spm_vecfun(P,@log);
% field names of random effects
%--------------------------------------------------------------------------
field = fieldnames(P);
str.field = field(rfx);
return