forked from cmmid/covid-uk
-
Notifications
You must be signed in to change notification settings - Fork 2
/
UK.R
489 lines (420 loc) · 24.2 KB
/
UK.R
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
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
# - - - - - - - - - - - - - - - - - - - - - - -
# UK model: load data and analyse scenarios
# - - - - - - - - - - - - - - - - - - - - - - -
library(rlang)
library(stringr)
# Set path
# Set this path to the base directory of the repository.
##covid_uk_path = "~/Dropbox/COVID-UK"
covid_uk_path = "./"
# covidm options
cm_path = paste0(covid_uk_path, "/covidm/");
if (grepl(Sys.info()["user"], pattern = "^adamkuchars(ki)?$")) { cm_path = "~/Documents/GitHub/covid-uk/covidm/" }
source(paste0(cm_path, "/R/covidm.R"))
# build parameters for entire UK, for setting R0.
parametersUK1 = cm_parameters_SEI3R(cm_uk_locations("UK", 0),
dE = cm_delay_gamma(4.0, 4.0, t_max = 60, t_step = 0.25)$p,
dIp = cm_delay_gamma(1.5, 4.0, t_max = 60, t_step = 0.25)$p,
dIs = cm_delay_gamma(3.5, 4.0, t_max = 60, t_step = 0.25)$p,
dIa = cm_delay_gamma(5.0, 4.0, t_max = 60, t_step = 0.25)$p,
deterministic = F);
# build parameters for regions of UK, down to the county level (level 3).
locations = cm_uk_locations("UK", 3);
parameters = cm_parameters_SEI3R(locations, date_start = "2020-01-29", date_end = "2021-12-31",
dE = cm_delay_gamma(4.0, 4.0, t_max = 60, t_step = 0.25)$p, # 6.5 day serial interval.
dIp = cm_delay_gamma(1.5, 4.0, t_max = 60, t_step = 0.25)$p, # 1.5 days w/o symptoms
dIs = cm_delay_gamma(3.5, 4.0, t_max = 60, t_step = 0.25)$p, # 5 days total of infectiousness
dIa = cm_delay_gamma(5.0, 4.0, t_max = 60, t_step = 0.25)$p, # 5 days total of infectiousness here as well.
deterministic = F);
# Split off the elderly (70+, age groups 15 and 16) so their contact matrices can be manipulated separately
parameters = cm_split_matrices_ex_in(parameters, 15);
# Create additional matrix for child-elderly contacts
for (j in seq_along(parameters$pop))
{
# Recover home/other contact matrix
mat_ref = parameters$pop[[j]]$matrices[[1]] + parameters$pop[[j]]$matrices[[4]] +
parameters$pop[[j]]$matrices[[5]] + parameters$pop[[j]]$matrices[[8]];
gran = 5/7; # adjustment for weekdays only.
N = nrow(mat_ref);
popsize = parameters$pop[[j]]$size;
mat = matrix(0, ncol = N, nrow = N);
# Add child-grandparent contacts: under 15s to 55+s
for (a in 1:3) {
dist = c(rep(0, 10 + a), mat_ref[a, (11 + a):N]);
dist = dist/sum(dist);
mat[a, ] = mat[a, ] + gran * dist;
mat[, a] = mat[, a] + (gran * dist) * (popsize[a] / popsize);
}
# Add child-grandparent contact matrix to population
parameters$pop[[j]]$matrices$gran = mat;
parameters$pop[[j]]$contact = c(parameters$pop[[j]]$contact, 0);
}
# Health burden processes
probs = fread(
"Age,Prop_symptomatic,IFR,Prop_inf_hosp,Prop_inf_critical,Prop_critical_fatal,Prop_noncritical_fatal,Prop_symp_hospitalised,Prop_hospitalised_critical
10,0.66,8.59E-05,0.002361009,6.44E-05,0.5,0,0,0.3
20,0.66,0.000122561,0.003370421,9.19E-05,0.5,9.47E-04,0.007615301,0.3
30,0.66,0.000382331,0.010514103,0.000286748,0.5,0.001005803,0.008086654,0.3
40,0.66,0.000851765,0.023423527,0.000638823,0.5,0.001231579,0.009901895,0.3
50,0.66,0.001489873,0.0394717,0.001117404,0.5,0.002305449,0.018535807,0.3
60,0.66,0.006933589,0.098113786,0.005200192,0.5,0.006754596,0.054306954,0.3
70,0.66,0.022120421,0.224965092,0.016590316,0.5,0.018720727,0.150514645,0.3
80,0.66,0.059223786,0.362002579,0.04441784,0.5,0.041408882,0.332927412,0.3
100,0.66,0.087585558,0.437927788,0.065689168,0.5,0.076818182,0.617618182,0.3")
reformat = function(P)
{
# 70-74,3388.488 75-79,2442.147 80-84,1736.567 85-89,1077.555 90-94,490.577 95-99,130.083 100+,15.834
x = c(P[1:7], weighted.mean(c(P[8], P[9]), c(3388.488 + 2442.147, 1736.567 + 1077.555 + 490.577 + 130.083 + 15.834)));
return (rep(x, each = 2))
}
P.icu_symp = reformat(probs[, Prop_symp_hospitalised * Prop_hospitalised_critical]);
P.nonicu_symp = reformat(probs[, Prop_symp_hospitalised * (1 - Prop_hospitalised_critical)]);
P.death_icu = reformat(probs[, Prop_critical_fatal]);
P.death_nonicu = reformat(probs[, Prop_noncritical_fatal]);
burden_processes = list(
list(source = "Ip", type = "multinomial", names = c("to_icu", "to_nonicu", "null"), report = c("", "", ""),
prob = matrix(c(P.icu_symp, P.nonicu_symp, 1 - P.icu_symp - P.nonicu_symp), nrow = 3, ncol = 16, byrow = T),
delays = matrix(c(cm_delay_gamma(7, 7, 60, 0.25)$p, cm_delay_gamma(7, 7, 60, 0.25)$p, cm_delay_skip(60, 0.25)$p), nrow = 3, byrow = T)),
list(source = "to_icu", type = "multinomial", names = "icu", report = "p",
prob = matrix(1, nrow = 1, ncol = 16, byrow = T),
delays = matrix(cm_delay_gamma(10, 10, 60, 0.25)$p, nrow = 1, byrow = T)),
list(source = "to_nonicu", type = "multinomial", names = "nonicu", report = "p",
prob = matrix(1, nrow = 1, ncol = 16, byrow = T),
delays = matrix(cm_delay_gamma(8, 8, 60, 0.25)$p, nrow = 1, byrow = T)),
list(source = "Ip", type = "multinomial", names = c("death", "null"), report = c("o", ""),
prob = matrix(c(P.death_nonicu, 1 - P.death_nonicu), nrow = 2, ncol = 16, byrow = T),
delays = matrix(c(cm_delay_gamma(22, 22, 60, 0.25)$p, cm_delay_skip(60, 0.25)$p), nrow = 2, byrow = T))
)
parameters$processes = burden_processes
# Observer for lockdown scenarios
observer_lockdown = function(lockdown_trigger) function(time, dynamics)
{
# Get current icu prevalence
icu_prevalence = dynamics[t == time, sum(icu_p)];
# Determine lockdown trigger
trigger = lockdown_trigger;
# If ICU prevalence exceeds a threshold, turn on lockdown
if (icu_prevalence >= trigger) {
return (list(csv = paste(time, "trace_lockdown", "All", 2, sep = ","),
changes = list(contact_lowerto = c(1, 0.1, 0.1, 0.1, 1, 0.1, 0.1, 0.1, 1))));
} else {
return (list(csv = paste(time, "trace_lockdown", "All", 1, sep = ","),
changes = list(contact_lowerto = c(1, 1, 1, 1, 1, 1, 1, 1, 1))));
}
return (list(csv = paste(time, "trace_lockdown", "All", 1, sep = ",")))
}
# Load age-varying symptomatic rate
covid_scenario = qread(paste0(covid_uk_path, "/data/2-linelist_symp_fit_fIa0.5.qs"));
# Identify London boroughs for early seeding, and regions of each country for time courses
london = cm_structure_UK[match(str_sub(locations, 6), Name), Geography1 %like% "London"]
england = cm_structure_UK[match(str_sub(locations, 6), Name), Code %like% "^E" & !(Geography1 %like% "London")]
wales = cm_structure_UK[match(str_sub(locations, 6), Name), Code %like% "^W"]
scotland = cm_structure_UK[match(str_sub(locations, 6), Name), Code %like% "^S"]
nireland = cm_structure_UK[match(str_sub(locations, 6), Name), Code %like% "^N"]
westmid = cm_structure_UK[match(str_sub(locations, 6), Name), Name == "West Midlands (Met County)"]
cumbria = cm_structure_UK[match(str_sub(locations, 6), Name), Name == "Cumbria"]
save = function(run)
{
# if (analysis == 3) {
# filename = paste0("~/Dropbox/COVID-UK Storage/", run$dynamics$scenario[1], "-", run$dynamics$run[1], ".qs");
# cm_save(run, filename);
# }
}
add_totals = function(run, totals)
{
regions = run$dynamics[, unique(population)];
# totals by age
totals0 = run$dynamics[, .(total = sum(value)), by = .(scenario, run, compartment, group)];
return (rbind(totals, totals0))
}
add_dynamics = function(run, dynamics, iv)
{
regions = run$dynamics[, unique(population)];
interv = data.table(scenario = run$dynamics$scenario[1], run = run$dynamics$run[1], t = unique(run$dynamics$t),
compartment = "trace_school", region = "All", value = unlist(iv$trace_school));
if (!is.null(iv$trace_intervention)) {
interv = rbind(interv,
data.table(scenario = run$dynamics$scenario[1], run = run$dynamics$run[1], t = unique(run$dynamics$t),
compartment = "trace_intervention", region = "All", value = unlist(iv$trace_intervention)));
} else {
interv = rbind(interv,
data.table(scenario = run$dynamics$scenario[1], run = run$dynamics$run[1], t = unique(run$dynamics$t),
compartment = "trace_intervention", region = "All", value = 1));
}
csvlines = NULL;
if (nchar(run$csv[[1]]) > 0) {
csvlines = fread(run$csv[[1]], header = F);
csvlines = cbind(run$dynamics$scenario[1], run$dynamics$run[1], csvlines);
names(csvlines) = c("scenario", "run", "t", "compartment", "region", "value");
csvlines = unique(csvlines);
}
# time courses
return (rbind(dynamics,
run$dynamics[population %in% locations[westmid], .(region = "West Midlands", value = sum(value)), by = .(scenario, run, t, compartment)],
run$dynamics[population %in% locations[cumbria], .(region = "Cumbria", value = sum(value)), by = .(scenario, run, t, compartment)],
run$dynamics[population %in% locations[london], .(region = "London", value = sum(value)), by = .(scenario, run, t, compartment)],
run$dynamics[population %in% locations[england], .(region = "England", value = sum(value)), by = .(scenario, run, t, compartment)],
run$dynamics[population %in% locations[wales], .(region = "Wales", value = sum(value)), by = .(scenario, run, t, compartment)],
run$dynamics[population %in% locations[scotland], .(region = "Scotland", value = sum(value)), by = .(scenario, run, t, compartment)],
run$dynamics[population %in% locations[nireland], .(region = "Northern Ireland", value = sum(value)), by = .(scenario, run, t, compartment)],
run$dynamics[, .(region = "United Kingdom", value = sum(value)), by = .(scenario, run, t, compartment)],
interv,
csvlines
))
}
#############
# MAIN CODE #
#############
argv = commandArgs(trailingOnly = T);
argc = length(argv);
if (argc != 2) {
stop("Must provide two arguments: analysis set and number of runs.");
}
analysis = as.numeric(argv[argc-1]);
n_runs = as.numeric(argv[argc]);
if (analysis == 1) {
# Define school terms, base versus intervention (both same here)
school_close_b = c("2020-2-16", "2020-4-05", "2020-5-24", "2020-7-22", "2020-10-25", "2020-12-20", "2021-02-14", "2021-04-01", "2021-05-30", "2021-07-25");
school_reopen_b = c("2020-2-22", "2020-4-18", "2020-5-30", "2020-9-01", "2020-10-31", "2021-01-02", "2021-02-20", "2021-04-17", "2021-06-05", "2021-09-01");
school_close_i = c("2020-2-16", "2020-4-05", "2020-5-24", "2020-7-22", "2020-10-25", "2020-12-20", "2021-02-14", "2021-04-01", "2021-05-30", "2021-07-25");
school_reopen_i = c("2020-2-22", "2020-4-18", "2020-5-30", "2020-9-01", "2020-10-31", "2021-01-02", "2021-02-20", "2021-04-17", "2021-06-05", "2021-09-01");
# Define interventions to be used
interventions = list(
`School Closures` = list(contact = c(1.0, 1.0, 0.0, 1.0, 1.0, 1.0, 0.0, 1.0, 0)),
`Social Distancing` = list(contact = c(1.0, 0.5, 1.0, 0.5, 1.0, 0.5, 1.0, 0.5, 0)),
`Elderly Shielding` = list(contact = c(1.0, 1.0, 1.0, 1.0, 1.0, 0.25, 1.0, 0.25, 0)),
`Self-Isolation` = list(fIs = rep(0.65, 16)),
`Combination` = list(contact = c(1.0, 0.5, 0.0, 0.5, 1.0, 0.25, 0.0, 0.25, 0), fIs = rep(0.65, 16))
);
# Set options
option.trigger = "national";
option.duration = 7 * 12;
option.lockdown = NA;
option.intervention_shift = 0;
} else if (analysis == 2.1) {
# Define school terms, base versus intervention (both same here)
school_close_b = c("2020-2-16", "2020-4-05", "2020-5-24", "2020-7-22", "2020-10-25", "2020-12-20", "2021-02-14", "2021-04-01", "2021-05-30", "2021-07-25");
school_reopen_b = c("2020-2-22", "2020-4-18", "2020-5-30", "2020-9-01", "2020-10-31", "2021-01-02", "2021-02-20", "2021-04-17", "2021-06-05", "2021-09-01");
school_close_i = c("2020-2-16", "2020-4-05", "2020-5-24", "2020-7-22", "2020-10-25", "2020-12-20", "2021-02-14", "2021-04-01", "2021-05-30", "2021-07-25");
school_reopen_i = c("2020-2-22", "2020-4-18", "2020-5-30", "2020-9-01", "2020-10-31", "2021-01-02", "2021-02-20", "2021-04-17", "2021-06-05", "2021-09-01");
# Define interventions to be used
interventions = list(
`Combination` = list(contact = c(1.0, 0.5, 0.0, 0.5, 1.0, 0.25, 0.0, 0.25, 0), fIs = rep(0.65, 16))
);
# Set options
option.trigger = "national";
option.duration = 7 * 12;
option.lockdown = NA;
option.intervention_shift = c(0, 14, 28, 56);
} else if (analysis == 2.2) {
# Define school terms, base versus intervention (both same here)
school_close_b = c("2020-2-16", "2020-4-05", "2020-5-24", "2020-7-22", "2020-10-25", "2020-12-20", "2021-02-14", "2021-04-01", "2021-05-30", "2021-07-25");
school_reopen_b = c("2020-2-22", "2020-4-18", "2020-5-30", "2020-9-01", "2020-10-31", "2021-01-02", "2021-02-20", "2021-04-17", "2021-06-05", "2021-09-01");
school_close_i = c("2020-2-16", "2020-4-05", "2020-5-24", "2020-7-22", "2020-10-25", "2020-12-20", "2021-02-14", "2021-04-01", "2021-05-30", "2021-07-25");
school_reopen_i = c("2020-2-22", "2020-4-18", "2020-5-30", "2020-9-01", "2020-10-31", "2021-01-02", "2021-02-20", "2021-04-17", "2021-06-05", "2021-09-01");
# Define interventions to be used
interventions = list(
`Combination` = list(contact = c(1.0, 0.5, 0.0, 0.5, 1.0, 0.25, 0.0, 0.25, 0), fIs = rep(0.65, 16))
);
# Set options
option.trigger = "local";
option.duration = 7 * 12;
option.lockdown = NA;
option.intervention_shift = c(0, 14, 28, 56);
} else if (analysis == 3) {
# Define school terms, base versus intervention (schools close from 23 March)
school_close_b = c("2020-2-16", "2020-4-05", "2020-5-24", "2020-7-22", "2020-10-25", "2020-12-20", "2021-02-14", "2021-04-01", "2021-05-30", "2021-07-25");
school_reopen_b = c("2020-2-22", "2020-4-18", "2020-5-30", "2020-9-01", "2020-10-31", "2021-01-02", "2021-02-20", "2021-04-17", "2021-06-05", "2021-09-01");
school_close_i = c("2020-2-16", "2020-3-23", "2020-10-25", "2020-12-20", "2021-02-14", "2021-04-01", "2021-05-30", "2021-07-25");
school_reopen_i = c("2020-2-22", "2020-9-01", "2020-10-31", "2021-01-02", "2021-02-20", "2021-04-17", "2021-06-05", "2021-09-01");
# Define interventions to be used
interventions = list(
`Intensive Interventions` = list(contact = c(1, 0.655, 1, 0.59155, 1, 0.25, 1, 0.157375, 0), fIs = rep(0.65, 16))
);
# Set options
option.trigger = "2020-03-17";
option.duration = 364;
option.lockdown = c(NA, 1000, 2000, 5000);
option.intervention_shift = 0;
} else if (analysis == 4) {
# Define school terms, base versus intervention (both same here)
school_close_b = c("2020-2-16", "2020-4-05", "2020-5-24", "2020-7-22", "2020-10-25", "2020-12-20", "2021-02-14", "2021-04-01", "2021-05-30", "2021-07-25");
school_reopen_b = c("2020-2-22", "2020-4-18", "2020-5-30", "2020-9-01", "2020-10-31", "2021-01-02", "2021-02-20", "2021-04-17", "2021-06-05", "2021-09-01");
school_close_i = c("2020-2-16", "2020-4-05", "2020-5-24", "2020-7-22", "2020-10-25", "2020-12-20", "2021-02-14", "2021-04-01", "2021-05-30", "2021-07-25");
school_reopen_i = c("2020-2-22", "2020-4-18", "2020-5-30", "2020-9-01", "2020-10-31", "2021-01-02", "2021-02-20", "2021-04-17", "2021-06-05", "2021-09-01");
# Define interventions to be used
interventions = list(
`Intensive` = list(contact = c(1, 0.655, 1, 0.59155, 1, 0.25, 1, 0.157375, 0.0), fIs = rep(0.65, 16)),
`Intensive + School` = list(contact = c(1, 0.655, 0, 0.59155, 1, 0.25, 0, 0.157375, 0.0), fIs = rep(0.65, 16)),
`Intensive + School + G20` = list(contact = c(1, 0.655, 0, 0.59155, 1, 0.25, 0, 0.157375, 0.2), fIs = rep(0.65, 16)),
`Intensive + School + G50` = list(contact = c(1, 0.655, 0, 0.59155, 1, 0.25, 0, 0.157375, 0.5), fIs = rep(0.65, 16)),
`Intensive + School + G100` = list(contact = c(1, 0.655, 0, 0.59155, 1, 0.25, 0, 0.157375, 1.0), fIs = rep(0.65, 16))
);
# Set options
option.trigger = "2020-03-17";
option.duration = 125;
option.lockdown = NA;
option.intervention_shift = 0;
parameters$time1 = "2020-07-20";
} else if (analysis == 6) {
# Define school terms, base versus intervention (both same here)
school_close_b = c("2020-2-16", "2020-4-05", "2020-5-24", "2020-7-22", "2020-10-25", "2020-12-20", "2021-02-14", "2021-04-01", "2021-05-30", "2021-07-25");
school_reopen_b = c("2020-2-22", "2020-4-18", "2020-5-30", "2020-9-01", "2020-10-31", "2021-01-02", "2021-02-20", "2021-04-17", "2021-06-05", "2021-09-01");
school_close_i = c("2020-2-16", "2020-4-05", "2020-5-24", "2020-7-22", "2020-10-25", "2020-12-20", "2021-02-14", "2021-04-01", "2021-05-30", "2021-07-25");
school_reopen_i = c("2020-2-22", "2020-4-18", "2020-5-30", "2020-9-01", "2020-10-31", "2021-01-02", "2021-02-20", "2021-04-17", "2021-06-05", "2021-09-01");
# Define interventions to be used
interventions = list(
`Background` = list(contact = c(1, 1, 0, 1, 1, 0.25, 0, 0.25, 0), fIs = rep(0.65, 16)),
`Background + 0% Sports` = list(contact = c(1, 1, 0, 1 - 0.0041, 1, 0.25, 0, 0.25, 0), fIs = rep(0.65, 16)),
`Background + 25% Leisure` = list(contact = c(1, 1, 0, 1 - 0.362, 1, 0.25, 0, 0.25, 0), fIs = rep(0.65, 16))
);
# Set options
option.trigger = "2020-03-17";
option.duration = 168;
option.lockdown = NA;
option.intervention_shift = 0;
parameters$time1 = "2020-09-01";
}
# Pick R0s
set.seed(9876);
R0s = rnorm(n_runs, mean = 2.675739, sd = 0.5719293)
# Do runs
dynamics = data.table()
totals = data.table()
print(Sys.time())
set.seed(1234567);
for (r in 1:n_runs) {
cat(paste0(r, ": R0 = ", R0s[r], "\n"));
# 1. Pick age-varying symptomatic rate
covy = unname(unlist(covid_scenario[sample.int(nrow(covid_scenario), 1), f_00:f_70]));
covy = rep(covy, each = 2);
# 2. Calculate R0 adjustment needed
parametersUK1$pop[[1]]$y = covy;
u_adj = R0s[r] / cm_calc_R0(parametersUK1, 1);
# 3. Pick seeding times
seed_start = ifelse(london, sample(0:6, length(london), replace = T), sample(0:20, length(london), replace = T));
# 4. Do base model
# 4a. Set parameters
params = duplicate(parameters);
for (j in seq_along(params$pop)) {
params$pop[[j]]$u = params$pop[[j]]$u * u_adj;
params$pop[[j]]$y = covy;
params$pop[[j]]$seed_times = rep(seed_start[j] + 0:27, each = 2);
params$pop[[j]]$dist_seed_ages = cm_age_coefficients(25, 50, 5 * 0:16);
}
# CALCULATE IMPACT ON R0
if (analysis == 5) {
interventions = list(
`Base` = list(),
`School Closures` = list(contact = c(1.0, 1.0, 0.0, 1.0, 1.0, 1.0, 0.0, 1.0, 0)),
`Social Distancing` = list(contact = c(1.0, 0.5, 1.0, 0.5, 1.0, 0.5, 1.0, 0.5, 0)),
`Elderly Shielding` = list(contact = c(1.0, 1.0, 1.0, 1.0, 1.0, 0.25, 1.0, 0.25, 0)),
`Self-Isolation` = list(fIs = rep(0.65, 16)),
`Combination` = list(contact = c(1.0, 0.5, 0.0, 0.5, 1.0, 0.25, 0.0, 0.25, 0), fIs = rep(0.65, 16)),
`Intensive, schools open` = list(contact = c(1, 0.655, 1, 0.59155, 1, 0.25, 1, 0.157375, 0), fIs = rep(0.65, 16)),
`Intensive, schools closed` = list(contact = c(1, 0.655, 0, 0.59155, 1, 0.25, 0, 0.157375, 0), fIs = rep(0.65, 16)),
`Lockdown` = list(contact = c(1, 0.1, 0.1, 0.1, 1, 0.1, 0.1, 0.1, 0), fIs = rep(0.65, 16))
);
for (i in seq_along(interventions))
{
iR0s = rep(0, length(params$pop));
iweights = rep(0, length(params$pop));
for (j in seq_along(params$pop))
{
for (k in seq_along(interventions[[i]]))
{
params$pop[[j]][[names(interventions[[i]])[k]]] = interventions[[i]][[k]];
}
iR0s[j] = cm_calc_R0(params, j);
iweights[j] = sum(params$pop[[j]]$size);
}
weighted_R0 = weighted.mean(iR0s, iweights);
dynamics = rbind(dynamics, data.table(run = r, scenario = names(interventions)[i], R0 = weighted_R0));
}
next;
}
# 4b. Set school terms
iv = cm_iv_build(params)
cm_iv_set(iv, school_close_b, school_reopen_b, contact = c(1, 1, 0, 1, 1, 1, 0, 1, 1), trace_school = 2);
params = cm_iv_apply(params, iv);
# 4c. Run model
run = cm_simulate(params, 1, r);
run$dynamics[, run := r];
run$dynamics[, scenario := "Base"];
run$dynamics[, R0 := R0s[r]];
save(run);
totals = add_totals(run, totals);
dynamics = add_dynamics(run, dynamics, iv);
peak_t = run$dynamics[compartment == "cases", .(total_cases = sum(value)), by = t][, t[which.max(total_cases)]];
peak_t_bypop = run$dynamics[compartment == "cases", .(total_cases = sum(value)), by = .(t, population)][, t[which.max(total_cases)], by = population]$V1;
rm(run)
gc()
# 5. Run interventions
for (i in seq_along(interventions)) {
for (duration in option.duration) {
for (trigger in option.trigger) {
for (intervention_shift in option.intervention_shift) {
for (lockdown in option.lockdown) {
cat(paste0(names(interventions)[i], "...\n"))
# 5a. Make parameters and adjust R0
params = duplicate(parameters);
for (j in seq_along(params$pop)) {
params$pop[[j]]$u = params$pop[[j]]$u * u_adj;
params$pop[[j]]$y = covy;
if (!is.na(lockdown)) {
params$pop[[j]]$observer = observer_lockdown(lockdown);
}
}
# 5b. Set interventions
if (trigger == "national") {
intervention_start = peak_t - duration / 2 + intervention_shift;
} else if (trigger == "local") {
intervention_start = peak_t_bypop - duration / 2 + intervention_shift;
} else {
intervention_start = as.numeric(ymd(trigger) - ymd(params$date0));
}
if (trigger == "local") {
# Apply interventions to one population at a time.
for (pi in seq_along(params$pop)) {
ymd_start = ymd(params$date0) + intervention_start[pi];
ymd_end = ymd_start + duration - 1;
iv = cm_iv_build(params)
cm_iv_set(iv, school_close_i, school_reopen_i, contact = c(1, 1, 0, 1, 1, 1, 0, 1, 1), trace_school = 2);
cm_iv_set(iv, ymd_start, ymd_end, interventions[[i]]);
cm_iv_set(iv, ymd_start, ymd_end, trace_intervention = 2);
params = cm_iv_apply(params, iv, pi);
}
} else {
# Apply interventions to entire population.
ymd_start = ymd(params$date0) + intervention_start;
ymd_end = ymd_start + duration - 1;
iv = cm_iv_build(params)
cm_iv_set(iv, school_close_i, school_reopen_i, contact = c(1, 1, 0, 1, 1, 1, 0, 1, 1), trace_school = 2);
cm_iv_set(iv, ymd_start, ymd_end, interventions[[i]]);
cm_iv_set(iv, ymd_start, ymd_end, trace_intervention = 2);
params = cm_iv_apply(params, iv);
}
# 5c. Run model
run = cm_simulate(params, 1, r);
tag = "";
if (length(option.duration) > 1) { tag = paste0(tag, " ", duration + 1, " day"); }
if (length(option.lockdown) > 1) { tag = paste0(tag, " ", ifelse(lockdown >= 0, lockdown, "variable"), " lockdown"); }
if (length(option.trigger) > 1) { tag = paste0(tag, " ", trigger, " trigger"); }
if (length(option.intervention_shift) > 1) { tag = paste0(tag, " ", intervention_shift, " shift"); }
run$dynamics[, run := r];
run$dynamics[, scenario := paste0(names(interventions)[i], tag)];
run$dynamics[, R0 := R0s[r]];
save(run);
totals = add_totals(run, totals);
dynamics = add_dynamics(run, dynamics, iv);
rm(run)
gc()
}
}
}
}
}
}
cm_save(totals, paste0(covid_uk_path, analysis, "-totals.qs"));
cm_save(dynamics, paste0(covid_uk_path, analysis, "-dynamics.qs"));
print(Sys.time())