-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathbpamz_results.R
2359 lines (1997 loc) · 137 KB
/
bpamz_results.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
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
<<<<<<< HEAD
setwd("C:/Users/Administrator/OneDrive - Johns Hopkins University/Research/universal regimen/universal")
date <- "20190226"
setting <- "SEA"
require(plyr)
source("bpamz_cohort.R")
source("bpamz_result_utils.R")
load(file = paste0("impact.",date,".Rdata"))
setting <- "SEA"
load(paste0("cohortfreqs.",date,".Rdata"))
c <- runcohort
# make bpamz outcome table
tableSreg <- function(impact, c, save)
{
# # proportion cured after 1st and 2nd round:
l4 <- loutcomeboot(individualoutcomefunction = step4812cure, simoutput = impact, c = c)
if(save) save(l4, file=paste0("l4.",date,".Rdata"))
# load(file=paste0("l4.",date,".Rdata"))
# cured first round: overall, rif-s, rif-r, and fq-r:
cure1 <- rbind(
unlist(lapply(l4, function(y) round(c(mean(apply(y[,1,],2,mean)),
sd(apply(y[,1,], 2, mean))), 3))),
unlist(lapply(l4, function(y) round(c(mean(apply(y[,1,]*(y[,which(patientvars=="RIF")+3,]==0),2,sum)/apply(y[,which(patientvars=="RIF")+3,]==0,2,sum)),
sd(apply(y[,1,]*(y[,which(patientvars=="RIF")+3,]==0),2,sum)/apply(y[,which(patientvars=="RIF")+3,]==0,2,sum))), 3))),
unlist(lapply(l4, function(y) round(c(mean(apply(y[,1,]*(y[,which(patientvars=="RIF")+3,]==1),2,sum)/apply(y[,which(patientvars=="RIF")+3,]==1,2,sum)),
sd(apply(y[,1,]*(y[,which(patientvars=="RIF")+3,]==1),2,sum)/apply(y[,which(patientvars=="RIF")+3,]==1,2,sum))), 3)) ),
unlist(lapply(l4, function(y) round(c(mean(apply(y[,1,]*(y[,which(patientvars=="MOXI")+3,]==1),2,sum)/apply(y[,which(patientvars=="MOXI")+3,]==1,2,sum)),
sd(apply(y[,1,]*(y[,which(patientvars=="MOXI")+3,]==1),2,sum)/apply(y[,which(patientvars=="MOXI")+3,]==1,2,sum))), 3))) )
rownames(cure1) <- c("All", "RIF-S", "RIF-R", "MOXI-R")
cure1
# cured second round: overall, rif-s, rif-r, and fq-r:
cure2 <- rbind(
unlist(lapply(l4, function(y) round(c(mean(apply(y[,2,],2,mean)),
sd(apply(y[,2,], 2, mean))), 3))),
unlist(lapply(l4, function(y) round(c(mean(apply(y[,2,]*(y[,which(patientvars=="RIF")+3,]==0),2,sum)/apply(y[,which(patientvars=="RIF")+3,]==0,2,sum)),
sd(apply(y[,2,]*(y[,which(patientvars=="RIF")+3,]==0),2,sum)/apply(y[,which(patientvars=="RIF")+3,]==0,2,sum))), 3))),
unlist(lapply(l4, function(y) round(c(mean(apply(y[,2,]*(y[,which(patientvars=="RIF")+3,]==1),2,sum)/apply(y[,which(patientvars=="RIF")+3,]==1,2,sum)),
sd(apply(y[,2,]*(y[,which(patientvars=="RIF")+3,]==1),2,sum)/apply(y[,which(patientvars=="RIF")+3,]==1,2,sum))), 3)) ),
unlist(lapply(l4, function(y) round(c(mean(apply(y[,2,]*(y[,which(patientvars=="MOXI")+3,]==1),2,sum)/apply(y[,which(patientvars=="MOXI")+3,]==1,2,sum)),
sd(apply(y[,2,]*(y[,which(patientvars=="MOXI")+3,]==1),2,sum)/apply(y[,which(patientvars=="MOXI")+3,]==1,2,sum))), 3))) )
rownames(cure2) <- c("All", "RIF-S", "RIF-R", "MOXI-R")
cure2
#cured by N=36 months
# lc30 <- loutcomeboot(individualoutcomefunction = still.in.state, states=7, t=30, simoutput=impact, c=c)
lc36 <- loutcomeboot(individualoutcomefunction = still.in.state, states=7, t=36, simoutput=impact, c=c)
if(save) save(lc36, file=paste0("lc36.",date,".Rdata"))
# load(paste0("lc36.",date,".Rdata"))
cure36m <- rbind(
unlist(lapply(lc36, function(y) round(c(mean(apply(y[,1,],2,mean)),
sd(apply(y[,1,],2,mean))), 3))),
unlist(lapply(lc36, function(y) round(c(mean(apply(y[,1,]*(y[,which(patientvars=="RIF")+2,]==0),2,sum)/apply(y[,which(patientvars=="RIF")+2,]==0,2,sum)),
sd(apply(y[,1,]*(y[,which(patientvars=="RIF")+2,]==0),2,sum)/apply(y[,which(patientvars=="RIF")+2,]==0,2,sum))), 3))),
unlist(lapply(lc36, function(y) round(c(mean(apply(y[,1,]*(y[,which(patientvars=="RIF")+2,]==1),2,sum)/apply(y[,which(patientvars=="RIF")+2,]==1,2,sum)),
sd(apply(y[,1,]*(y[,which(patientvars=="RIF")+2,]==1),2,sum)/apply(y[,which(patientvars=="RIF")+2,]==1,2,sum))), 3))),
unlist(lapply(lc36, function(y) round(c(mean(apply(y[,1,]*(y[,which(patientvars=="MOXI")+2,]==1),2,sum)/apply(y[,which(patientvars=="MOXI")+2,]==1,2,sum)),
sd(apply(y[,1,]*(y[,which(patientvars=="MOXI")+2,]==1),2,sum)/apply(y[,which(patientvars=="MOXI")+2,]==1,2,sum))), 3))) )
rownames(cure36m) <- c("All", "RIF-S", "RIF-R", "MOXI-R")
cure36m
# cure first round if treated
lsuccess1 <- loutcomeboot(individualoutcomefunction = treatmentsuccess, simoutput=impact, c=c)
if(save) save(lsuccess1, file=paste0("lsuccess1.",date,".Rdata"))
success <- rbind(
unlist(lapply(lsuccess1, function(y) round(c(mean(apply(y[,2,],2,sum)/apply(y[,1,],2,sum)),
sd(apply(y[,2,],2,sum)/apply(y[,1,],2,sum))), 3))),
unlist(lapply(lsuccess1, function(y) round(c(mean(apply(y[,2,]*(y[,3,]==0),2,sum)/apply(y[,1,]*(y[,3,]==0),2,sum)),
sd(apply(y[,2,]*(y[,3,]==0),2,sum)/apply(y[,1,]*(y[,3,]==0),2,sum))), 3))),
unlist(lapply(lsuccess1, function(y) round(c(mean(apply(y[,2,]*(y[,3,]==1),2,sum)/apply(y[,1,]*(y[,3,]==1),2,sum)),
sd(apply(y[,2,]*(y[,3,]==1),2,sum)/apply(y[,1,]*(y[,3,]==1),2,sum))), 3))),
unlist(lapply(lsuccess1, function(y) round(c(mean(apply(y[,2,]*(y[,4,]==1),2,sum)/apply(y[,1,]*(y[,4,]==1),2,sum)),
sd(apply(y[,2,]*(y[,4,]==1),2,sum)/apply(y[,1,]*(y[,4,]==1),2,sum))), 3))) )
rownames(success) <- c("All", "RIF-S", "RIF-R", "MOXI-R")
success
# total TB/infectious time:
li <- loutcomeboot(individualoutcomefunction =time.in.state, states=c(1,2,5,9,10), cutofftime=12*20, carryforward=T,
simoutput=impact, c=c)
if(save) save(li, file=paste0("li.",date,".Rdata"))
# load(file=paste0("li.",date,".Rdata"))
tbtime <- rbind(
unlist(lapply(li, function(x) round(c(mean(apply(x[,1,],2,mean)), sd(apply(x[,1,],2,mean))),3))),
unlist(lapply(li, function(y) round(c(mean(apply(y[,1,]*(y[,which(patientvars=="RIF")+2,]==0),2,sum)/apply(y[,which(patientvars=="RIF")+2,]==0,2,sum)),
sd(apply(y[,1,]*(y[,which(patientvars=="RIF")+2,]==0),2,sum)/apply(y[,which(patientvars=="RIF")+2,]==0,2,sum))),3))),
unlist(lapply(li, function(y) round(c(mean(apply(y[,1,]*(y[,which(patientvars=="RIF")+2,]==1),2,sum)/apply(y[,which(patientvars=="RIF")+2,]==1,2,sum)),
sd(apply(y[,1,]*(y[,which(patientvars=="RIF")+2,]==1),2,sum)/apply(y[,which(patientvars=="RIF")+2,]==1,2,sum))),3))),
unlist(lapply(li, function(y) round(c(mean(apply(y[,1,]*(y[,which(patientvars=="MOXI")+2,]==1),2,sum)/apply(y[,which(patientvars=="MOXI")+2,]==1,2,sum)),
sd(apply(y[,1,]*(y[,which(patientvars=="MOXI")+2,]==1),2,sum)/apply(y[,which(patientvars=="MOXI")+2,]==1,2,sum))),3)))
)
rownames(tbtime) <- c("All", "RIF-S", "RIF-R", "MOXI-R")
tbtime
# time on treatment (of some kind)
followyears <- 20
ltall <- loutcomeboot(individualoutcomefunction =time.in.state, states=c(statetypes$treating, statetypes$treating_adr), cutofftime=12*followyears, carryforward=T,
simoutput=impact, c=c)
if(save) save(ltall, file=paste0("ltall.",date,".Rdata"))
# load(file=paste0("ltall.",date,".Rdata"))
rxtime <- rbind(
unlist(lapply(ltall, function(x) round(c(mean(apply(x[,1,],2,mean)), sd(apply(x[,1,],2,mean))),3))),
unlist(lapply(ltall, function(y) round(c(mean(apply(y[,1,]*(y[,which(patientvars=="RIF")+2,]==0),2,sum)/apply(y[,which(patientvars=="RIF")+2,]==0,2,sum)),
sd(apply(y[,1,]*(y[,which(patientvars=="RIF")+2,]==0),2,sum)/apply(y[,which(patientvars=="RIF")+2,]==0,2,sum))),3))),
unlist(lapply(ltall, function(y) round(c(mean(apply(y[,1,]*(y[,which(patientvars=="RIF")+2,]==1),2,sum)/apply(y[,which(patientvars=="RIF")+2,]==1,2,sum)),
sd(apply(y[,1,]*(y[,which(patientvars=="RIF")+2,]==1),2,sum)/apply(y[,which(patientvars=="RIF")+2,]==1,2,sum))),3))),
unlist(lapply(ltall, function(y) round(c(mean(apply(y[,1,]*(y[,which(patientvars=="MOXI")+2,]==1),2,sum)/apply(y[,which(patientvars=="MOXI")+2,]==1,2,sum)),
sd(apply(y[,1,]*(y[,which(patientvars=="MOXI")+2,]==1),2,sum)/apply(y[,which(patientvars=="MOXI")+2,]==1,2,sum))),3)))
)
rownames(rxtime) <- c("All", "RIF-S", "RIF-R", "MOXI-R")
rxtime
return(rbind(cure1, cure2, cure36m, success, tbtime, rxtime))
}
# make table showing outcomes, summary of full generation and recurrent matrix and adr matrix which will be included in appendix.
rec <- make.recurrence.matrix()[c("HR(ZE)","R(ZE)","BPaMZ","BPamZ","BPaM","BPaZ", "BPam","BPa","MDR, FQ-S", "MDR, FQ-R"),c("4","6","18")]
# (rec <- round(rbind((rec[1,]*sum(c$Freq*(1-c$RIF)*(1-c$INH))+rec[2,]*sum(c$Freq*(1-c$RIF)*c$INH))/sum(c$Freq*(1-c$RIF)),
# rec[3:nrow(rec),]),3)
# )
# to check: all displayed entries should be less than this, else need a min of the two:
round(rowSums(make.adr.matrix()[c("HR(ZE)","R(ZE)","BPaMZ","BPamZ","BPaM","BPaZ", "BPam","BPa","MDR, FQ-S", "MDR, FQ-R"),], na.rm=T)*100, 1)
write.table(round(rec,3), "recurrencetable.csv", sep=",", row.names = TRUE, col.names = TRUE)
# details for supplement:
write.table(rbind(unlist(set.scenario("0")), unlist(set.scenario("1a")), unlist(set.scenario("1x")),unlist(set.scenario("3")),
unlist(set.scenario("3x")),unlist(set.scenario("4x")), unlist(set.scenario("5x"))), file="scenarios.csv", sep=",")
drugs
make.active.regimen.matrix()[,1:(length(drugs)+length(regimens))] # describe, and refer to published code
a <- make.recurrence.matrix(); a[c(1:4,8:nrow(a)),c("9","12","18")] <- NA; round(a,3)
write.table(a, file="recurrencematrix.csv", sep=",")
a <- make.adr.matrix(); a[a==0]<-NA; round(a,3)
write.table(a, file="adrmatrix.csv", sep=",")
tSEA <- tableSreg(impact=impact, c=runcohort, save=T)
save(tSEA, file=paste0("tSEA.",date,".Rdata"))
write.table(tSEA, file=paste0("tSEA.",date,".csv"), sep = ",")
tSEA[1:16,] <- round(tSEA[1:16,]*100,1)
tSEA[17:24,] <- round(tSEA[17:24,],2)
tSEA2 <- cbind(paste0(tSEA[,1], " \u00b1 ",tSEA[,2]), paste0(tSEA[,3], " \u00b1 ",tSEA[,4]), paste0(tSEA[,5], " \u00b1 ",tSEA[,6]), paste0(tSEA[,7], " \u00b1 ",tSEA[,8]))
Encoding(tSEA2)<-"UTF-8"
tSEA2[1:16,] <- paste0(tSEA2[1:16,],"%")
tSEA2
write.table(tSEA2, file=paste0("tSEA2_success.",date,".csv"), sep = ",")
save(tSEA2, file=paste0("tSEA2.",date,".Rdata"))
# additional results for text
# % with initial regimen as hrze for rif-r in expanded xpert:
dxrr <- function(patiententry) {c(patiententry["Currentregimen",3]==4,patiententry["Currentregimen",3]>0,patiententry["RIF",1])} # treated rifr, treated at all, has rifr
lreg <- loutcomeboot(individualoutcomefunction = dxrr, simoutput = list("novelrr"=impact$novelrr, "novelrrx"=impact$novelrrx), c = c)
save(lreg, file=paste0("lreg.",date,".Rdata"))
# load("lreg.20190226.Rdata")
lapply(lreg, function(y) mean(apply(y[,1,]*y[,2,]*y[,3,],2,sum)/apply(y[,2,]*y[,3,],2,sum)))
# proportion of overall cohort cured in 3 rounds:
load(paste0("l4.",date,".Rdata"))
lapply(l4, function(y) mean(apply(y[,3,],2,mean)))
lapply(l4, function(y) sd(apply(y[,3,],2,mean)))
rm(l4)
# % reduction in time with TB
load("li.20190226.Rdata")
lapply(li, function(x) apply(x[,1,], 2, mean))
lapply(li, function(x) mean(apply(x[,1,]*(x[,which(patientvars=="RIF")+2,]==1), 2, sum)/apply((x[,which(patientvars=="RIF")+2,]==1), 2, sum)))
lapply(li, function(x) sd(apply(x[,1,]*(x[,which(patientvars=="RIF")+2,]==1), 2, sum)/apply((x[,which(patientvars=="RIF")+2,]==1), 2, sum)))
reduction <- lapply(li, function(x) apply((li$baseline[,1,]-x[,1,]), 2, sum)/apply(li$baseline[,1,],2,sum))
lapply(reduction, mean); lapply(reduction, sd)
rrreduction <- lapply(li, function(x) (apply(li$baseline[,1,]*(li$baseline[,which(patientvars=="RIF")+2,]==1),2,sum)-
apply(x[,1,]*(x[,which(patientvars=="RIF")+2,]==1), 2, sum))/
apply(li$baseline[,1,]*(li$baseline[,which(patientvars=="RIF")+2,]==1),2,sum))
univrrreduction <- lapply(li, function(x) (apply(li$novelrr[,1,]*(li$novelrr[,which(patientvars=="RIF")+2,]==1),2,sum)-
apply(x[,1,]*(x[,which(patientvars=="RIF")+2,]==1), 2, sum))/
apply(li$novelrr[,1,]*(li$novelrr[,which(patientvars=="RIF")+2,]==1),2,sum))
rrreduction
lapply(rrreduction, mean); lapply(rrreduction, sd)
lapply(univrrreduction, mean); lapply(univrrreduction, sd)
rm(li)
# for supplement, South Africa
minfreq <- 1/1e8
SAfc <- SAfcohort[cohort$Freq>minfreq|SAfcohort$Freq>minfreq,]
nrow(SAfc)
## and now run the above to amke new table (without saving)
tSAf <- tableSreg(impact=impact, c=SAfc, save=F)
# save(tSAf, file=paste0("tSAf.",date,".Rdata"))
write.table(tSAf, file=paste0("tSAf.",date,".csv"), sep = ",")
save(tSAf, file=paste0("tSAf.",date,".Rdata"))
tSAf[1:16,] <- round(tSAf[1:16,]*100,1)
tSAf[17:24,] <- round(tSAf[17:24,],2)
tSAf2 <- cbind(paste0(tSAf[,1], " \u00b1 ",tSAf[,2]), paste0(tSAf[,3], " \u00b1 ",tSAf[,4]), paste0(tSAf[,5], " \u00b1 ",tSAf[,6]), paste0(tSAf[,7], " \u00b1 ",tSAf[,8]))
tSAf2[1:16,] <- paste0(tSAf2[1:16,],"%")
Encoding(tSAf2)<-"UTF-8"
write.table(tSAf2, file=paste0("tSAf2.",date,".csv"), sep = ",")
# for supplement, same duration for all
load(file=paste0("moreimpact.",date,".Rdata")) # 4 or 6 months for all
tmore <- tableSreg(impact=moreimpact, c=runcohort, save=F)
save(tmore, file=paste0("tmore.",date,".Rdata"))
load(paste0("tSEA.",date,".Rdata"))
tmore <- cbind(tSEA[,7:8], tmore)
write.table(tmore, file=paste0("tmore",date,".csv"), sep = ",")
tmore[1:16,] <- round(tmore[1:16,]*100,1)
tmore[17:24,] <- round(tmore[17:24,],2)
tmore2 <- cbind(paste0(tmore[,1], " \u00b1 ",tmore[,2]), paste0(tmore[,3], " \u00b1 ",tmore[,4]), paste0(tmore[,5], " \u00b1 ",tmore[,6]))
tmore2[1:16,] <- paste0(tmore2[1:16,],"%")
Encoding(tmore2)<-"UTF-8"
write.table(tmore2, file=paste0("tmore2.",date,".csv"), sep = ",")
save(tmore2, file=paste0("tmore2.",date,".Rdata"))
rm(moreimpact)
# for sensitivity analysis, change prevalence of moxi/PZA resistance:
# Higher FQ R setting -- will be putting in supplement for part 1, but including in Fig for part 2 (and in text for part 3?)
params2 <- params
highparams <- as.numeric(allparams[,"UR.high"])
names(highparams) <- allparams[,1]
# use high end estimates of FQ-R for SEA
params2["MOXI-R-any_in_RIF-S"] <- highparams["MOXI-R-any_in_RIF-S"]
params2["MOXI-R-highlevel_in_RIF-S"] <- highparams["MOXI-R-highlevel_in_RIF-S"]
highFQScohort <- cohort.probs(params = params2, patientvars = patientvars)
# will use the same "impact" and "dst", but this different highFQcohort, restricted to same patient types as base model since those are the ones we ran.
highFQScohort <- highFQScohort[cohort$Freq>minfreq|SAfcohort$Freq>minfreq,]
# only care about some of the regimen impact outcomes:
#(and also, don't want to emphasize the hgher cure in rif-r due to more prior treatment and faster rediagnosis)
# l4 (proportion cured in 1 and 2 rounds), and li
# overall cohort, pantb,
# for higher RS MOXI-R, higher RS and RR MOXI-R, and higher MOXI-R and PZA+R in MOXI-R
# so need three new cohorts:
# highFQScohort as defined above, and
params2["MOXI-R-any_in_RIF-R"] <- highparams["MOXI-R-any_in_RIF-R"]
params2["MOXI-R-highlevel_in_RIF-R"] <- highparams["MOXI-R-highlevel_in_RIF-R"]
highFQRcohort <- cohort.probs(params = params2, patientvars = patientvars)
highFQRcohort <- highFQRcohort[cohort$Freq>minfreq|SAfcohort$Freq>minfreq,]
params2["PZA-R_in_RIF-R"] <- highparams["PZA-R_in_RIF-R"]
highFQZRcohort <- cohort.probs(params = params2, patientvars = patientvars)
highFQZRcohort <- highFQZRcohort[cohort$Freq>minfreq|SAfcohort$Freq>minfreq,]
regdiffs <- function(impact, c)
{
l4 <- loutcomeboot(individualoutcomefunction = step4812cure, simoutput = impact, c = c)
# cured first round overall:
cure1 <-t(array(unlist(lapply(l4, function(y) round(c(mean(apply(y[,1,],2,mean)),
sd(apply(y[,1,], 2, mean))), 3)))[c(1:2,7:8)],dim=c(2,2)))
cure1diff <- cbind(round(c(mean(apply(l4$novelpantb[,1,],2,mean) - apply(l4$baseline[,1,],2,mean)),
sd(apply(l4$novelpantb[,1,],2,mean) - apply(l4$baseline[,1,],2,mean))), 3),
round(c(mean((apply(l4$novelpantb[,1,],2,mean) - apply(l4$baseline[,1,],2,mean))/apply(l4$baseline[,1,],2,mean)),
sd((apply(l4$novelpantb[,1,],2,mean) - apply(l4$baseline[,1,],2,mean))/apply(l4$baseline[,1,],2,mean))), 3))
li <- loutcomeboot(individualoutcomefunction =time.in.state, states=c(1,2,5,9,10), cutofftime=12*20, carryforward=T,
simoutput=impact, c=c)
tbtime <- t(array(unlist(lapply(li, function(x) round(c(mean(apply(x[,1,],2,mean)), sd(apply(x[,1,],2,mean))),3)))[c(1:2,7:8)],dim=c(2,2)))
tbtimediff <- cbind(round(c(mean(apply(li$novelpantb[,1,],2,mean) - apply(li$baseline[,1,],2,mean)),
sd(apply(li$novelpantb[,1,],2,mean) - apply(li$baseline[,1,],2,mean))), 3),
round(c(mean((apply(li$novelpantb[,1,],2,mean) - apply(li$baseline[,1,],2,mean))/apply(li$baseline[,1,],2,mean)),
sd((apply(li$novelpantb[,1,],2,mean) - apply(li$baseline[,1,],2,mean))/apply(li$baseline[,1,],2,mean))), 3) )
tbtimerr <- t(array(unlist(lapply(li, function(x) round(c(mean(apply(y[,1,]*(y[,which(patientvars=="RIF")+2,]==1),2,sum)/apply(y[,which(patientvars=="RIF")+2,]==1,2,sum)),
sd(apply(y[,1,]*(y[,which(patientvars=="RIF")+2,]==1),2,sum)/apply(y[,which(patientvars=="RIF")+2,]==1,2,sum))),3)))[c(1:2,7:8)],dim=c(2,2)))
tbtimediffrr <- cbind(round(c(mean(apply(li$novelpantb[,1,]*(li$novelpantb[,which(patientvars=="RIF")+2,]==1),2,sum)/apply(li$novelpantb[,which(patientvars=="RIF")+2,]==1,2,sum) -
apply(li$novelrrx[,1,]*(li$novelrrx[,which(patientvars=="RIF")+2,]==1),2,sum)/apply(li$novelrrx[,which(patientvars=="RIF")+2,]==1,2,sum)),
sd(apply(li$novelpantb[,1,]*(li$novelpantb[,which(patientvars=="RIF")+2,]==1),2,sum)/apply(li$novelpantb[,which(patientvars=="RIF")+2,]==1,2,sum) -
apply(li$novelrrx[,1,]*(li$novelrrx[,which(patientvars=="RIF")+2,]==1),2,sum)/apply(li$novelrrx[,which(patientvars=="RIF")+2,]==1,2,sum))), 3))
return(rbind(cure1, cure1diff, tbtime, tbtimediff , tbtimerr, tbtimediffrr)[c(1,5,2,6,3,4,7,8, 9,10),])
}
tresist <- cbind(regdiffs(impact, c), regdiffs(impact, highFQScohort), regdiffs(impact, highFQRcohort), regdiffs(impact, highFQZRcohort))
write.table(tresist, file=paste0("tresist",date,".csv"), sep = ",")
save(tresist, file=paste0("tresist.",date,".Rdata"))
tresist[c(1,3,5,6,8),] <- round(tresist[c(1,3,5,6,8),]*100,1)
tresist[c(2,4,7),] <- round(tresist[c(2,4,7),],2)
tresist2 <- cbind(paste0(tresist[,1], " \u00b1 ",tresist[,2]), paste0(tresist[,3], " \u00b1 ",tresist[,4]), paste0(tresist[,5], " \u00b1 ",tresist[,6]), paste0(tresist[,7], " \u00b1 ",tresist[,8]))
Encoding(tresist2)<-"UTF-8"
tresist2[c(1,3,5,6,8),] <- paste0(tresist2[c(1,3,5,6,8),],"%")
write.table(tresist2, file=paste0("tresist2",date,".csv"), sep = ",")
#
# for multiple scenarios, bar plots, :
yl <- loutcomeboot(function(x) x[c("eventtime","TBstate","RIF"),], impact, c, desiredsize=1e4)
zl <- lapply(yl, function(y)
abind(y[,seq(1,52,by=3),], y[,seq(2,53,by=3),], y[,seq(3,54,by=3),],
along = 4, new.names = list("patient"=1:1e4, "step"=1:18, "copy"=1:50, "a"=c("time","state","RIF"))))
zlp <- lapply(zl, function(z) aperm(z, perm = c(1,3,2,4)))
rm(zl)
save(yl, zlp, file=paste0("fig2statestatus.",date,".Rdata"))
rm(yl)
maxtime <- 30; times <- seq(0,maxtime,by=0.5)
statestatus <- function(z1, times=0:maxtime,rifs = c(1,0))
{ return(t(array(do.call('rbind', lapply(times, function(t)
tabulate(unlist( z1[,,"state"][
cbind(1:dim(z1)[[1]], c(apply(z1[,,"time"], c(1), function(x) ifelse(max(x)>t, which.max(x>t)-1, length(x)))))][z1[,1,"RIF"] %in% rifs]), nbins = length(statetypes))
)), dim=c(length(times), length(statetypes)), dimnames=list("time"=times, "state"=names(statetypes)))))
}
statestatusarray <- function(z2, times=0:30, rifs=c(0,1))
{
aaply(z2, 2, function(z) statestatus(z,times, rifs))
}
fullstatenames <- c(" \nTB not yet diagnosed\n ", "[shouldn't show up]", "On treatment", " \nOn treatment &\nacquired resistance",
"Treatment failed", "Will relapse", "Cured", "Deceased", "Relapsed TB","Lost before (re)treatment")
rbind(unlist(statetypes), fullstatenames)
plotorder <- c(1,10,5,9,6,3,4,7,8)
pdf("Fig2.pdf", width=12, height=8)
layout(matrix(c(1,3,2,4,5,5), ncol=3), widths = c(5,5,2))
par(mar=c(3,3,1,5), oma=c(2,2,2,2))
require(RColorBrewer)
colors <- c("gray", palette(brewer.pal(8, name = "Accent")), "black")
colors <- c("gray", palette(brewer.pal(8, name = "Accent")), "black")
downshift <- 0.5
mapply( function(z, title)
{
meanstates <- apply(statestatusarray(z[,1:5,,], times, rifs=1), c(2,3), mean)
b <- barplot(xlab = "", ylab="", yaxt='n', xaxt='n', border = NA, space=0,
meanstates[plotorder,],
col=colors, beside=F)
mtext(title, side=3, font=2,line=0.5, cex=1)
axis(side = 2, at = seq(0,sum(z[,1:5,1,"RIF"])/5,length=6),labels = seq(0,100,by=20), cex.axis=0.8, las=2)
axis(side = 1, at = b[seq(1,length(b),length.out = maxtime/3+1)],labels = seq(0,maxtime,by=3), cex.axis=0.8)
label <- meanstates[plotorder,ncol(meanstates)]/sum(z[,1,1,"RIF"]==1)*100 > 2
# text(x = (b[length(b)]+c(0,3*ncol(meanstates)/maxtime,rep(0,sum(label)-2))),
text(x = rep(b[length(b)], sum(label)),
y = (cumsum(meanstates[plotorder,ncol(meanstates)])- meanstates[plotorder,ncol(meanstates)]/2)[label]-downshift,pos=4, xpd=NA,
labels = paste0(round(meanstates[plotorder,ncol(meanstates)]/sum(z[,1,1,"RIF"]==1)*100,1)[label], "%"), cex=1)
# segments(b[length(b)],(cumsum(meanstates[plotorder,ncol(meanstates)])[2]- meanstates[plotorder[2],ncol(meanstates)]/2) -1,
# b[length(b)]+3*ncol(meanstates)/maxtime,(cumsum(meanstates[plotorder,ncol(meanstates)])[2]- meanstates[plotorder[2],ncol(meanstates)]/2) -1, xpd=NA)
},
zlp, c("Current practice", "Novel regimen for RIF-R", "Novel RIF-R regimen + expanded RIF DST", "Novel regimen for all"))
mtext("Months elapsed since TB onset", 1, outer=T, cex=1, adj = .35)
mtext("% of RIF-R TB cohort", 2, outer=T, cex=1, line=-0.5)
# mtext("Status of RIF-R cohort over time", side=3, font=2, outer=T, line=0.5)
plot.new()
legend("center", legend = rev(fullstatenames[plotorder]), fill=rev(colors[1:9]), cex=1.2, xpd=NA,title = expression(bold("Status")))
dev.off()
# sensitivity analysis: delays
rm(impact)
load(file=paste0("delays.",date,".Rdata"))
li <- loutcomeboot(individualoutcomefunction =time.in.state, states=c(1,2,5,9,10), cutofftime=12*20, carryforward=T,
simoutput=delays, c=c)
tbtime <- t(array(unlist(lapply(li, function(x) round(c(mean(apply(x[,1,],2,mean)), sd(apply(x[,1,],2,mean))),3)))[c(1:2,7:8)],dim=c(2,2)))
tbtimediff <- cbind(round(c(mean(apply(li$novelpantb[,1,],2,mean) - apply(li$novelrrx[,1,],2,mean)),
sd(apply(li$novelpantb[,1,],2,mean) - apply(li$novelrrx[,1,],2,mean))), 3),
round(c(mean((apply(li$novelpantb[,1,],2,mean) - apply(li$novelrrx[,1,],2,mean))/apply(li$novelrrx[,1,],2,mean)),
sd((apply(li$novelpantb[,1,],2,mean) - apply(li$novelrrx[,1,],2,mean))/apply(li$novelrrx[,1,],2,mean))), 3) )
tbtimerr <- t(array(unlist(lapply(li, function(y) round(c(mean(apply(y[,1,]*(y[,which(patientvars=="RIF")+2,]==1),2,sum)/apply(y[,which(patientvars=="RIF")+2,]==1,2,sum)),
sd(apply(y[,1,]*(y[,which(patientvars=="RIF")+2,]==1),2,sum)/apply(y[,which(patientvars=="RIF")+2,]==1,2,sum))),3)))[c(1:2,7:8)],dim=c(2,2)))
tbtimediffrr <- cbind(round(c(mean((apply(li$novelpantb[,1,]*(li$novelpantb[,which(patientvars=="RIF")+2,]==1),2,sum) -
apply(li$novelrrx[,1,]*(li$novelrrx[,which(patientvars=="RIF")+2,]==1),2,sum))/
apply(li$novelrrx[,1,]*(li$novelrrx[,which(patientvars=="RIF")+2,]==1),2,sum)),
sd((apply(li$novelpantb[,1,]*(li$novelpantb[,which(patientvars=="RIF")+2,]==1),2,sum) -
apply(li$novelrrx[,1,]*(li$novelrrx[,which(patientvars=="RIF")+2,]==1),2,sum))/
apply(li$novelrrx[,1,]*(li$novelrrx[,which(patientvars=="RIF")+2,]==1),2,sum)) ), 3))
save(tbtime, tbtimediff, tbtimerr, tbtimediffrr, file=paste0("delayresults.",data,".Rdata"))
rm(delays)
# sensitivity analysis: BPAMZ efficacy
load(paste0("lowerefficacy.",date,".Rdata"))
tloeff <- tableSreg(impact=lowerefficacy, c=c, save=F)
write.table(tloeff, file=paste0("tloeff.",date,".csv"), sep = ",")
save(tloeff, file=paste0("tloeff.",date,".Rdata"))
tloeff[1:16,] <- round(tloeff[1:16,]*100,1)
tloeff[17:24,] <- round(tloeff[17:24,],2)
tloeff2 <- cbind(paste0(tloeff[,1], " \u00b1 ",tloeff[,2])) #, paste0(tloeff[,3], " \u00b1 ",tloeff[,4]), paste0(tloeff[,5], " \u00b1 ",tloeff[,6]))
tloeff2[1:12,] <- paste0(tloeff2[1:12,],"%")
load(paste0("tSEA2.",date,".Rdata"))
tloeff2 <- cbind(tSEA2[,4],tloeff2)
Encoding(tloeff2)<-"UTF-8"
write.table(tloeff2, file=paste0("tloeff2.",date,".csv"), sep = ",")
rm(lowerefficacy)
#############################################
# outcomes for moxi resistant and Xpert XDR
# let's tally total unsuccessful treatments (fail + relapse + death during treatment),
# and tally infectious years (starting from TB onset),
# for each of FQ-R RR and FQ-R RS (each weighted according to their prevalence in an overall TB cohort of 1e5),
# for each of FQ DST (simultaneous) and no,
# and for SEA plus a hypothetical setting with FQ-R in RS on the high end based on pakistan and Mumbai data (Assuming same prev PZA-R).
#( for this last, just need to define a new freq vector, order it like the original one, and use the same courses otherwise)
# and in the text, will report the increase in cure rates amogn RIF-R versus __ in RIF-S when FQ DST is added,
# but that given RIF R only comprise __% of the overall cohort, the overall cure rate increases from __ to __ with stepwise and __ with simult.
# The number of Xpert XDR tests per additional cure is __ for stepwise and __ for simult.
# Also give all results for higher prev setting.
# moxi-r fractions
sum(c$Freq[c$MOXI==1])/sum(c$Freq)
sum(c$Freq[c$MOXI==1])/sum(c$Freq)*1e5
sum(c$Freq[c$MOXI==1&c$RIF==1])/sum(c$Freq)
sum(c$Freq[c$MOXI==1&c$RIF==0])/sum(c$Freq)
sum(c$Freq[c$MOXI==1&c$RIF==1])/sum(c$Freq[c$RIF==1])
sum(c$Freq[c$MOXI==1&c$RIF==0])/sum(c$Freq[c$RIF==0])
sum(c$Freq[c$MOXI==1&c$RIF==1])/sum(c$Freq[c$MOXI==0&c$RIF==1])/
sum(c$Freq[c$MOXI==1&c$RIF==0])/sum(c$Freq[c$MOXI==0&c$RIF==0])
sum(c$Freq[c$PZA==1&c$RIF==1])/sum(c$Freq[c$PZA==0&c$RIF==1])/
sum(c$Freq[c$PZA==1&c$RIF==0])/sum(c$Freq[c$PZA==0&c$RIF==0])
sum(c$Freq[c$MOXI==1&c$RIF==1&c$PZA==0&c$partialmoxi==1])/sum(c$Freq[c$MOXI==1&c$RIF==1])
sum(c$Freq[c$MOXI==1&c$RIF==1&c$PZA==1&c$partialmoxi==1])/sum(c$Freq[c$MOXI==1&c$RIF==1])
# MOXI-R do worse:
load("lsuccess1.",date,".Rdata")
unlist(lapply(lsuccess1, function(y) round(c(mean(apply(y[,2,]*(y[,3,]==1)*(y[,4,]==1),2,sum)/apply(y[,1,]*(y[,3,]==1)*(y[,4,]==1),2,sum)),
sd(apply(y[,2,]*(y[,3,]==1)*(y[,4,]==1),2,sum)/apply(y[,1,]*(y[,3,]==1)*(y[,4,]==1),2,sum))), 3)))
unlist(lapply(lsuccess1, function(y) round(c(mean(apply(y[,2,]*(y[,3,]==0)*(y[,4,]==1),2,sum)/apply(y[,1,]*(y[,3,]==0)*(y[,4,]==1),2,sum)),
sd(apply(y[,2,]*(y[,3,]==0)*(y[,4,]==1),2,sum)/apply(y[,1,]*(y[,3,]==0)*(y[,4,]==1),2,sum))), 3)))
rm(l4)
# outcomes with DST approaches:
load(paste0("dst.",date,".Rdata"))
# cures with fq dst
l4dst <- loutcomeboot(individualoutcomefunction = step4812cure, simoutput = dst, c = c)
save(l4dst, file=paste0("l4dst.",date,".Rdata"))
dstcure1RS <- lapply(l4dst, function(y) apply(y[,1,]*(y[,which(patientvars=="RIF")+3,]==0)*(y[,which(patientvars=="MOXI")+3,]==1),2,sum)/apply((y[,which(patientvars=="RIF")+3,]==0)*(y[,which(patientvars=="MOXI")+3,]==1),2,sum))
dstcure1RR <- lapply(l4dst, function(y) apply(y[,1,]*(y[,which(patientvars=="RIF")+3,]==1)*(y[,which(patientvars=="MOXI")+3,]==1),2,sum)/apply((y[,which(patientvars=="RIF")+3,]==1)*(y[,which(patientvars=="MOXI")+3,]==1),2,sum))
dstcure1all <- lapply(l4dst, function(y) apply(y[,1,]*(y[,which(patientvars=="MOXI")+3,]==1),2,sum)/apply((y[,which(patientvars=="MOXI")+3,]==1),2,sum))
lapply(dstcure1RS, mean); lapply(dstcure1RS, sd)
lapply(dstcure1RR, mean); lapply(dstcure1RR, sd)
rm(l4dst)
# unsuccessful treatments first round
unsuccessfulrx <- function(patiententry) # without, then with, prerxltfu; and append rif and moxi status, and partialmoxi and pza
{ return(c(patiententry["eventtype",3]==eventtypes$treatmentstart &
patiententry["TBstate",4] %in% c(statetypes$failed, statetypes$pendingrelapse, statetypes$deceased ),
(patiententry["eventtype",2]==eventtypes$TBdiagnosis &
patiententry["TBstate",4] %in% c(statetypes$failed, statetypes$pendingrelapse, statetypes$deceased,
statetypes$pretreatmentlost ) ),
patiententry["RIF",1],
patiententry["MOXI",1],
patiententry["partialmoxi",1],
patiententry["PZA",1]) ) }
ludst <- loutcomeboot(individualoutcomefunction = unsuccessfulrx, simoutput= dst, c = c)
save(ludst, file=paste0("ludst.",date,".Rdata"))
# load(file=paste0("ludst.",date,".Rdata"))
dstu1RS <- lapply(ludst, function(y) apply(y[,1,]*(y[,3,]==0)*(y[,4,]==1),2,sum))
dstu1RR <- lapply(ludst, function(y) apply(y[,1,]*(y[,3,]==1)*(y[,4,]==1),2,sum))
dstu1all <- lapply(ludst, function(y) apply(y[,1,]*(y[,4,]==1),2,sum))
rm(ludst)
lidst <- loutcomeboot(individualoutcomefunction =time.in.state, states=c(1,2,5,9,10), cutofftime=12*20, carryforward=T,
simoutput=dst, c=c)
save(lidst, file=paste0("lidst.",date,".Rdata"))
# load(file=paste0("lidst.",date,".Rdata"))
dsttimeRS <- lapply(lidst, function(y) apply(y[,1,]*(y[,which(patientvars=="RIF")+2,]==0)*(y[,which(patientvars=="MOXI")+2,]==1),2,sum))
dsttimeRR <- lapply(lidst, function(y) apply(y[,1,]*(y[,which(patientvars=="RIF")+2,]==1)*(y[,which(patientvars=="MOXI")+2,]==1),2,sum))
dsttimeall <- lapply(lidst, function(y) apply(y[,1,]*(y[,which(patientvars=="MOXI")+2,]==1),2,sum))
rm(lidst)
# will boxplot these:
# colors/shading for DST none vs step vs sim side by side, larger x axis divisions for rs vs rr vs overall,
# and panels for u and i
require(RColorBrewer)
pdf("Fig3.pdf", width=6, height=10)
layout(c(1,2,3), heights = c(6,6,1))
# par(mfcol=c(2,1))
par(mar=c(4,4,1,1), oma=rep(0.5,4), lwd=0.3, cex=1)
cols <- brewer.pal(3, name = "Accent")
b <- boxplot(list("No FQ DST" = dstu1RS$noxxdr, "FQ DST in RIF-R" = dstu1RS$stepxxdr, "FQ DST for all" = dstu1RS$fullxxdr,
"No FQ DST" = dstu1RR$noxxdr, "FQ DST in RIF-R" = dstu1RR$stepxxdr, "FQ DST for all" = dstu1RR$fullxxdr,
"No FQ DST" = dstu1all$noxxdr, "FQ DST in RIF-R" = dstu1all$stepxxdr, "FQ DST for all" = dstu1all$fullxxdr),
at=c(1,2,3,6,7,8, 11,12,13),
col=cols, notch=FALSE,
ylab = "Unsuccessful treatments of MOXI-R TB",
ylim=c(0,max(unlist(dstu1all))*1.05),
xaxt ='n')
text(x=c(1,2,3, 6,7,8, 11,12,13, 16,17,18), y=b$stats[5,]-max(b$stats[5,])/100, pos=3,
labels = paste0(round(100*b$stats[3,]/rep(c(sum(c$Freq[c$MOXI==1&c$RIF==0]), sum(c$Freq[c$MOXI==1&c$RIF==1]), sum(c$Freq[c$MOXI==1])), each=3)/1e5, 1),"%"), cex=0.8)
axis(1, at = c(2,7,12), labels = c("Among RIF-S,\nMOXI-R TB", "Among RIF-R,\nMOXI-R TB", "Among all\nMOXI-R TB"),
lwd.ticks = F)
legend("topleft", "A", bty="n", cex = 1.5)
b2 <- boxplot(list("No FQ DST" = dsttimeRS$noxxdr, "FQ DST in RIF-R" = dsttimeRS$stepxxdr, "FQ DST for all" = dsttimeRS$fullxxdr,
"No FQ DST" = dsttimeRR$noxxdr, "FQ DST in RIF-R" = dsttimeRR$stepxxdr, "FQ DST for all" = dsttimeRR$fullxxdr,
"No FQ DST" = dsttimeall$noxxdr, "FQ DST in RIF-R" = dsttimeall$stepxxdr, "FQ DST for all" = dsttimeall$fullxxdr),
at=c(1,2,3,6,7,8, 11,12,13),
col=cols, notch=FALSE,
ylab = "Person-months of active MOXI-R TB",
ylim=c(0,max(unlist(dsttimeall))),
xaxt ='n')
axis(1, at = c(2,7,12), labels = c("Among RIF-S,\nMOXI-R TB", "Among RIF-R,\nMOXI-R TB", "Among all\nMOXI-R TB"),
lwd.ticks = F)
legend("topleft", "B", bty="n", cex = 1.5)
par(mar=c(1,1,1,1))
plot.new()
# par(mar=c(1,1,1,1))
legend("center", legend = c("None", "RIF-R", "All"),
pch=15, col=cols, title = expression(bold("FQ DST performed for:")), xpd=NA, ncol=3, pt.cex = 3)
dev.off()
# sensitivity analysis plot, higher moxi and/or pza resistance:
ludsthighFQS <- loutcomeboot(individualoutcomefunction = unsuccessfulrx, simoutput= dst, c = highFQScohort)
dstu1allhighFQS <- lapply(ludsthighFQS, function(y) apply(y[,1,]*(y[,4,]==1),2,sum))
rm(ludsthighFQS)
ludsthighFQR <- loutcomeboot(individualoutcomefunction = unsuccessfulrx, simoutput= dst, c = highFQRcohort)
dstu1allhighFQR <- lapply(ludsthighFQR, function(y) apply(y[,1,]*(y[,4,]==1),2,sum))
rm(ludsthighFQR)
ludsthighFQZR <- loutcomeboot(individualoutcomefunction = unsuccessfulrx, simoutput= dst, c = highFQZRcohort)
dstu1allhighFQZR <- lapply(ludsthighFQZR, function(y) apply(y[,1,]*(y[,4,]==1),2,sum))
rm(ludsthighFQZR)
pdf("Fig3S.pdf", width=11, height=8)
layout(c(1,2), heights = c(6,1))
par(mar=c(4,4,1,1), oma=rep(0.5,4), lwd=0.3, cex=1)
cols <- brewer.pal(3, name = "Accent")
b <- boxplot(list("No FQ DST" = dstu1all$noxxdr, "FQ DST in RIF-R" = dstu1all$stepxxdr, "FQ DST for all" = dstu1all$fullxxdr,
"No FQ DST" = dstu1allhighFQS$noxxdr, "FQ DST in RIF-R" = dstu1allhighFQS$stepxxdr, "FQ DST for all" = dstu1allhighFQS$fullxxdr,
"No FQ DST" = dstu1allhighFQR$noxxdr, "FQ DST in RIF-R" = dstu1allhighFQR$stepxxdr, "FQ DST for all" = dstu1allhighFQR$fullxxdr,
"No FQ DST" = dstu1allhighFQZR$noxxdr, "FQ DST in RIF-R" = dstu1allhighFQZR$stepxxdr, "FQ DST for all" = dstu1allhighFQZR$fullxxdr),
at=c(1,2,3, 5,6,7, 9,10,11, 13,14,15),
col=cols, notch=FALSE,
ylab = "Unsuccessful treatments of MOXI-R TB",
ylim=c(0,max(unlist(dstu1allhighFQZR))),
xaxt ='n')
axis(1, at = c(2,6,10,14), labels = c("With original\nMOXI-R and PZA-R\n\n", "Increased MOXI-R\namong RIF-S\n\n", "Increased MOXI-R\namong RIF-S and RIF-R\n\n", "Increased MOXI-R\namong RIF-S and RIF-R,\nand increased PZA-R\namong RIF-R"),
lwd.ticks = F, padj=0.5)
plot.new()
# par(mar=c(1,1,1,1))
legend("center", legend = c("None", "RIF-R", "All"),
pch=15, col=cols, title = expression(bold("FQ DST performed for:")), xpd=NA, ncol=3, pt.cex = 3)
dev.off()
# Sensitivity alaysis, DST in SAf
load("dst.20190226.Rdata")
ludstSAf <- loutcomeboot(individualoutcomefunction = unsuccessfulrx, simoutput= dst, c = SAfcohort)
save(ludstSAf, file=paste0("ludstSAf.",date,".Rdata"))
# load(file=paste0("ludst.",date,".Rdata"))
dstu1RS <- lapply(ludstSAf, function(y) apply(y[,1,]*(y[,3,]==0)*(y[,4,]==1),2,sum))
dstu1RR <- lapply(ludstSAf, function(y) apply(y[,1,]*(y[,3,]==1)*(y[,4,]==1),2,sum))
dstu1all <- lapply(ludstSAf, function(y) apply(y[,1,]*(y[,4,]==1),2,sum))
rm(ludstSAf)
lidstSAf <- loutcomeboot(individualoutcomefunction =time.in.state, states=c(1,2,5,9,10), cutofftime=12*20, carryforward=T,
simoutput=dst, c=SAfcohort)
save(lidstSAf, file=paste0("lidstSAf.",date,".Rdata"))
# load(file=paste0("lidst.",date,".Rdata"))
dsttimeRS <- lapply(lidstSAf, function(y) apply(y[,1,]*(y[,which(patientvars=="RIF")+2,]==0)*(y[,which(patientvars=="MOXI")+2,]==1),2,sum))
dsttimeRR <- lapply(lidstSAf, function(y) apply(y[,1,]*(y[,which(patientvars=="RIF")+2,]==1)*(y[,which(patientvars=="MOXI")+2,]==1),2,sum))
dsttimeall <- lapply(lidstSAf, function(y) apply(y[,1,]*(y[,which(patientvars=="MOXI")+2,]==1),2,sum))
rm(lidstSAf)
rm(dst)
pdf("Fig3S2.pdf", width=6, height=10)
layout(c(1,2,3), heights = c(6,6,1))
# par(mfcol=c(2,1))
par(mar=c(4,4,1,1), oma=rep(0.5,4), lwd=0.3, cex=1)
cols <- brewer.pal(3, name = "Accent")
b <- boxplot(list("No FQ DST" = dstu1RS$noxxdr, "FQ DST in RIF-R" = dstu1RS$stepxxdr, "FQ DST for all" = dstu1RS$fullxxdr,
"No FQ DST" = dstu1RR$noxxdr, "FQ DST in RIF-R" = dstu1RR$stepxxdr, "FQ DST for all" = dstu1RR$fullxxdr,
"No FQ DST" = dstu1all$noxxdr, "FQ DST in RIF-R" = dstu1all$stepxxdr, "FQ DST for all" = dstu1all$fullxxdr),
at=c(1,2,3,6,7,8, 11,12,13),
col=cols, notch=FALSE,
ylab = "Unsuccessful treatments of MOXI-R TB",
ylim=c(0,max(unlist(dstu1all))*1.05),
xaxt ='n')
text(x=c(1,2,3, 6,7,8, 11,12,13, 16,17,18), y=b$stats[5,]-max(b$stats[5,])/100, pos=3,
labels = paste0(round(100*b$stats[3,]/rep(c(sum(c$Freq[c$MOXI==1&c$RIF==0]), sum(c$Freq[c$MOXI==1&c$RIF==1]), sum(c$Freq[c$MOXI==1])), each=3)/1e5, 1),"%"), cex=0.8)
axis(1, at = c(2,7,12), labels = c("Among RIF-S,\nMOXI-R TB", "Among RIF-R,\nMOXI-R TB", "Among all\nMOXI-R TB"),
lwd.ticks = F)
legend("topleft", "A", bty="n", cex = 1.5)
b2 <- boxplot(list("No FQ DST" = dsttimeRS$noxxdr, "FQ DST in RIF-R" = dsttimeRS$stepxxdr, "FQ DST for all" = dsttimeRS$fullxxdr,
"No FQ DST" = dsttimeRR$noxxdr, "FQ DST in RIF-R" = dsttimeRR$stepxxdr, "FQ DST for all" = dsttimeRR$fullxxdr,
"No FQ DST" = dsttimeall$noxxdr, "FQ DST in RIF-R" = dsttimeall$stepxxdr, "FQ DST for all" = dsttimeall$fullxxdr),
at=c(1,2,3,6,7,8, 11,12,13),
col=cols, notch=FALSE,
ylab = "Person-months of active MOXI-R TB",
ylim=c(0,max(unlist(dsttimeall))),
xaxt ='n')
axis(1, at = c(2,7,12), labels = c("Among RIF-S,\nMOXI-R TB", "Among RIF-R,\nMOXI-R TB", "Among all\nMOXI-R TB"),
lwd.ticks = F)
legend("topleft", "B", bty="n", cex = 1.5)
par(mar=c(1,1,1,1))
plot.new()
# par(mar=c(1,1,1,1))
legend("center", legend = c("None", "RIF-R", "All"),
pch=15, col=cols, title = expression(bold("FQ DST performed for:")), xpd=NA, ncol=3, pt.cex = 3)
dev.off()
# rr vs rs moxi-r:
sum((c$RIF==0)*c$Freq)/sum(c$Freq)
sum((c$RIF==0)*c$MOXI*c$Freq)/sum(c$MOXI*c$Freq)
# fold-diffs in unsuccessful outcomes averted, simultaneous vs stepwise
(dstu1all$fullxxdr-dstu1all$noxxdr)
(dstu1all$stepxxdr-dstu1all$noxxdr)
mean(dstu1all$stepxxdr-dstu1all$noxxdr)
median(dstu1all$stepxxdr-dstu1all$noxxdr)
mean((dstu1all$fullxxdr-dstu1all$noxxdr)/mean(dstu1all$stepxxdr-dstu1all$noxxdr))
(dsttimeall$fullxxdr-dsttimeall$noxxdr)
(dsttimeall$stepxxdr-dsttimeall$noxxdr)
mean((dsttimeall$fullxxdr-dsttimeall$noxxdr)/mean(dsttimeall$stepxxdr-dsttimeall$noxxdr))
mean((dsttimeall$fullxxdr-dsttimeall$noxxdr)/mean(dsttimeall$stepxxdr-dsttimeall$noxxdr))
# dsts performed:
dsts <- function(patiententry) # xperts, then xxdrs, each first round then overall; and append rif and moxi status, and partialmoxi and pza
{ return(c(floor(patiententry["DSTs",4]),
round(10*(patiententry["DSTs",4]%%1)),
floor(patiententry["DSTs",ncol(patiententry)]),
round(10*(patiententry["DSTs",ncol(patiententry)]%%1)),
patiententry["RIF",1],
patiententry["MOXI",1],
patiententry["partialmoxi",1],
patiententry["PZA",1]) )
}
dstdst <- loutcomeboot(individualoutcomefunction = dsts, simoutput=dst, c=c)
save(dstdst, file=paste0("dstdst.",date,".Rdata"))
# total increased FQ DSTs from universal approach:
apply(dstdst$fullxxdr[,4,] - dstdst$stepxxdr[,4,] , 2, sum)
mean(apply(dstdst$fullxxdr[,2,] - dstdst$stepxxdr[,2,] , 2, sum)); sd(apply(dstdst$fullxxdr[,2,] - dstdst$stepxxdr[,2,] , 2, sum))
mean(apply(dstdst$fullxxdr[,2,] - dstdst$stepxxdr[,2,], 2, sum)/apply(dstdst$stepxxdr[,2,] - dstdst$noxxdr[,2,], 2, sum))
sd(apply(dstdst$fullxxdr[,2,] - dstdst$stepxxdr[,2,], 2, sum)/apply(dstdst$stepxxdr[,2,] - dstdst$noxxdr[,2,], 2, sum))
# dsts per added successful treatment (first treatment round only), comparing step to none and full to step:
(apply(dstdst$stepxxdr[,2,],2,sum) - apply(dstdst$fullxxdr[,2,],2,sum))/
(dstu1all$fullxxdr - dstu1all$stepxxdr)
mean((apply(dstdst$stepxxdr[,2,],2,sum) - apply(dstdst$fullxxdr[,2,],2,sum))/
(dstu1all$fullxxdr - dstu1all$stepxxdr));
sd((apply(dstdst$stepxxdr[,2,],2,sum) - apply(dstdst$fullxxdr[,2,],2,sum))/
(dstu1all$fullxxdr - dstu1all$stepxxdr))
(apply(dstdst$noxxdr[,2,],2,sum) - apply(dstdst$stepxxdr[,2,],2,sum))/
(dstu1all$stepxxdr - dstu1all$noxxdr)
median((apply(dstdst$noxxdr[,2,],2,sum) - apply(dstdst$stepxxdr[,2,],2,sum))/
(dstu1all$stepxxdr - dstu1all$noxxdr))
apply(dstdst$noxxdr[,2,],2,sum) - apply(dstdst$stepxxdr[,2,],2,sum)
mean(dstu1all$stepxxdr - dstu1all$noxxdr)
median(dstu1all$stepxxdr - dstu1all$noxxdr)
mean((apply(dstdst$noxxdr[,2,],2,sum) - apply(dstdst$stepxxdr[,2,],2,sum))/
median(dstu1all$stepxxdr - dstu1all$noxxdr))
mean((apply(dstdst$noxxdr[,2,],2,sum) - apply(dstdst$stepxxdr[,2,],2,sum))/
mean(dstu1all$stepxxdr - dstu1all$noxxdr))
sd((apply(dstdst$noxxdr[,2,],2,sum) - apply(dstdst$stepxxdr[,2,],2,sum))/
mean(dstu1all$stepxxdr - dstu1all$noxxdr))
rm(dstdst)
#####################################
# Acquired resistance
# load(paste0("dst.",date,".Rdata"))
dst <- dst[c("noxxdr","fullxxdr")]
load(paste0("impact.",date,".Rdata"))
combined <- list("baseline"=impact$baseline, "novelrrx"=impact$novelrrx, "noxxdr"=dst$noxxdr, "fullxxdr"=dst$fullxxdr)
save(combined, file=paste0("combined.",date,".Rdata"))
rm(impact)
rm(dst)
# # estimated reduction in infectious time and force of infection (not accounting for infectiousness during treatment, changes in a person's infectiousness over time, etc):
# l <- loutcomeboot(individualoutcomefunction =time.in.state, states=c(1,2,5,9,10), cutofftime=12*10, simoutput=combined, c=c)
# # overall
# lapply(l, function(x) apply((l$baseline[,1,]-x[,1,]), 2, sum)/apply(l$baseline[,1,],2,sum))
# ## Tracking acquired resistance
# adrevents <- function(patiententry, drugs=c("RIF","INH","MOXI","BDQ","PA")) # total events, which drugs acquired, and initial resistance
# {
# return(c(
# sum(
# patiententry["eventtype",]==eventtypes$treatmentstart &
# patiententry["TBstate",]==statetypes$treating_adr &
# colSums(patiententry[drugs,] != cbind(patiententry[drugs,1],patiententry[drugs,1:(ncol(patiententry)-1)]))>=1 ),
# rowSums(patiententry[drugs,] != cbind(patiententry[drugs,1],patiententry[drugs,1:(ncol(patiententry)-1)])),
# patiententry[drugs,1] ))
# }
#
# ladr <- loutcomeboot(individualoutcomefunction = adrevents, simoutput = combined, c = c)
#
# lapply(ladr, function(y) apply(y[,1,],2,sum)) #total events
# lapply(ladr, function(y) apply(y[,1,],2,sum)/1e5) #total event rate
# lapply(ladr, function(y) apply(y[,1,]*y[,1+5+1,],2,sum)/apply(y[,1+5+1,],2,sum)) #rate of events among initial RIF-R
# lapply(ladr, function(y) apply(y[,1,]*y[,1+5+3,],2,sum)/apply(y[,1+5+3,],2,sum)) #rate of events among initial moxi-R
# lapply(ladr, function(y) apply(y[,1,]*y[,1+5+4,],2,sum)/apply(y[,1+5+1,],2,sum)) #rate of events among initial BDQ-R
# lapply(ladr, function(y) apply(y[,1,]*y[,1+1,],2,mean)) #acquisitions of (any) RIF-R
# lapply(ladr, function(y) apply(y[,1,]*y[,1+3,],2,mean)) #acquisitions of (any) MOXI-R ??
# lapply(ladr, function(y) apply(y[,1,]*y[,1+4,],2,mean)) #acquisitions of (any) BDQ-R ??
#
#
# infectious time with drug resistance:
infectioustime <- function(patiententry,
resdrugs=list(
"rif"=list(c("RIF"), 1),
"moxi"=list(c("MOXI"), 1),
"himoxi"=list(c("MOXI","partialmoxi"), c(1,0), "and"),
"bdq"=list(c("BDQ"), 1),
"pa"=list(c("PA"), 1),
"noveldrug"=list(c("BDQ","PA"), c(1,1), "or"),
"none"=list(c())),
includeinitial=TRUE, cutofftime=12*10) # follow up to 4th diagnosis ie step 13. exclude untreated (states 1 and 2) if includeinitial==F.
{
elapsedtimes <- patiententry["eventtime",2:13] - patiententry["eventtime",1:12]
elapsedtimes[patiententry["eventtime",2:13]>cutofftime] <- cutofftime - (patiententry["eventtime",1:12])[patiententry["eventtime",2:13]>cutofftime]
elapsedtimes[elapsedtimes<0] <- 0
if (includeinitial) indices <- patiententry["TBstate",1:12] %in% c(1,2,5,9,10) else
indices <- patiententry["TBstate",1:12] %in% c(5,9,10)
indexlist <- lapply(resdrugs, function(y) {
if(length(y[[1]])==1) (indices & patiententry[y[[1]],1:12]==y[[2]]) else
if(length(y[[1]])>1) {if(y[[3]]=="and") indices & apply(patiententry[y[[1]],1:12]==y[[2]], 2, all) else
if(y[[3]]=="or") indices & apply(patiententry[y[[1]],1:12]==y[[2]], 2, any) } else
indices
})
t <- lapply(indexlist, function(y) sum(elapsedtimes*y))
return(c(unlist(t), patiententry[patientvars,1]))
}
# names(combined) <- c("baseline", "novelrrx", "noxxdr", "fullxxdr")
lres <- loutcomeboot(individualoutcomefunction = infectioustime, includeinitial=TRUE, simoutput = combined, c=c)
# lanytb <- loutcomeboot(individualoutcomefunction = infectioustime, resdrugs=list(c()), includeinitial=TRUE, simoutput = combined, c=c)
lrestreated <- loutcomeboot(individualoutcomefunction = infectioustime, includeinitial=FALSE, simoutput = combined, c=c)
# lanytbtreated <- loutcomeboot(individualoutcomefunction = infectioustime, resdrugs=list(c()), includeinitial=FALSE, simoutput = combined, c=c)
save(lres, lrestreated, file=paste0("lres_etc.",date,".Rdata"))
rm(combined)
# note that if n% of the cohort is BDQ-R, then average outcomes will be a weighted average of outcomes in the original cohort, and those in the bdq-r version of the cohort.
# but a weighted average won't capture the variability among that 5%.
# so if I want to who what happens with 4% BDQ-R (i.e. same as current rif-r), then I should run loutcomeboot with a desiredsize of 4/(100-4)*1e5, using combinedbdq,
# and then do a weighted average of each result with lres etc. Can just use the same c for the freq's, and I'll weight it later.
load(paste0("combinedbdq.",date,".Rdata"))
lbdqres <- loutcomeboot(individualoutcomefunction = infectioustime, includeinitial=TRUE, simoutput = combinedbdq, c=c, desiredsize = 1e5*4/(100-4))
lbdqrestreated <- loutcomeboot(individualoutcomefunction = infectioustime, includeinitial=FALSE, simoutput = combinedbdq, c=c, desiredsize = 1e5*4/(100-4))
save(lbdqres, lbdqrestreated, file=paste0("lbdqres_etc.",date,".Rdata"))
# rifr_bdq <- mapply(function(a,b) apply(a[,1,],2,mean)*.96 + apply(b[,1,],2,mean)*.04, a=lres, b=lbdqres)
# rifrtreated_bdq <- mapply(function(a,b) mean(apply(a[,1,],2,mean)*.96 + apply(b[,1,],2,mean)*.04), a=lrestreated, b=lbdqrestreated)
# boxplot(cbind(mapply(function(x,y) apply(y[,4,],2,mean), x=1, y=lres), mapply(function(a,b) apply(a[,4,],2,mean)*.96 + apply(b[,4,],2,mean)*.04, a=lres, b=lbdqres)), ylim=c(0,1))
# will mention bdq sensitiivty analysis in text, plan to put a figure in appendix (with a focux on post-treatment time, and what that means for the next round of trnamission)
rm(combinedbdq)
load("highres.20190226.Rdata")
lhighres <- loutcomeboot(individualoutcomefunction = infectioustime, includeinitial=TRUE, simoutput = highres, c=c)
lhighrestreated <- loutcomeboot(individualoutcomefunction = infectioustime, includeinitial=FALSE, simoutput = highres, c=c)
save(lhighres, lhighrestreated, file=paste0("lhighres_etc.",date,".Rdata"))
boxplot(lapply(lrestreated, function(x) apply(x[,6,],2,mean)/apply(x[,7,],2,mean)))
boxplot(lapply(lhighrestreated, function(x) apply(x[,6,],2,mean)/apply(x[,7,],2,mean)))
rm(highres)
###### FIGURE 4 #############
load(paste0("lres_etc.",date,".Rdata"))
load(paste0("lhighres_etc.",date,".Rdata"))
names(lhighres) <- names(lhighrestreated) <- c( "baseline", "novelrrx", "noxxdr", "fullxxdr")
a1 <- data.frame(lapply(lrestreated, function(y) apply(y[,1,],2,mean)/apply(y[,7,],2,mean)))
a2 <- data.frame(lapply(lrestreated, function(y) apply(y[,2,],2,mean)/apply(y[,7,],2,mean)))
a3 <- data.frame(lapply(lrestreated, function(y) apply(y[,4,],2,mean)/apply(y[,7,],2,mean)))
a4 <- data.frame(lapply(lrestreated, function(y) apply(y[,5,],2,mean)/apply(y[,7,],2,mean)))
a1$drug <- "Rifampin"
a2$drug <- "Moxifloxacin"
a3$drug <- "Bedaquiline"
a4$drug <- "Pretomanid"
a <- bind_rows(a1,a2,a3,a4)
am <- melt(a, id.vars=c("drug"))
am$params <- "Low ADR"
c1 <- data.frame(lapply(lres, function(y) apply(y[,1,],2,mean)/apply(y[,7,],2,mean)))
c2 <- data.frame(lapply(lres, function(y) apply(y[,2,],2,mean)/apply(y[,7,],2,mean)))
c3 <- data.frame(lapply(lres, function(y) apply(y[,4,],2,mean)/apply(y[,7,],2,mean)))
c4 <- data.frame(lapply(lres, function(y) apply(y[,5,],2,mean)/apply(y[,7,],2,mean)))
c1$drug <- "Rifampin"
c2$drug <- "Moxifloxacin"
c3$drug <- "Bedaquiline"
c4$drug <- "Pretomanid"
c <- bind_rows(c1,c2,c3,c4)
cm <- melt(a, id.vars=c("drug"))
cm$params <- "Low ADR"
b1 <- data.frame(lapply(lhighrestreated, function(y) apply(y[,1,],2,mean)/apply(y[,7,],2,mean)))
b2 <- data.frame(lapply(lhighrestreated, function(y) apply(y[,2,],2,mean)/apply(y[,7,],2,mean)))
b3 <- data.frame(lapply(lhighrestreated, function(y) apply(y[,4,],2,mean)/apply(y[,7,],2,mean)))
b4 <- data.frame(lapply(lhighrestreated, function(y) apply(y[,5,],2,mean)/apply(y[,7,],2,mean)))
b1$drug <- "Rifampin"
b2$drug <- "Moxifloxacin"
b3$drug <- "Bedaquiline"
b4$drug <- "Pretomanid"
b <- bind_rows(b1,b2,b3,b4)
bm <- melt(b, id.vars=c("drug"))
bm$params <- "High ADR"
require(dplyr)
require(data.table)
mm <- bind_rows(am,bm)
colnames(mm) <- c("Drug", "Scenario", "Infectious_Time", "Parameters")
mm$Parameters <- factor(mm$Parameters, levels=c("Low ADR", "High ADR"))
levels(mm$Parameters) <- c("Low", "High")
# levels(mm$Parameters) <- c("Post-Treatment TB, Low BPaMZ ADR Risk", "Post-Treatment TB, High BPaMZ ADR Risk")
d1 <- data.frame(lapply(lhighres, function(y) apply(y[,1,],2,mean)/apply(y[,7,],2,mean)))
d2 <- data.frame(lapply(lhighres, function(y) apply(y[,2,],2,mean)/apply(y[,7,],2,mean)))
d3 <- data.frame(lapply(lhighres, function(y) apply(y[,4,],2,mean)/apply(y[,7,],2,mean)))
d4 <- data.frame(lapply(lhighres, function(y) apply(y[,5,],2,mean)/apply(y[,7,],2,mean)))
d1$drug <- "Rifampin"
d2$drug <- "Moxifloxacin"
d3$drug <- "Bedaquiline"
d4$drug <- "Pretomanid"
d <- bind_rows(d1,d2,d3,d4)
dm <- melt(d, id.vars=c("drug"))
dm$params <- "High ADR"
require(dplyr)
require(data.table)
mm2 <- bind_rows(cm,dm)
colnames(mm2) <- c("Drug", "Scenario", "Infectious_Time", "Parameters")
mm2$Parameters <- factor(mm2$Parameters, levels=c("Low ADR", "High ADR"))
levels(mm2$Parameters) <- c("Low", "High")
mm$tbtime <- "Post-treatment time with TB"
mm2$tbtime <- "All time with TB"
mm3 <- rbind(mm,mm2)
mm3$Drug <- factor(mm3$Drug)
mm3$Scenario <- factor(mm3$Scenario, levels=rev(levels(mm3$Scenario)))
levels(mm3$Scenario) <- rev(c("Current Practice",
"BPaMZ for RIF-R, no FQ DST",
"BPaMZ for all, no FQ DST",
"BPaMZ for all, universal FQ DST"))
save(am, bm, cm, dm, mm, mm2, mm3, file="Fig4data.Rdata")
# total time, not proportion:
e1 <- data.frame(lapply(lrestreated, function(y) apply(y[,1,],2,sum)))
e2 <- data.frame(lapply(lrestreated, function(y) apply(y[,2,],2,sum)))
e3 <- data.frame(lapply(lrestreated, function(y) apply(y[,4,],2,sum)))
e4 <- data.frame(lapply(lrestreated, function(y) apply(y[,5,],2,sum)))
e1$drug <- "Rifampin"
e2$drug <- "Moxifloxacin"
e3$drug <- "Bedaquiline"
e4$drug <- "Pretomanid"
e <- bind_rows(e1,e2,e3,e4)
em <- melt(e, id.vars=c("drug"))
em$params <- "Low"
g1 <- data.frame(lapply(lres, function(y) apply(y[,1,],2,sum)))
g2 <- data.frame(lapply(lres, function(y) apply(y[,2,],2,sum)))
g3 <- data.frame(lapply(lres, function(y) apply(y[,4,],2,sum)))
g4 <- data.frame(lapply(lres, function(y) apply(y[,5,],2,sum)))
g1$drug <- "Rifampin"
g2$drug <- "Moxifloxacin"
g3$drug <- "Bedaquiline"
g4$drug <- "Pretomanid"
g <- bind_rows(g1,g2,g3,g4)
gm <- melt(g, id.vars=c("drug"))
gm$params <- "Low"
names(lhighres) <- names(lhighrestreated) <- c( "baseline", "novelrrx", "noxxdr", "fullxxdr")
f1 <- data.frame(lapply(lhighrestreated, function(y) apply(y[,1,],2,sum)))
f2 <- data.frame(lapply(lhighrestreated, function(y) apply(y[,2,],2,sum)))
f3 <- data.frame(lapply(lhighrestreated, function(y) apply(y[,4,],2,sum)))
f4 <- data.frame(lapply(lhighrestreated, function(y) apply(y[,5,],2,sum)))
f1$drug <- "Rifampin"
f2$drug <- "Moxifloxacin"
f3$drug <- "Bedaquiline"
f4$drug <- "Pretomanid"
f <- bind_rows(f1,f2,f3,f4)
fm <- melt(f, id.vars=c("drug"))
fm$params <- "High"
require(dplyr)
require(data.table)
mmt <- bind_rows(em,fm)
colnames(mmt) <- c("Drug", "Scenario", "Infectious_Time", "Parameters")
mmt$Parameters <- factor(mmt$Parameters, levels=c("Low", "High"))
h1 <- data.frame(lapply(lhighres, function(y) apply(y[,1,],2,sum)))
h2 <- data.frame(lapply(lhighres, function(y) apply(y[,2,],2,sum)))
h3 <- data.frame(lapply(lhighres, function(y) apply(y[,4,],2,sum)))
h4 <- data.frame(lapply(lhighres, function(y) apply(y[,5,],2,sum)))
h1$drug <- "Rifampin"
h2$drug <- "Moxifloxacin"
h3$drug <- "Bedaquiline"
h4$drug <- "Pretomanid"
h <- bind_rows(h1,h2,h3,h4)
hm <- melt(h, id.vars=c("drug"))
hm$params <- "High"
require(dplyr)
require(data.table)
mm2t <- bind_rows(gm,hm)
colnames(mm2t) <- c("Drug", "Scenario", "Infectious_Time", "Parameters")
mm2t$Parameters <- factor(mm2t$Parameters, levels=c("Low", "High"))
mmt$tbtime <- "Post-treatment time with TB"
mm2t$tbtime <- "All time with TB"
mm3t <- rbind(mmt,mm2t)
mm3t$Drug <- factor(mm3t$Drug)
mm3t$Scenario <- factor(mm3t$Scenario, levels=rev(levels(mm3t$Scenario)))
levels(mm3t$Scenario) <- rev(c("Current Practice",
"BPaMZ for RIF-R, no FQ DST",
"BPaMZ for all, no FQ DST",
"BPaMZ for all, universal FQ DST"))
save(em, fm, gm, hm, mmt, mm2t, mm3t, file="Fig4data_t.Rdata")
require(ggplot2)
load("Fig4data_t.Rdata")
pdf("Fig4fixed.pdf", width=8, height=10)
# better would be total time (person-months in cohort, not %) with dr-tb
ggplot(data = transform(mm3t, Drug=factor(Drug, levels=c("Rifampin","Moxifloxacin","Bedaquiline","Pretomanid"))),
aes(x=Scenario, y=Infectious_Time)) +
geom_col(aes(fill=Parameters), position = position_dodge(), width=0.8) +
facet_grid(Drug ~ tbtime) +
coord_flip() +
# ggtitle("Potential for drug resistance transmission") +
theme(plot.title = element_text(hjust = 0.5)) +
xlab("BPaMZ implementation scenario") +
ylab("Person-months with active TB resistant to the specified drug, per 100,000 TB cases") +
theme(legend.position="bottom") + guides(fill=guide_legend(ncol=2)) + labs(fill = "Risk of acquiring resistance during BPaMZ:")
dev.off()
# ggplot(data = transform(mm, Drug=factor(Drug, levels=c("Moxifloxacin","Rifampin","Bedaquiline","Pretomanid"))),
# aes(x=Scenario, y=Infectious_Time)) +
# geom_col(aes(fill=Parameters), position = position_dodge(0.8), width=0.5) +
# facet_wrap( ~ Drug, ncol=2) +
# coord_flip() +
# ggtitle("Potential for drug resistance transmission") + theme(plot.title = element_text(hjust = 0.5)) +
# ylab("TB in cohort after first treatment attempt:\nFraction resistant to the specified drug") +
# theme(legend.position="bottom") + guides(fill=guide_legend(ncol=2)) + labs(fill = "Fraction of TB\nresistant to drug") +
# geom_col(data=transform(mm2, Drug=factor(Drug, levels=c("Moxifloxacin","Rifampin","Bedaquiline","Pretomanid"))),
# aes(x=as.numeric(Scenario)+0.1, fill=Parameters), position= position_dodge(0.8), width=0.5) +
# scale_fill_manual(values=c("black", "darkblue", "gray","lightblue"))