-
Notifications
You must be signed in to change notification settings - Fork 0
/
app_function.R
1387 lines (1285 loc) · 89.9 KB
/
app_function.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
### Shiny app: Eliciting Prior Opinion on External Biases
## Run a Shiny app which allows for the elicitation of expert opinion
## regarding external biases as defined in Wadsworth et al. (Under review)
# The 'OUT_dir' argument specifies the directory where output files should be saved.
# elicit_func() will run the Shiny app.
# - Once the elicitation is completed, the final elicitation plots and
# final elicited bias prior parameter values will be saved to OUT_dir.
elicit_func <- function(OUT_dir="") {
pixels = "41px"
shinyApp(
ui = fluidPage(titlePanel("Eliciting opinion on Dose-Response relationships with topiramate"),
bootstrapPage(tabsetPanel(
tabPanel("Outline",
br(),
fluidRow(
column(2,
br(),
br(),
br(),
textInput("expert","Please enter your name here.")
),
column(9,
h3(HTML(text=c('<span style="color:blue">Objective: We wish to elicit your opinion on how similar the dose-response relationship in children (4-10 years) in a new study will be to the dose-response relationships in adolescents (11-17 years) and adults that have been observed in a previous study.</span>
<br/>
<br/>
<span style="color:red">Old study:</span> Explored the efficacy of adjunctive topiramate in patients with partial seizures or primary generalised tonic-clonic seizures aged 11-85.
<br/>
<br/>
<span style="color:red">New study:</span> Exploring the efficacy of adjunctive topiramate in patients with partial seizures or primary generalised tonic-clonic seizures aged 4-10 years.
<br/>
<br/>
The response here is percent change in seizure frequency from baseline.')))
)
)
),
tabPanel("Existing data",
checkboxInput("check_AdultData","Tick to show adult data"),
checkboxInput("check_AdolescentData","Tick to show adolescent data"),
h4("This plot presents the line of best fit for adults and adolescents, found from existing data.
The line of best fit represents our best estimate of what the average response on a dose is."),
plotOutput("ExistingPlot")
),
tabPanel("Dose-response in children",
fluidRow(
column(8,
conditionalPanel( condition = "input.Tick_AvgResp == false",
br(),
h4("Imagine you have treated a large number of paediatric patients (say 100). Can you give us your opinion about
their average response?"),
h4("Pick the shape of the line of best fit for children which most closely matches your belief."),
plotOutput("ShapePlot1")),
conditionalPanel( condition = "input.Tick_AvgResp == true",
h4("This plot illustrates your best guess on the line of best fit for the dose-response relationship in children."),
plotOutput("ERPlot"))),
column(4,
br(),
br(),
checkboxInput("Tick_AvgResp","Elicit expert opinion on the shape of the average response."),
conditionalPanel( condition = "input.Tick_AvgResp == false",
radioButtons("ShapeAvgResp", "Select the shape that best matches your belief",
c("A", "B","C","D","E"))
),
conditionalPanel( condition = "input.Tick_AvgResp == true",
h4("Given the existing data, adjust the sliders to give your best guess at the average response amongst children
aged 2-11 years on placebo and a high dose of the test treatment."),
uiOutput("sliderAvgPlac"),
uiOutput("sliderAvgMax")
))
)
),
tabPanel("Uncertainty regarding dose-response in children",
fluidRow(
column(8,
conditionalPanel( condition = "input.Tick_PCtiles == false",
h4("These plots present the options for the shape which most closely matches your belief regarding your opinion on where the line of best fit for the dose-response relationship in children will lie when we model data from the new study."),
br(),
plotOutput("ShapePlot2")),
conditionalPanel( condition = "input.Tick_PCtiles == true",
h4("This plot illustrates your opinion on where the line of best fit for the dose-response relationship in children will lie when we model data from the new study."),
plotOutput("ERPlot2"))),
column(4,
br(),
br(),
checkboxInput("Tick_PCtiles","Elicit expert opinion on the shape of 5th and 95th percentiles."),
conditionalPanel( condition = "input.Tick_PCtiles == false",
radioButtons("ShapePCtiles", "Select the shape that best matches your belief",
c("A", "B","C","D"))
),
conditionalPanel( condition = "input.Tick_PCtiles == true",
br(),
uiOutput("sliderDose0"),
uiOutput("sliderDosemid"),
uiOutput("sliderDosehigh")))
)
),
tabPanel("Opinion on placebo",
h4("Imagine you have treated 100 paediatric patients, can you give us your opinion regarding
their average percent seizure frequency change from baseline, when on placebo. Please
fill in the text boxes under the histogram below to mark your belief. Every text box takes a number representing the % chance and matches with the
corresponding interval it is below."),
h4("The more stongly you believe the average response
would lie in a particular interval, the more % chance you should add. If you believe it
impossible that the average response would lie in a particular interval, then add no % chance
to that interval. If you are certain the average response would lie in a certain interval,
100% should be in that interval."),
h4(div(textOutput("Meantext1"), style = "color:red")),
h4(div(textOutput("Uncerttext1"), style = "color:green")),
br(),
fluidRow(
column(3,
checkboxInput("Tick_DensHist0","Show density")
),
column(5,
div(h3(textOutput("text1")), style = "color:red")
)
),
plotOutput("HistPlot0", click = "plot_click1"),
br(),
br(),
br(),
br(),
div(style="display:inline-block",class="outer",width="100%",
div(style="display:inline-block",h5("_________")),
div(style="display:inline-block",textInput("n100_n95_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n95_n90_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n90_n85_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n85_n80_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n80_n75_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n75_n70_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n70_n65_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n65_n60_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n60_n55_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n55_n50_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n50_n45_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n45_n40_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n40_n35_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n35_n30_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n30_n25_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n25_n20_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n20_n15_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n15_n10_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n10_n05_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n05_n00_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n00_p05_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p05_p10_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p10_p15_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p15_p20_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p20_p25_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p25_p30_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p30_p35_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p35_p40_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p40_p45_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p45_p50_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p50_p55_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p55_p60_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p60_p65_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p65_p70_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p70_p75_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p75_p80_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p80_p85_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p85_p90_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p90_p95_0", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p95_p100_0", label = "", value = 0,width=pixels))
)
),
tabPanel("Opinion on moderate dose",
h4("Imagine you have treated 100 paediatric patients, can you give us your opinion regarding
their average percent seizure frequency change from baseline, when on a moderate dose of
8 mg/kg/day. Please
fill in the text boxes under the histogram below to mark your belief. Every text box takes a number representing the % chance and matches with the
corresponding interval it is below."),
h4("The more stongly you believe the average response
would lie in a particular interval, the more % chance you should add. If you believe it
impossible that the average response would lie in a particular interval, then add no % chance
to that interval. If you are certain the average response would lie in a certain interval,
100% should be in that interval."),
h4(div(textOutput("Meantext2"), style = "color:red")),
h4(div(textOutput("Uncerttext2"), style = "color:green")),
br(),
fluidRow(
column(3,
checkboxInput("Tick_DensHistMid","Show density")
),
column(5,
div(h3(textOutput("text2")), style = "color:red")
)
),
plotOutput("HistPlotMid", click = "plot_click2"),
br(),
br(),
br(),
br(),
div(style="display:inline-block",class="outer",width="100%",
div(style="display:inline-block",h5("_________")),
div(style="display:inline-block",textInput("n100_n95_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n95_n90_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n90_n85_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n85_n80_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n80_n75_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n75_n70_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n70_n65_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n65_n60_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n60_n55_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n55_n50_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n50_n45_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n45_n40_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n40_n35_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n35_n30_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n30_n25_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n25_n20_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n20_n15_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n15_n10_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n10_n05_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n05_n00_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n00_p05_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p05_p10_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p10_p15_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p15_p20_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p20_p25_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p25_p30_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p30_p35_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p35_p40_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p40_p45_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p45_p50_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p50_p55_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p55_p60_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p60_p65_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p65_p70_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p70_p75_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p75_p80_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p80_p85_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p85_p90_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p90_p95_Mid", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p95_p100_Mid", label = "", value = 0,width=pixels))
)
),
tabPanel("Opinion on high dose",
h4("Imagine you have treated 100 paediatric patients, can you give us your opinion regarding
their average percent seizure frequency change from baseline, when on a high dose of
16 mg/kg/day. Please
fill in the text boxes under the histogram below to mark your belief. Every text box takes a number representing the % chance and matches with the
corresponding interval it is below."),
h4("The more stongly you believe the average response
would lie in a particular interval, the more % chance you should add. If you believe it
impossible that the average response would lie in a particular interval, then add no % chance
to that interval. If you are certain the average response would lie in a certain interval,
100% should be in that interval."),
h4(div(textOutput("Meantext3"), style = "color:red")),
h4(div(textOutput("Uncerttext3"), style = "color:green")),
br(),
fluidRow(
column(3,
checkboxInput("Tick_DensHistHigh","Show density")
),
column(5,
div(h3(textOutput("text3")), style = "color:red")
)
),
plotOutput("HistPlotHigh", click = "plot_click3"),
br(),
br(),
br(),
br(),
div(style="display:inline-block",class="outer",width="100%",
div(style="display:inline-block",h5("_________")),
div(style="display:inline-block",textInput("n100_n95_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n95_n90_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n90_n85_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n85_n80_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n80_n75_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n75_n70_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n70_n65_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n65_n60_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n60_n55_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n55_n50_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n50_n45_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n45_n40_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n40_n35_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n35_n30_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n30_n25_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n25_n20_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n20_n15_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n15_n10_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n10_n05_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n05_n00_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("n00_p05_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p05_p10_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p10_p15_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p15_p20_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p20_p25_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p25_p30_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p30_p35_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p35_p40_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p40_p45_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p45_p50_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p50_p55_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p55_p60_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p60_p65_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p65_p70_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p70_p75_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p75_p80_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p80_p85_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p85_p90_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p90_p95_High", label = "", value = 0,width=pixels)),
div(style="display:inline-block",textInput("p95_p100_High", label = "", value = 0,width=pixels))
)
),
tabPanel("Uncertainty regarding dose-response in children",
fluidRow(
column(3,
checkboxInput("Tick_ShowUncert","Tick to show uncertainty")
),
column(5,
conditionalPanel(condition = 'input.Tick_ShowUncert == true',
checkboxInput("Tick_ShapeUncertShow","Show uncertainty from earlier shape choice"))
)
),
fluidRow(
column(3,
checkboxInput("Tick_HistShow","Show histograms.")
),
column(3,
conditionalPanel(condition = 'input.Tick_HistShow == true',
checkboxInput("Tick_DensShow","Show densities."))
)
),
h4("This plot illustrates your opinion on where the line of best fit for the dose-response relationship in children will lie when we model data from the new study."),
plotOutput("ERPlot3")
),
tabPanel("Comparison of opinion and existing data",
fluidRow(
column(3,
checkboxInput("Tick_LogFinal","Show the log plot.")
),
column(3,
checkboxInput("check_ChildrenData","Tick to show simulated data for younger children based on expert opinion")
)
), conditionalPanel( condition = "input.Tick_LogFinal == false",
h4("This plot compares fitted dose-response relationships in adults and adolescents fitted to historical data compared with your opinion on the dose-response relationship in children in the new study."),
br(),
plotOutput("FinalPlot")
),
conditionalPanel( condition = "input.Tick_LogFinal == true",
h4("This plot compares fitted dose-response relationships in adults and adolescents fitted to historical data compared with your opinion on the dose-response relationship in children in the new study, on the log scale."),
br(),
plotOutput("FinalLogPlot"))
)
))
),
server = function(input, output,session) {
# Model specifications
alpha <- 4.4469 #adult placebo effect (from Girgis et al. paper)
betap <- -0.0627 #adult concentration effect (from Girgis et al. paper)
gamma1 <- 0.0361 #difference between adult and paediatric placebo effect (from Girgis et al. paper)
gamma2 <- 0.0048 #difference between adult and paediatric concentration effect (from Girgis et al. paper)
dose <- 0:16 #dose range
dose2 <- dose^2
dnum1 = 1 #index in 'dose' of the first dose we elicit expert opinion about (here dose = 0)
dnum2 = 9 #index in 'dose' of the second dose we elicit expert opinion about (here dose = 8)
dnum3 = 17 #index in 'dose' of the third dose we elicit expert opinion about (here dose = 16)
dnums = c(dnum1,dnum2,dnum3)
dose.optim <- dose[dnums]
#PD is log(% change from baseline + 110)
PD_age0 <- alpha + betap*dose #adults
PD_age1 <- alpha + betap*dose + gamma1 + gamma2*dose #adolescents
#Define the 'line of best fit' shapes that will be inititially displayed
bias_meanshape1 <- (PD_age0+PD_age1)/2 + (PD_age0+PD_age1)/45
bias_meanshape2 <- PD_age1
bias_meanshape3 <- (PD_age0+PD_age1)/2
bias_meanshape4 <- PD_age0
bias_meanshape5 <- (PD_age0+PD_age1)/2 - (PD_age0+PD_age1)/45
#Define histogram bins
binwidth = 5
bins <- seq(-100,100,by=binwidth)
bins1 <- seq(-100,95,by=binwidth)
bins2 <- seq(-95,100,by=binwidth)
center_bin <- seq(-97.5,97.5,by=binwidth)
#Define functions for transforming between % change and log(% change) scales
back_trans <- function(x){
exp(x)-110
}
forw_trans <- function(x){
log(x+110)
}
#Define function for finding bias prior parameters from elicited opinion, for reactive plots
inits1 <- c(0,0,0)
ofn_q95 <- function(x,q95,ChosenAvgLine){
ex1 = exp(x[1])
ex2 = x[2]
ex3 = exp(x[3])
bias_var <- ex1^2 + (dose.optim^2)*(ex2^2 + ex3^2) + 2*dose.optim*ex1*ex2
bias.q95 <- qnorm(0.95,ChosenAvgLine[dnums],sqrt(bias_var))
y = sum(abs(bias.q95 - q95))
return(y)
}
#Define the line of best fit for children
mean_line <- reactive({
bias_mean <- c(AvgInt(),AvgSlo())
ChosenAvgLine <- PD_age1 + bias_mean[1] + dose*bias_mean[2]
return(ChosenAvgLine)
})
# Define reactive values for starting points for Placebo and Max dose, based on choice of mean line
reactiveAvgPlac <- reactive({
if(input$ShapeAvgResp == "A"){
ChoiceAvgInt <- back_trans(bias_meanshape1[dnum1])
}else if(input$ShapeAvgResp == "B"){
ChoiceAvgInt <- back_trans(bias_meanshape2[dnum1])
}else if(input$ShapeAvgResp == "C"){
ChoiceAvgInt <- back_trans(bias_meanshape3[dnum1])
}else if(input$ShapeAvgResp == "D"){
ChoiceAvgInt <- back_trans(bias_meanshape4[dnum1])
}else if(input$ShapeAvgResp == "E"){
ChoiceAvgInt <- back_trans(bias_meanshape5[dnum1])
}
})
reactiveAvgMax <- reactive({
if(input$ShapeAvgResp == "A"){
ChoiceAvgMax <- back_trans(bias_meanshape1[dnum3])
}else if(input$ShapeAvgResp == "B"){
ChoiceAvgMax <- back_trans(bias_meanshape2[dnum3])
}else if(input$ShapeAvgResp == "C"){
ChoiceAvgMax <- back_trans(bias_meanshape3[dnum3])
}else if(input$ShapeAvgResp == "D"){
ChoiceAvgMax <- back_trans(bias_meanshape4[dnum3])
}else if(input$ShapeAvgResp == "E"){
ChoiceAvgMax <- back_trans(bias_meanshape5[dnum3])
}
})
# Define reactive values for starting points for percentiles
reactiveRespByDose <- reactive({
ChosenAvgLine <- mean_line()
if(input$ShapePCtiles == "A"){
qmfit <- c(ChosenAvgLine[dnum1] + 0.126, ChosenAvgLine[dnum2] + 0.3676, ChosenAvgLine[dnum3] + 1.055)
osol <- optim(par=inits1, ofn_q95, q95 = qmfit,ChosenAvgLine=mean_line())
bias_var <- exp(osol$par[1])^2 + (dose^2)*(osol$par[2]^2 + exp(osol$par[3])^2) + 2*dose*exp(osol$par[1])*osol$par[2]
ChoiceRespByDose <- c(qnorm(0.95,ChosenAvgLine,sqrt(bias_var)),qnorm(0.05,ChosenAvgLine,sqrt(bias_var)))
return(ChoiceRespByDose)
}else if(input$ShapePCtiles == "B"){
qmfit <- c(ChosenAvgLine[dnum1] + 0.2, ChosenAvgLine[dnum2] + 0.2, ChosenAvgLine[dnum3] + 0.2)
osol <- optim(par=inits1, ofn_q95, q95 = qmfit,ChosenAvgLine=mean_line())
bias_var <- exp(osol$par[1])^2 + (dose^2)*(osol$par[2]^2 + exp(osol$par[3])^2) + 2*dose*exp(osol$par[1])*osol$par[2]
ChoiceRespByDose <- c(qnorm(0.95,ChosenAvgLine,sqrt(bias_var)),qnorm(0.05,ChosenAvgLine,sqrt(bias_var)))
return(ChoiceRespByDose)
}else if(input$ShapePCtiles == "C"){
qmfit <- c(ChosenAvgLine[dnum1] + 0.7, ChosenAvgLine[dnum2] + 0.35, ChosenAvgLine[dnum3] + 0.3)
osol <- optim(par=inits1, ofn_q95, q95 = qmfit,ChosenAvgLine=mean_line())
bias_var <- exp(osol$par[1])^2 + (dose^2)*(osol$par[2]^2 + exp(osol$par[3])^2) + 2*dose*exp(osol$par[1])*osol$par[2]
ChoiceRespByDose <- c(qnorm(0.95,ChosenAvgLine,sqrt(bias_var)),qnorm(0.05,ChosenAvgLine,sqrt(bias_var)))
return(ChoiceRespByDose)
}else if(input$ShapePCtiles == "D"){
qmfit <- c(ChosenAvgLine[dnum1] + 0.7, ChosenAvgLine[dnum2] + 0.35, ChosenAvgLine[dnum3] + 1.2)
osol <- optim(par=inits1, ofn_q95, q95 = qmfit,ChosenAvgLine=mean_line())
bias_var <- exp(osol$par[1])^2 + (dose^2)*(osol$par[2]^2 + exp(osol$par[3])^2) + 2*dose*exp(osol$par[1])*osol$par[2]
ChoiceRespByDose <- c(qnorm(0.95,ChosenAvgLine,sqrt(bias_var)),qnorm(0.05,ChosenAvgLine,sqrt(bias_var)))
return(ChoiceRespByDose)
}
})
#Define reactive optim
reactiveOptim <- reactive({
ChosenAvgLine <- mean_line()
Chosenqmfit <- forw_trans(c(input$EliDose0,input$EliDosemid,input$EliDosehigh))
Chosenosol <- optim(par=inits1, ofn_q95, q95 = Chosenqmfit, ChosenAvgLine=mean_line(), control=list(maxit = 1000))
return(Chosenosol)
})
#Define reactive values for final bias covariance matrix
reactiveFinalCov <- reactive({
Chosenosol <- reactiveOptim()
Choicevar1 <- exp(Chosenosol$par[1])^2
Choicevar2 <- Chosenosol$par[2]^2 + exp(Chosenosol$par[3])^2
Choicecov <- exp(Chosenosol$par[1])*Chosenosol$par[2]
return(c(Choicevar1,Choicevar2,Choicecov,Chosenosol$par))
})
#Define the sliders
output$sliderAvgPlac <- renderUI({
sliderInput("AvgPlac", "Average response on placebo:",
min=-100, max=100, value=reactiveAvgPlac(), round=-1, step=0.1)
})
output$sliderAvgMax <- renderUI({
sliderInput("AvgMax", "Maximum effect of topiramate (average response on the highest dose):",
min=-100, max=100, value=reactiveAvgMax(), round=-1, step=0.1)
})
AvgInt <- reactive({
forw_trans(input$AvgPlac) - PD_age1[1]
})
AvgSlo <- reactive({ #this is: (slope for elicited children) - (slope for adolescent data)
(forw_trans(input$AvgMax) - forw_trans(input$AvgPlac))/dose[dnum3] - (PD_age1[2]-PD_age1[1])/(dose[2]-dose[1])
})
output$sliderDose0 <- renderUI({
sliderInput("EliDose0", "Given the existing data, state a value which you are 95% sure the average response
amongst younger children on placebo will lie below:",
min=round(back_trans(PD_age1[dnum1] + AvgInt() + dose[dnum1]*AvgSlo())), max=100, value=back_trans(reactiveRespByDose()[dnum1]), round=-1, step=0.1)
})
output$sliderDosemid <- renderUI({
sliderInput("EliDosemid", "Given the existing data, state a value which you are 95% sure the average response
amongst younger children on a moderate dose of the test treatment will lie below:",
min=round(back_trans(PD_age1[dnum2] + AvgInt() + dose[dnum2]*AvgSlo())), max=100, value=back_trans(reactiveRespByDose()[dnum2]), round=-1, step=0.1)
})
output$sliderDosehigh <- renderUI({
sliderInput("EliDosehigh", "Given the existing data, state a value which you are 95% sure the average response
amongst younger children on a high dose of the test treatment will lie below:",
min=round(back_trans(PD_age1[dnum3] + AvgInt() + dose[dnum3]*AvgSlo())), max=100, value=back_trans(reactiveRespByDose()[dnum3]), round=-1, step=0.1)
})
#define the reactive plots
output$ExistingPlot <- renderPlot({
par(mar=c(15,5,1.1,2.1), xpd=TRUE)
plot(dose,back_trans(PD_age0),ylim=c(-100,0),lty=1,lwd=2,cex.lab=2,cex.axis=1.8,type='l',
xaxp = c(0,16,2), xlab='Dose (mg/kg/day)',ylab='% change from baseline in seizure frequency')
lines(dose,back_trans(PD_age1),type='l',col='blue',lty=1,lwd=2)
legend(-0.47,-117,cex=1.5,c("Adult line of best fit", "Adolescent line of best fit","Adult data","Adolescent data"),
col=c(1,'blue',1,'blue'), lty=c(1,1,NA,NA),pch=c(NA,NA,15,16),lwd=2)
par(mar=c(15,5,1.1,2.1), xpd=F)
abline(h=seq(-100,0,by=20), v=c(0,8,16), col="gray", lty=3)
if(input$check_AdolescentData == TRUE){
set.seed(123)
index_Adoles <- sample(1:length(PD_age1),size=170,replace = T)
sampAdoles <- PD_age1[index_Adoles]
AdolesData <- sampAdoles + rnorm(170,0,sqrt(0.03))
WhichAdolesDose <- dose[index_Adoles]
points(WhichAdolesDose,back_trans(AdolesData),pch=9,col='blue',cex=1)
}
if(input$check_AdultData == TRUE){
set.seed(100)
index_Adult <- sample(1:length(PD_age0),size=170,replace = T)
sampAdult <- PD_age0[index_Adult]
AdultData <- sampAdult + rnorm(170,0,sqrt(0.03))
WhichAdultDose <- dose[index_Adult]
points(WhichAdultDose,back_trans(AdultData),pch=15,cex=1)
}
}, height = 820, width = 1200)
ERPlot_func <- function(){
par(mar=c(15,5,1.1,2.1), xpd=TRUE)
plot(dose,back_trans(PD_age0),ylim=c(-100,max(c(0,back_trans(PD_age1 + AvgInt() + dose*AvgSlo())))),lty=1,lwd=2,cex.lab=2,cex.axis=1.8,type='l',
xaxp = c(0,16,2), xlab='Dose (mg/kg/day)',ylab='% change from baseline in seizure frequency')
lines(dose,back_trans(PD_age1),type='l',col='blue',lty=1,lwd=2)
legend(-0.47,-117,cex=1.5,col=c(1,'blue','red'),lty=c(1,1,1),lwd=2,
c("Adult line of best fit","Adolescent line of best fit",
"Your best guess at the dose-response relationship in children"))
par(mar=c(15,5,1.1,2.1), xpd=F)
abline(h=seq(-100,max(c(0,back_trans(PD_age1 + AvgInt() + dose*AvgSlo()))),by=20),
v=c(0,8,16), col="gray", lty=3)
lines(dose,back_trans(PD_age1 + AvgInt() + dose*AvgSlo()),type='l',col='red',lty=1,lwd=2)
}
ERPlot_plot <- reactive({
pdf(paste(OUT_dir,input$expert,"_meanPlot.pdf",sep=''),pointsize=10,paper="a4")
ERPlot_func()
dev.off()
})
output$ERPlot <- renderPlot({
ERPlot_func()
ERPlot_plot()
}, height = 820, width = 1200)
ERPlot2_func <- function(){
ChosenAvgLine <- mean_line()
Chosenosol <- reactiveOptim()
Chosenbias_var <- exp(Chosenosol$par[1])^2 + (dose^2)*(Chosenosol$par[2]^2 + exp(Chosenosol$par[3])^2) + 2*dose*exp(Chosenosol$par[1])*Chosenosol$par[2]
ChoicePlotqmfit95 <- qnorm(0.95,ChosenAvgLine,sqrt(Chosenbias_var))
ChoicePlotqmfit05 <- qnorm(0.05,ChosenAvgLine,sqrt(Chosenbias_var))
par(mar=c(15,5,1.1,2.1), xpd=TRUE)
plot(dose,back_trans(ChosenAvgLine),
ylim=c(-100,max(c(back_trans(ChosenAvgLine),back_trans(ChoicePlotqmfit95),
input$EliDose0,input$EliDosemid,input$EliDosehigh))),
lty=1,lwd=2,cex.lab=2,cex.axis=1.8,type='l',col='red', xaxp = c(0,16,2),
xlab='Dose (mg/kg/day)',ylab='% change from baseline in seizure frequency')
legend(-0.47,-125,cex=1.5, lty=c(1,2,NA),lwd=2, pch=c(NA,NA,16),
c("Your best guess at the dose-response relationship in children",
"Lines indicating your 90% credibility interval around the line of best fit",
"Values you gave that you were 95% sure the true average response would lie below"),col=c('red','darkgreen','darkblue'))
par(mar=c(15,5,1.1,2.1), xpd=F)
abline(h=seq(-100,max(c(back_trans(ChosenAvgLine),back_trans(ChoicePlotqmfit95),
input$EliDose0,input$EliDosemid,input$EliDosehigh)),by=20), v=c(0,8,16), col="gray", lty=3)
lines(dose,back_trans(ChoicePlotqmfit95),type='l',col='darkgreen',lty=2,lwd=2)
lines(dose,back_trans(ChoicePlotqmfit05),type='l',col='darkgreen',lty=2,lwd=2)
points(dose[dnum1],input$EliDose0,cex=2.5,col="darkblue",pch=16)
points(dose[dnum2],input$EliDosemid,cex=2.5,col="darkblue",pch=16)
points(dose[dnum3],input$EliDosehigh,cex=2.5,col="darkblue",pch=16)
}
ERPlot2_plot <- reactive({
pdf(paste(OUT_dir,input$expert,"_uncertPlot.pdf",sep=''),pointsize=10,paper="a4")
ERPlot2_func()
dev.off()
})
output$ERPlot2 <- renderPlot({
ERPlot2_func()
ERPlot2_plot()
}, height = 820, width = 1200)
output$ShapePlot1 <- renderPlot({
par(mfrow=c(3,2),oma = c(0, 0, 2, 0))
#Mean shape 1
plot(dose,back_trans(PD_age0),ylim=c(-100,0),lty=1,lwd=2,cex.lab=2,cex.axis=1.8,type='l',
xlab='Dose (mg/kg/day)',ylab='% change',main='A',cex.main=3)
lines(dose,back_trans(PD_age1),type='l',col='blue',lty=1,lwd=2)
lines(dose,back_trans(bias_meanshape1),type='l',col='red',lty=2,lwd=2)
#Mean shape 2
plot(dose,back_trans(PD_age0),ylim=c(-100,0),lty=1,lwd=2,cex.lab=2,cex.axis=1.8,type='l',
xlab='Dose (mg/kg/day)',ylab='% change',main='B',cex.main=3)
lines(dose,back_trans(PD_age1),type='l',col='blue',lty=1,lwd=2)
lines(dose,back_trans(bias_meanshape2),type='l',col='red',lty=2,lwd=2)
#Mean shape 3
plot(dose,back_trans(PD_age0),ylim=c(-100,0),lty=1,lwd=2,cex.lab=2,cex.axis=1.8,type='l',
xlab='Dose (mg/kg/day)',ylab='% change',main='C',cex.main=3)
lines(dose,back_trans(PD_age1),type='l',col='blue',lty=1,lwd=2)
lines(dose,back_trans(bias_meanshape3),type='l',col='red',lty=2,lwd=2)
#Mean shape 4
plot(dose,back_trans(PD_age0),ylim=c(-100,0),lty=1,lwd=2,cex.lab=2,cex.axis=1.8,type='l',
xlab='Dose (mg/kg/day)',ylab='% change',main='D',cex.main=3)
lines(dose,back_trans(PD_age1),type='l',col='blue',lty=1,lwd=2)
lines(dose,back_trans(bias_meanshape4),type='l',col='red',lty=2,lwd=2)
#Mean shape 5
plot(dose,back_trans(PD_age0),ylim=c(-100,0),lty=1,lwd=2,cex.lab=2,cex.axis=1.8,type='l',
xlab='Dose (mg/kg/day)',ylab='% change',main='E',cex.main=3)
lines(dose,back_trans(PD_age1),type='l',col='blue',lty=1,lwd=2)
lines(dose,back_trans(bias_meanshape5),type='l',col='red',lty=2,lwd=2)
#Legend
plot(1, type="n", axes=F, xlab="", ylab="")
legend("topleft",c('Adults','Adolescents','Your best guess at the dose-response relationship in children'),cex=2,col=c(1,'blue','red'),lty=c(1,1,2))
}, height = 900, width = 1200)
ShapePlot2_func <- function(){
ChosenAvgLine <- mean_line()
#Choice A
qmfitA <- c(ChosenAvgLine[dnum1] + 0.126, ChosenAvgLine[dnum2] + 0.3676, ChosenAvgLine[dnum3] + 1.055)
osolA <- optim(par=inits1, ofn_q95, q95 = qmfitA, ChosenAvgLine=mean_line())
bias_varA <- exp(osolA$par[1])^2 + (dose^2)*(osolA$par[2]^2 + exp(osolA$par[3])^2) + 2*dose*exp(osolA$par[1])*osolA$par[2]
ChoiceqmfitA95 <- qnorm(0.95,ChosenAvgLine,sqrt(bias_varA))
ChoiceqmfitA05 <- qnorm(0.05,ChosenAvgLine,sqrt(bias_varA))
#Choice B
qmfitB <- c(ChosenAvgLine[dnum1] + 0.2, ChosenAvgLine[dnum2] + 0.2, ChosenAvgLine[dnum3] + 0.2)
osolB <- optim(par=inits1, ofn_q95, q95 = qmfitB, ChosenAvgLine=mean_line())
bias_varB <- exp(osolB$par[1])^2 + (dose^2)*(osolB$par[2]^2 + exp(osolB$par[3])^2) + 2*dose*exp(osolB$par[1])*osolB$par[2]
ChoiceqmfitB95 <- qnorm(0.95,ChosenAvgLine,sqrt(bias_varB))
ChoiceqmfitB05 <- qnorm(0.05,ChosenAvgLine,sqrt(bias_varB))
#Choice C
qmfitC <- c(ChosenAvgLine[dnum1] + 0.7, ChosenAvgLine[dnum2] + 0.35, ChosenAvgLine[dnum3] + 0.3)
osolC <- optim(par=inits1, ofn_q95, q95 = qmfitC, ChosenAvgLine=mean_line())
bias_varC <- exp(osolC$par[1])^2 + (dose^2)*(osolC$par[2]^2 + exp(osolC$par[3])^2) + 2*dose*exp(osolC$par[1])*osolC$par[2]
ChoiceqmfitC95 <- qnorm(0.95,ChosenAvgLine,sqrt(bias_varC))
ChoiceqmfitC05 <- qnorm(0.05,ChosenAvgLine,sqrt(bias_varC))
#Choice D
qmfitD <- c(ChosenAvgLine[dnum1] + 0.7, ChosenAvgLine[dnum2] + 0.35, ChosenAvgLine[dnum3] + 1.2)
osolD <- optim(par=inits1, ofn_q95, q95 = qmfitD, ChosenAvgLine=mean_line())
bias_varD <- exp(osolD$par[1])^2 + (dose^2)*(osolD$par[2]^2 + exp(osolD$par[3])^2) + 2*dose*exp(osolD$par[1])*osolD$par[2]
ChoiceqmfitD95 <- qnorm(0.95,ChosenAvgLine,sqrt(bias_varD))
ChoiceqmfitD05 <- qnorm(0.05,ChosenAvgLine,sqrt(bias_varD))
par(mfrow=c(3,2),oma = c(0, 0, 2, 0))
#Uncertainty shape A
plot(dose,back_trans(ChosenAvgLine),type='l',col='red',lty=1,lwd=2,main='A',cex.main=3,
ylim=c(-100,80),cex.lab=2,cex.axis=1.8,xlab='Dose (mg/kg/day)',ylab='% change')
lines(dose,back_trans(ChoiceqmfitA95),type='l',col='darkgreen',lty=2,lwd=2)
lines(dose,back_trans(ChoiceqmfitA05),type='l',col='darkgreen',lty=2,lwd=2)
#Uncertainty shape B
plot(dose,back_trans(ChosenAvgLine),type='l',col='red',lty=1,lwd=2,main='B',cex.main=3,
ylim=c(-100,80),cex.lab=2,cex.axis=1.8,xlab='Dose (mg/kg/day)',ylab='% change')
lines(dose,back_trans(ChoiceqmfitB95),type='l',col='darkgreen',lty=2,lwd=2)
lines(dose,back_trans(ChoiceqmfitB05),type='l',col='darkgreen',lty=2,lwd=2)
#Uncertainty shape C
plot(dose,back_trans(ChosenAvgLine),type='l',col='red',lty=1,lwd=2,main='C',cex.main=3,
ylim=c(-100,80),cex.lab=2,cex.axis=1.8,xlab='Dose (mg/kg/day)',ylab='% change')
lines(dose,back_trans(ChoiceqmfitC95),type='l',col='darkgreen',lty=2,lwd=2)
lines(dose,back_trans(ChoiceqmfitC05),type='l',col='darkgreen',lty=2,lwd=2)
#Uncertainty shape D
plot(dose,back_trans(ChosenAvgLine),type='l',col='red',lty=1,lwd=2,main='D',cex.main=3,
ylim=c(-100,80),cex.lab=2,cex.axis=1.8,xlab='Dose (mg/kg/day)',ylab='% change')
lines(dose,back_trans(ChoiceqmfitD95),type='l',col='darkgreen',lty=2,lwd=2)
lines(dose,back_trans(ChoiceqmfitD05),type='l',col='darkgreen',lty=2,lwd=2)
#Legend
plot(1, type="n", axes=F, xlab="", ylab="")
legend("topleft",cex=2,col=c('red','darkgreen'),lty=c(1,2),
c('Your best guess at the dose-response relationship in children',
'Lines indicating your 90% credibility interval around the line of best fit'))
}
ShapePlot2_plot <- reactive({
pdf(paste(OUT_dir,input$expert,"_uncertShapePlot.pdf",sep=''),pointsize=10,paper="a4")
ShapePlot2_func()
dev.off()
})
output$ShapePlot2 <- renderPlot({
ShapePlot2_func()
ShapePlot2_plot()
}, height = 900, width = 1200)
ChoicePlotqmfit05_reactive <- reactive({
ChosenAvgLine <- mean_line()
Chosenosol <- reactiveOptim()
Chosenbias_var <- exp(Chosenosol$par[1])^2 + (dose^2)*(Chosenosol$par[2]^2 + exp(Chosenosol$par[3])^2) + 2*dose*exp(Chosenosol$par[1])*Chosenosol$par[2]
ChoicePlotqmfit05 <- qnorm(0.05,ChosenAvgLine,sqrt(Chosenbias_var))
})
ChoicePlotqmfit95_reactive <- reactive({
ChosenAvgLine <- mean_line()
Chosenosol <- reactiveOptim()
Chosenbias_var <- exp(Chosenosol$par[1])^2 + (dose^2)*(Chosenosol$par[2]^2 + exp(Chosenosol$par[3])^2) + 2*dose*exp(Chosenosol$par[1])*Chosenosol$par[2]
ChoicePlotqmfit95 <- qnorm(0.95,ChosenAvgLine,sqrt(Chosenbias_var))
})
############# HISTOGRAMS #########~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~##########
# Histogram and percentiles for dose 0
FreqHistInput0 <- reactive({
FreqHistInput1 <- c(input$n100_n95_0,input$n95_n90_0,input$n90_n85_0,input$n85_n80_0,input$n80_n75_0,input$n75_n70_0,
input$n70_n65_0,input$n65_n60_0,input$n60_n55_0,input$n55_n50_0,input$n50_n45_0,input$n45_n40_0,
input$n40_n35_0,input$n35_n30_0,input$n30_n25_0,input$n25_n20_0,input$n20_n15_0,input$n15_n10_0,
input$n10_n05_0,input$n05_n00_0,input$n00_p05_0,input$p05_p10_0,input$p10_p15_0,input$p15_p20_0,
input$p20_p25_0,input$p25_p30_0,input$p30_p35_0,input$p35_p40_0,input$p40_p45_0,input$p45_p50_0,
input$p50_p55_0,input$p55_p60_0,input$p60_p65_0,input$p65_p70_0,input$p70_p75_0,input$p75_p80_0,
input$p80_p85_0,input$p85_p90_0,input$p90_p95_0,input$p95_p100_0)
return(FreqHistInput1)
})
HistInput0 <- reactive({
HistInput1 <- NULL
for(i in 1:length(FreqHistInput0())){
HistInput1 <- c(HistInput1,rep(center_bin[i],FreqHistInput0()[i]))
}
return(HistInput1)
})
HistPlot0_func <- function(){
ChosenAvgLine <- mean_line()
Chosenosol <- reactiveOptim()
Chosenbias_var <- exp(Chosenosol$par[1])^2 + (dose^2)*(Chosenosol$par[2]^2 + exp(Chosenosol$par[3])^2) + 2*dose*exp(Chosenosol$par[1])*Chosenosol$par[2]
ChoicePlotqmfit95 <- qnorm(0.95,ChosenAvgLine,sqrt(Chosenbias_var))
ChoicePlotqmfit05 <- qnorm(0.05,ChosenAvgLine,sqrt(Chosenbias_var))
par(mar=c(11,5,1.1,2.1), xpd=TRUE)
hist(HistInput0(),breaks=bins,ylim=c(0,max(50, max(freqBins1()))),axes=F,col="lightgrey",main='',
xlab='% change in seizure frequency from baseline',ylab='% chance of average response',xaxs='i')
axis(2,at=seq(0,max(50, max(freqBins1())),5))
axis(1, at=seq(-100,100,binwidth))
arrows(-1, -16, -100, -16, lwd=2)
arrows(1, -16, 100, -16, lwd=2)
text(-35, -19, labels = "Improvement in condition", cex=1.5, col="blue")
text(35, -19, labels = "Deterioration in condition", cex=1.5, col="red")
par(mar=c(9,5,1.1,2.1), xpd=F)
abline(h=(1:max(50, max(freqBins1())))*5, v=seq(-100,100,binwidth), col="gray", lty=3)
abline(v=round(back_trans(mean_line()[dnum1]),digits=1),lty=1,col='red',lwd=2)
abline(v=round(back_trans(ChoicePlotqmfit95[dnum1]),digits=1),lty=1,col='green',lwd=2)
abline(v=round(back_trans(ChoicePlotqmfit05[dnum1]),digits=1),lty=1,col='green',lwd=2)
}
HistPlot0_plot <- reactive({
pdf(paste(OUT_dir,input$expert,"_hist0Plot.pdf",sep=''),pointsize=10,paper="a4")
HistPlot0_func()
dev.off()
})
HistDensPlot0_func <- function(){
par(mar=c(11,5,1.1,2.1), xpd=TRUE)
Dens1 <- density(rep(HistInput0(),5),adjust=1.8)
hist(rep(HistInput0(),5),breaks=bins,col="lightgrey",main='',ylim=c(0,max(Dens1$y)),
xlab='% change in seizure frequency from baseline',ylab='Density',freq=F,xaxs='i')
lines(Dens1)
axis(1, at=seq(-100,100,binwidth))
par(mar=c(9,5,1.1,2.1), xpd=F)
abline(v=round(back_trans(mean_line()[dnum1]),digits=1),lty=1,col='red',lwd=2)
}
HistDensPlot0_plot <- reactive({
pdf(paste(OUT_dir,input$expert,"_histDens0Plot.pdf",sep=''),pointsize=10,paper="a4")
HistDensPlot0_func()
dev.off()
})
output$HistPlot0 <- renderPlot({
if(input$Tick_DensHist0==FALSE){
HistPlot0_func()
HistPlot0_plot()
}else if(input$Tick_DensHist0==TRUE){
HistDensPlot0_func()
HistDensPlot0_plot()
}
}, height = 600, width = 1900)
pctiles1 <- reactive({
v1_choice <- HistInput0()
freq1 <- NULL
for(i in 1:length(bins1)){
freq1[i] <- sum(v1_choice > bins1[i] & v1_choice < bins2[i])
}
numin05_1 <- 0.05*sum(freq1)
numin25_1 <- 0.25*sum(freq1)
numin75_1 <- 0.75*sum(freq1)
numin95_1 <- 0.95*sum(freq1)
low_05_1 <- bins1[min(which(cumsum(freq1/sum(freq1))>0.05))]
low_25_1 <- bins1[min(which(cumsum(freq1/sum(freq1))>0.25))]
low_75_1 <- bins1[min(which(cumsum(freq1/sum(freq1))>0.75))]
low_95_1 <- bins1[min(which(cumsum(freq1/sum(freq1))>0.95))]
sum_upto05_1 <- sum(freq1[c(which(cumsum(freq1/sum(freq1))<0.05))])
sum_upto25_1 <- sum(freq1[c(which(cumsum(freq1/sum(freq1))<0.25))])
sum_upto75_1 <- sum(freq1[c(which(cumsum(freq1/sum(freq1))<0.75))])
sum_upto95_1 <- sum(freq1[c(which(cumsum(freq1/sum(freq1))<0.95))])
bins_freq1 <- NULL
for(i in 1:length(freq1)){
bins_freq1 <- c(bins_freq1,rep(center_bin[i],freq1[i]))
}
pctiles_1 <- c(low_05_1 + (numin05_1 - sum_upto05_1)*binwidth/freq1[min(which(cumsum(freq1/sum(freq1))>0.05))],
low_25_1 + (numin25_1 - sum_upto25_1)*binwidth/freq1[min(which(cumsum(freq1/sum(freq1))>0.25))],
low_75_1 + (numin75_1 - sum_upto75_1)*binwidth/freq1[min(which(cumsum(freq1/sum(freq1))>0.75))],
low_95_1 + (numin95_1 - sum_upto95_1)*binwidth/freq1[min(which(cumsum(freq1/sum(freq1))>0.95))])
return(pctiles_1)
})
freqBins1 <- reactive({
v1_choice <- HistInput0()
freq1 <- NULL
for(i in 1:length(bins1)){
freq1[i] <- sum(v1_choice > bins1[i] & v1_choice < bins2[i])
}
return(freq1)
})
output$Meantext1 <- renderText({
paste("Your best guess, as shown by the red line, at the average response in children on placebo is ",
round(back_trans(mean_line()[dnum1]),digits=1),".",sep='')
})
output$Uncerttext1 <- renderText({
paste("Your best guesses, as shown by the green lines, at the 5th and 95th percentiles of the average response in children on placebo are ",
round(back_trans(ChoicePlotqmfit05_reactive()[dnum1]),digits=1)," and ",round(back_trans(ChoicePlotqmfit95_reactive()[dnum1]),digits=1),". You should assign 90% chance between these lines.",sep='')
})
# Histogram and percentiles for moderate dose
FreqHistInputMid <- reactive({
FreqHistInput2 <- c(input$n100_n95_Mid,input$n95_n90_Mid,input$n90_n85_Mid,input$n85_n80_Mid,input$n80_n75_Mid,input$n75_n70_Mid,
input$n70_n65_Mid,input$n65_n60_Mid,input$n60_n55_Mid,input$n55_n50_Mid,input$n50_n45_Mid,input$n45_n40_Mid,
input$n40_n35_Mid,input$n35_n30_Mid,input$n30_n25_Mid,input$n25_n20_Mid,input$n20_n15_Mid,input$n15_n10_Mid,
input$n10_n05_Mid,input$n05_n00_Mid,input$n00_p05_Mid,input$p05_p10_Mid,input$p10_p15_Mid,input$p15_p20_Mid,
input$p20_p25_Mid,input$p25_p30_Mid,input$p30_p35_Mid,input$p35_p40_Mid,input$p40_p45_Mid,input$p45_p50_Mid,
input$p50_p55_Mid,input$p55_p60_Mid,input$p60_p65_Mid,input$p65_p70_Mid,input$p70_p75_Mid,input$p75_p80_Mid,
input$p80_p85_Mid,input$p85_p90_Mid,input$p90_p95_Mid,input$p95_p100_Mid)
return(FreqHistInput2)
})
HistInputMid <- reactive({
HistInput2 <- NULL
for(i in 1:length(FreqHistInputMid())){
HistInput2 <- c(HistInput2,rep(center_bin[i],FreqHistInputMid()[i]))
}
return(HistInput2)
})
HistPlotMid_func <- function(){
ChosenAvgLine <- mean_line()
Chosenosol <- reactiveOptim()
Chosenbias_var <- exp(Chosenosol$par[1])^2 + (dose^2)*(Chosenosol$par[2]^2 + exp(Chosenosol$par[3])^2) + 2*dose*exp(Chosenosol$par[1])*Chosenosol$par[2]
ChoicePlotqmfit95 <- qnorm(0.95,ChosenAvgLine,sqrt(Chosenbias_var))
ChoicePlotqmfit05 <- qnorm(0.05,ChosenAvgLine,sqrt(Chosenbias_var))
par(mar=c(11,5,1.1,2.1), xpd=TRUE)
hist(HistInputMid(),breaks=bins,ylim=c(0,max(50, max(freqBins2()))),axes=F,col="lightgrey",main='',
xlab='% change in seizure frequency from baseline',ylab='% chance of average response',xaxs='i')
axis(2,at=seq(0,max(50, max(freqBins2())),5))
axis(1, at=seq(-100,100,binwidth))
arrows(-1, -16, -100, -16, lwd=2)
arrows(1, -16, 100, -16, lwd=2)
text(-35, -19, labels = "Improvement in condition", cex=1.5, col="blue")
text(35, -19, labels = "Deterioration in condition", cex=1.5, col="red")
par(mar=c(9,5,1.1,2.1), xpd=F)
abline(h=(1:max(50, max(freqBins2())))*5, v=seq(-100,100,binwidth), col="gray", lty=3)
abline(v=round(back_trans(mean_line()[dnum2]),digits=1),lty=1,col='red',lwd=2)
abline(v=round(back_trans(ChoicePlotqmfit95[dnum2]),digits=1),lty=1,col='green',lwd=2)
abline(v=round(back_trans(ChoicePlotqmfit05[dnum2]),digits=1),lty=1,col='green',lwd=2)
}
HistPlotMid_plot <- reactive({
pdf(paste(OUT_dir,input$expert,"_histMidPlot.pdf",sep=''),pointsize=10,paper="a4")
HistPlotMid_func()
dev.off()
})
HistDensPlotMid_func <- function(){
par(mar=c(11,5,1.1,2.1), xpd=TRUE)
Dens2 <- density(rep(HistInputMid(),5),adjust=1.8)
hist(rep(HistInputMid(),5),breaks=bins,col="lightgrey",main='',ylim=c(0,max(Dens2$y)),
xlab='% change in seizure frequency from baseline',ylab='Density',freq=F,xaxs='i')
lines(Dens2)
axis(1, at=seq(-100,100,binwidth))
par(mar=c(9,5,1.1,2.1), xpd=F)
abline(v=round(back_trans(mean_line()[dnum2]),digits=1),lty=1,col='red',lwd=2)
}
HistDensPlotMid_plot <- reactive({
pdf(paste(OUT_dir,input$expert,"_histDensMidPlot.pdf",sep=''),pointsize=10,paper="a4")
HistDensPlotMid_func()
dev.off()
})
output$HistPlotMid <- renderPlot({
if(input$Tick_DensHistMid==FALSE){
HistPlotMid_func()
HistPlotMid_plot()
}else if(input$Tick_DensHistMid==TRUE){
HistDensPlotMid_func()
HistDensPlotMid_plot()
}
}, height = 600, width = 1900)
pctiles2 <- reactive({
v2_choice <- HistInputMid()
freq2 <- NULL
for(i in 1:length(bins1)){
freq2[i] <- sum(v2_choice > bins1[i] & v2_choice < bins2[i])
}
numin05_2 <- 0.05*sum(freq2)
numin25_2 <- 0.25*sum(freq2)
numin75_2 <- 0.75*sum(freq2)
numin95_2 <- 0.95*sum(freq2)
low_05_2 <- bins1[min(which(cumsum(freq2/sum(freq2))>0.05))]
low_25_2 <- bins1[min(which(cumsum(freq2/sum(freq2))>0.25))]
low_75_2 <- bins1[min(which(cumsum(freq2/sum(freq2))>0.75))]
low_95_2 <- bins1[min(which(cumsum(freq2/sum(freq2))>0.95))]
sum_upto05_2 <- sum(freq2[c(which(cumsum(freq2/sum(freq2))<0.05))])
sum_upto25_2 <- sum(freq2[c(which(cumsum(freq2/sum(freq2))<0.25))])
sum_upto75_2 <- sum(freq2[c(which(cumsum(freq2/sum(freq2))<0.75))])
sum_upto95_2 <- sum(freq2[c(which(cumsum(freq2/sum(freq2))<0.95))])
bins_freq2 <- NULL
for(i in 1:length(freq2)){
bins_freq2 <- c(bins_freq2,rep(center_bin[i],freq2[i]))
}
pctiles_2 <- c(low_05_2 + (numin05_2 - sum_upto05_2)*binwidth/freq2[min(which(cumsum(freq2/sum(freq2))>0.05))],
low_25_2 + (numin25_2 - sum_upto25_2)*binwidth/freq2[min(which(cumsum(freq2/sum(freq2))>0.25))],
low_75_2 + (numin75_2 - sum_upto75_2)*binwidth/freq2[min(which(cumsum(freq2/sum(freq2))>0.75))],
low_95_2 + (numin95_2 - sum_upto95_2)*binwidth/freq2[min(which(cumsum(freq2/sum(freq2))>0.95))])
return(pctiles_2)
})
freqBins2 <- reactive({
v2_choice <- HistInputMid()
freq2 <- NULL
for(i in 1:length(bins1)){
freq2[i] <- sum(v2_choice > bins1[i] & v2_choice < bins2[i])
}
return(freq2)
})
output$Meantext2 <- renderText({
paste("Your best guess, as shown by the red line, at the average response in children on 8 mg/kg/day is ",
round(back_trans(mean_line()[dnum2]),digits=1),".",sep='')
})
output$Uncerttext2 <- renderText({
paste("Your best guesses, as shown by the green lines, at the 5th and 95th percentiles of the average response in children on 8 mg/kg/day are ",
round(back_trans(ChoicePlotqmfit05_reactive()[dnum2]),digits=1)," and ",round(back_trans(ChoicePlotqmfit95_reactive()[dnum2]),digits=1),". You should assign 90% chance between these lines.",sep='')
})
# Histogram and percentiles for high dose
FreqHistInputHigh <- reactive({
FreqHistInput3 <- c(input$n100_n95_High,input$n95_n90_High,input$n90_n85_High,input$n85_n80_High,input$n80_n75_High,input$n75_n70_High,
input$n70_n65_High,input$n65_n60_High,input$n60_n55_High,input$n55_n50_High,input$n50_n45_High,input$n45_n40_High,
input$n40_n35_High,input$n35_n30_High,input$n30_n25_High,input$n25_n20_High,input$n20_n15_High,input$n15_n10_High,
input$n10_n05_High,input$n05_n00_High,input$n00_p05_High,input$p05_p10_High,input$p10_p15_High,input$p15_p20_High,
input$p20_p25_High,input$p25_p30_High,input$p30_p35_High,input$p35_p40_High,input$p40_p45_High,input$p45_p50_High,
input$p50_p55_High,input$p55_p60_High,input$p60_p65_High,input$p65_p70_High,input$p70_p75_High,input$p75_p80_High,
input$p80_p85_High,input$p85_p90_High,input$p90_p95_High,input$p95_p100_High)
return(FreqHistInput3)
})
HistInputHigh <- reactive({
HistInput3 <- NULL
for(i in 1:length(FreqHistInputHigh())){
HistInput3 <- c(HistInput3,rep(center_bin[i],FreqHistInputHigh()[i]))
}