-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy path06_Additional_Analyses.R
More file actions
327 lines (278 loc) · 19.5 KB
/
06_Additional_Analyses.R
File metadata and controls
327 lines (278 loc) · 19.5 KB
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
##%######################################################%##
# #
#### 06 ADDITIONAL ANALYSES ####
# #
##%######################################################%##
# Analyses for:
# - sensitivity to change (pre-post)
# - RCIs
source("/Users/tim/NextcloudHGW/MEDI-Validierung/MEDI_German_Validation/MEDI_German_Validation_Study/medi_to_sedi.R")
# source("medi_to_sedi.R")
library(psych)
library(dplyr)
library(rciplot)
library(tidyr)
library(pROC)
library(gam)
alldata <- readRDS("/Users/tim/Documents/GitHub/new-explorer/caches/allData.RDS")
# alldata <- readRDS("C:/Users/steph/OneDrive/Desktop/Psychologie/Doktorat eHelp-MV-Projekt/GPNS/new-explorer/caches/allData.RDS")
alldata_medi_lv_scores <- readRDS("Data/alldata_medi_lv_scores.RDS") # validation sample (including only first session)
alldata_medi_lv_scores$sedi <- medi_to_sedi(alldata_medi_lv_scores)
alldata_medi_lv_scores$is.patient <- alldata_medi_lv_scores$sample == "patient"
#sedi.patient.roc <- pROC::roc(alldata_medi_lv_scores$is.patient ~ predict(glm(is.patient ~ sedi, data = alldata_medi_lv_scores, family = "binomial")))
#coords(sedi.patient.roc, "best", ret = c("threshold","ppv","npv"),
# best.method = "closest.topleft", best.weights = c(1, .25))
alldata$neurotic <- alldata$neurotic * 5
alldata$positive <- alldata$positive * 5
alldata$depression <- alldata$depression * 5
alldata$arousal <- alldata$arousal * 5
alldata$somatic <- alldata$somatic * 5
alldata$social <- alldata$social * 5
alldata$intrusive <- alldata$intrusive * 6
alldata$traumatic <- alldata$traumatic * 5
alldata$avoidance <- alldata$avoidance * 8
#-------- Sensitivity to Change in Pre-Post Data --------
# SEDI
#alldata$sedi <- medi_to_sedi(alldata)
# must be done with the data from the study, not with the current ZPP data:
#group_by(alldata, sample) %>% summarise(mSEDI = mean(sedi), sdSEDI = sd(sedi))
#criterion_c <- ((2.34*0.472) + (1.87*0.472)) / (.56 + .472)
#criterion_c
# select patient code, date, type of measurement, MEDI scales, and KODAP "problem" item
dat.pre.post <- alldata %>%
filter(type == "pre" | type == "post") %>%
dplyr::select(kkcode, type, ended, neurotic, positive, depression, arousal, somatic, intrusive, social, traumatic, avoidance, KPPost_problem, KPPost_change) %>%
group_by(kkcode, type) %>%
arrange(ended) %>%
# filter(row_number() == 1) %>%
ungroup()
dat.pre.post <- pivot_wider(dat.pre.post, id_cols = "kkcode", names_from = "type", values_from = colnames(dat.pre.post)[3:14])
dat.pre.post$neurotic_change <- 1 - dat.pre.post$neurotic_post / dat.pre.post$neurotic_pre
dat.pre.post$positive_change <- 1 - dat.pre.post$positive_post / dat.pre.post$positive_pre
dat.pre.post$depression_change <- 1 - dat.pre.post$depression_post / dat.pre.post$depression_pre
dat.pre.post$arousal_change <- 1 - dat.pre.post$arousal_post / dat.pre.post$arousal_pre
dat.pre.post$somatic_change <- 1 - dat.pre.post$somatic_post / dat.pre.post$somatic_pre
dat.pre.post$intrusive_change <- 1 - dat.pre.post$intrusive_post / dat.pre.post$intrusive_pre
dat.pre.post$social_change <- 1 - dat.pre.post$social_post / dat.pre.post$social_pre
dat.pre.post$traumatic_change <- 1 - dat.pre.post$traumatic_post / dat.pre.post$traumatic_pre
dat.pre.post$avoidance_change <- 1 - dat.pre.post$avoidance_post / dat.pre.post$avoidance_pre
dat.pre.post$neurotic_change[which(is.infinite(dat.pre.post$neurotic_change))] <- NA
dat.pre.post$positive_change[which(is.infinite(dat.pre.post$positive_change))] <- NA
dat.pre.post$depression_change[which(is.infinite(dat.pre.post$depression_change))] <- NA
dat.pre.post$arousal_change[which(is.infinite(dat.pre.post$arousal_change))] <- NA
dat.pre.post$somatic_change[which(is.infinite(dat.pre.post$somatic_change))] <- NA
dat.pre.post$intrusive_change[which(is.infinite(dat.pre.post$intrusive_change))] <- NA
dat.pre.post$social_change[which(is.infinite(dat.pre.post$social_change))] <- NA
dat.pre.post$traumatic_change[which(is.infinite(dat.pre.post$traumatic_change))] <- NA
dat.pre.post$avoidance_change[which(is.infinite(dat.pre.post$avoidance_change))] <- NA
dat.pre.post$neurotic_reduction <- dat.pre.post$neurotic_post - dat.pre.post$neurotic_pre
dat.pre.post$positive_reduction <- dat.pre.post$positive_post - dat.pre.post$positive_pre
dat.pre.post$depression_reduction <- dat.pre.post$depression_post - dat.pre.post$depression_pre
dat.pre.post$arousal_reduction <- dat.pre.post$arousal_post - dat.pre.post$arousal_pre
dat.pre.post$somatic_reduction <- dat.pre.post$somatic_post - dat.pre.post$somatic_pre
dat.pre.post$intrusive_reduction <- dat.pre.post$intrusive_post - dat.pre.post$intrusive_pre
dat.pre.post$social_reduction <- dat.pre.post$social_post - dat.pre.post$social_pre
dat.pre.post$traumatic_reduction <- dat.pre.post$traumatic_post - dat.pre.post$traumatic_pre
dat.pre.post$avoidance_reduction <- dat.pre.post$avoidance_post - dat.pre.post$avoidance_pre
# correlation of anchors with points reduction:
cor(dat.pre.post$neurotic_reduction, dat.pre.post$KPPost_problem_post, use = "complete", method = "spearman")
cor(dat.pre.post$depression_reduction, dat.pre.post$KPPost_problem_post, use = "complete", method = "spearman")
cor(dat.pre.post$arousal_reduction, dat.pre.post$KPPost_problem_post, use = "complete", method = "spearman")
cor(dat.pre.post$somatic_reduction, dat.pre.post$KPPost_problem_post, use = "complete", method = "spearman")
cor(dat.pre.post$intrusive_reduction, dat.pre.post$KPPost_problem_post, use = "complete", method = "spearman")
cor(dat.pre.post$traumatic_reduction, dat.pre.post$KPPost_problem_post, use = "complete", method = "spearman")
cor(dat.pre.post$avoidance_reduction, dat.pre.post$KPPost_problem_post, use = "complete", method = "spearman")
cor(dat.pre.post$neurotic_reduction, dat.pre.post$KPPost_change_post, use = "complete", method = "spearman")
cor(dat.pre.post$depression_reduction, dat.pre.post$KPPost_change_post, use = "complete", method = "spearman")
cor(dat.pre.post$arousal_reduction, dat.pre.post$KPPost_change_post, use = "complete", method = "spearman")
cor(dat.pre.post$somatic_reduction, dat.pre.post$KPPost_change_post, use = "complete", method = "spearman")
cor(dat.pre.post$intrusive_reduction, dat.pre.post$KPPost_change_post, use = "complete", method = "spearman")
cor(dat.pre.post$traumatic_reduction, dat.pre.post$KPPost_change_post, use = "complete", method = "spearman")
cor(dat.pre.post$avoidance_reduction, dat.pre.post$KPPost_change_post, use = "complete", method = "spearman")
# correlation of anchors with percentage reduction:
cor(dat.pre.post$neurotic_change, dat.pre.post$KPPost_problem_post, use = "complete", method = "spearman")
cor(dat.pre.post$depression_change, dat.pre.post$KPPost_problem_post, use = "complete", method = "spearman")
cor(dat.pre.post$arousal_change, dat.pre.post$KPPost_problem_post, use = "complete", method = "spearman")
cor(dat.pre.post$somatic_change, dat.pre.post$KPPost_problem_post, use = "complete", method = "spearman")
cor(dat.pre.post$intrusive_change, dat.pre.post$KPPost_problem_post, use = "complete", method = "spearman")
cor(dat.pre.post$traumatic_change, dat.pre.post$KPPost_problem_post, use = "complete", method = "spearman")
cor(dat.pre.post$avoidance_change, dat.pre.post$KPPost_problem_post, use = "complete", method = "spearman")
cor(dat.pre.post$neurotic_change, dat.pre.post$KPPost_change_post, use = "complete", method = "spearman")
cor(dat.pre.post$depression_change, dat.pre.post$KPPost_change_post, use = "complete", method = "spearman")
cor(dat.pre.post$arousal_change, dat.pre.post$KPPost_change_post, use = "complete", method = "spearman")
cor(dat.pre.post$somatic_change, dat.pre.post$KPPost_change_post, use = "complete", method = "spearman")
cor(dat.pre.post$intrusive_change, dat.pre.post$KPPost_change_post, use = "complete", method = "spearman")
cor(dat.pre.post$traumatic_change, dat.pre.post$KPPost_change_post, use = "complete", method = "spearman")
cor(dat.pre.post$avoidance_change, dat.pre.post$KPPost_change_post, use = "complete", method = "spearman")
# simple change score analysis
library(mgcv)
dat.pre.post$KPPost_problem.simple <- ifelse(dat.pre.post$KPPost_problem_post <= 2, 1, 0)
neurotic.gam <- gam(KPPost_problem.simple ~ s(neurotic_change) + s(neurotic_pre), method = "REML", select = TRUE, data = dat.pre.post, family = binomial(link = "logit"))
positive.gam <- gam(KPPost_problem.simple ~ s(positive_change) + s(positive_pre), method = "REML", select = TRUE, data = dat.pre.post, family = binomial(link = "logit"))
depression.gam <- gam(KPPost_problem.simple ~ s(depression_change) + s(depression_pre), method = "REML", select = TRUE, data = dat.pre.post, family = binomial(link = "logit"))
arousal.gam <- gam(KPPost_problem.simple ~ s(arousal_change) + s(arousal_pre), method = "REML", select = TRUE, data = dat.pre.post, family = binomial(link = "logit"))
somatic.gam <- gam(KPPost_problem.simple ~ s(somatic_change) + s(somatic_pre), method = "REML", select = TRUE, data = dat.pre.post, family = binomial(link = "logit"))
intrusive.gam <- gam(KPPost_problem.simple ~ s(intrusive_change) + s(intrusive_pre), method = "REML", select = TRUE, data = dat.pre.post, family = binomial(link = "logit"))
social.gam <- gam(KPPost_problem.simple ~ s(social_change) + s(social_pre), method = "REML", select = TRUE, data = dat.pre.post, family = binomial(link = "logit"))
traumatic.gam <- gam(KPPost_problem.simple ~ s(traumatic_change) + s(traumatic_pre), method = "REML", select = TRUE, data = dat.pre.post, family = binomial(link = "logit"))
avoidance.gam <- gam(KPPost_problem.simple ~ s(avoidance_change) + s(avoidance_pre), method = "REML", select = TRUE, data = dat.pre.post, family = binomial(link = "logit"))
return.criteria <- c("threshold","sensitivity","specificity","ppv","npv","auc")
coords(roc(KPPost_problem.simple ~ neurotic_change, data = dat.pre.post), "best", ret = return.criteria, method = "closest.topleft")
coords(roc(KPPost_problem.simple ~ positive_change, data = dat.pre.post), "best", ret = return.criteria, method = "closest.topleft")
coords(roc(KPPost_problem.simple ~ depression_change, data = dat.pre.post), "best", ret = return.criteria, method = "closest.topleft")
coords(roc(KPPost_problem.simple ~ arousal_change, data = dat.pre.post), "best", ret = return.criteria, method = "closest.topleft")
coords(roc(KPPost_problem.simple ~ somatic_change, data = dat.pre.post), "best", ret = return.criteria, method = "closest.topleft")
coords(roc(KPPost_problem.simple ~ intrusive_change, data = dat.pre.post), "best", ret = return.criteria, method = "closest.topleft")
coords(roc(KPPost_problem.simple ~ social_change, data = dat.pre.post), "best", ret = return.criteria, method = "closest.topleft")
coords(roc(KPPost_problem.simple ~ traumatic_change, data = dat.pre.post), "best", ret = return.criteria, method = "closest.topleft")
coords(roc(KPPost_problem.simple ~ avoidance_change, data = dat.pre.post), "best", ret = return.criteria, method = "closest.topleft")
coords(roc(KPPost_problem.simple ~ rowSums(dat.pre.post[,45:53]), data = dat.pre.post), "best")
roc(KPPost_problem.simple ~ neurotic_change, data = dat.pre.post)
roc(KPPost_problem.simple ~ positive_change, data = dat.pre.post)
roc(KPPost_problem.simple ~ depression_change, data = dat.pre.post)
roc(KPPost_problem.simple ~ arousal_change, data = dat.pre.post)
roc(KPPost_problem.simple ~ somatic_change, data = dat.pre.post)
roc(KPPost_problem.simple ~ intrusive_change, data = dat.pre.post)
roc(KPPost_problem.simple ~ social_change, data = dat.pre.post)
roc(KPPost_problem.simple ~ traumatic_change, data = dat.pre.post)
roc(KPPost_problem.simple ~ avoidance_change, data = dat.pre.post)
dat.pre.post$neurotic.acc <- dat.pre.post$neurotic_change >= .366
dat.pre.post$positive.acc <- dat.pre.post$positive_change <= -.106
dat.pre.post$depression.acc <- dat.pre.post$depression_change >= .592
dat.pre.post$arousal.acc <- dat.pre.post$arousal_change >= .634
dat.pre.post$somatic.acc <- dat.pre.post$somatic_change >= .222
dat.pre.post$intrusive.acc <- dat.pre.post$intrusive_change >= .455
dat.pre.post$social.acc <- dat.pre.post$social_change >= .212
dat.pre.post$traumatic.acc <- dat.pre.post$traumatic_change >= .567
dat.pre.post$avoidance.acc <- dat.pre.post$avoidance_change >= .508
prop.table(table(dat.pre.post$neurotic.acc | dat.pre.post$positive.acc | dat.pre.post$depression.acc | dat.pre.post$arousal.acc | dat.pre.post$somatic.acc | dat.pre.post$intrusive.acc | dat.pre.post$social.acc | dat.pre.post$traumatic.acc | dat.pre.post$avoidance.acc))
# equipercentile linking
library(equate)
eq.neurotic <- equate(as.freqtab(table(dat.pre.post$KPPost_problem_post, dat.pre.post$neurotic_change)), type = "equip", smoothmethod = "bump", jmin = 1)
eq.positive <- equate(as.freqtab(table(dat.pre.post$KPPost_problem_post, dat.pre.post$positive_change)), type = "equip", smoothmethod = "bump", jmin = 1)
eq.depression <- equate(as.freqtab(table(dat.pre.post$KPPost_problem_post, dat.pre.post$depression_change)), type = "equip", smoothmethod = "bump", jmin = 1)
eq.arousal <- equate(as.freqtab(table(dat.pre.post$KPPost_problem_post, dat.pre.post$arousal_change)), type = "equip", smoothmethod = "bump", jmin = 1)
eq.somatic <- equate(as.freqtab(table(dat.pre.post$KPPost_problem_post, dat.pre.post$somatic_change)), type = "equip", smoothmethod = "bump", jmin = 1)
eq.intrusive <- equate(as.freqtab(table(dat.pre.post$KPPost_problem_post, dat.pre.post$intrusive_change)), type = "equip", smoothmethod = "bump", jmin = 1)
eq.social <- equate(as.freqtab(table(dat.pre.post$KPPost_problem_post, dat.pre.post$social_change)), type = "equip", smoothmethod = "bump", jmin = 1)
eq.traumatic <- equate(as.freqtab(table(dat.pre.post$KPPost_problem_post, dat.pre.post$traumatic_change)), type = "equip", smoothmethod = "bump", jmin = 1)
eq.avoidance <- equate(as.freqtab(table(dat.pre.post$KPPost_problem_post, dat.pre.post$avoidance_change)), type = "equip", smoothmethod = "bump", jmin = 1)
# raw summary
group_by(dat.pre.post, KPPost_problem.simple) %>%
summarise(mChange.neur = mean(neurotic_change, na.rm = T),
mChange.dep = mean(depression_change, na.rm = T),
mChange.pos = mean(positive_change, na.rm = T),
mChange.aro = mean(arousal_change, na.rm = T),
mChange.som = mean(somatic_change, na.rm = T),
mChange.int = mean(intrusive_change, na.rm = T),
mChange.soc = mean(social_change, na.rm = T),
mChange.tra = mean(traumatic_change, na.rm = T),
mChange.avoid = mean(avoidance_change, na.rm = T),
n = n())
# get concordance tables
eq.neurotic$concordance
eq.positive$concordance
eq.depression$concordance
eq.arousal$concordance
eq.somatic$concordance
eq.intrusive$concordance
eq.social$concordance
eq.traumatic$concordance
eq.avoidance$concordance
# get group-wise summary statistics
group_by(alldata_medi_lv_scores, sample) %>%
summarise(
mNeurotic = mean(neurotic),
sdNeurotic = sd(neurotic),
mPositive = mean(positive),
sdPositive = sd(positive),
mDepression = mean(depression),
sdDepression = sd(depression),
mArousal = mean(arousal),
sdArousal = sd(arousal),
mSomatic = mean(somatic),
sdSomatic = sd(somatic),
mSocial = mean(social),
sdSocial = sd(social),
mIntrusive = mean(intrusive),
sdIntrusive = sd(intrusive),
mTraumatic = mean(traumatic),
sdTraumatic = sd(traumatic),
mAvoidance = mean(avoidance),
sdAvoidance = sd(avoidance)
) -> group.scores
# manual computation of Criterion C (midpoint of clinical and community sample)
round((group.scores$sdNeurotic[1]*group.scores$mNeurotic[2] + group.scores$sdNeurotic[2]*group.scores$mNeurotic[1]) / sum(group.scores$sdNeurotic)) # 18
round((group.scores$sdPositive[1]*group.scores$mPositive[2] + group.scores$sdPositive[2]*group.scores$mPositive[1]) / sum(group.scores$sdPositive)) # 23
round((group.scores$sdDepression[1]*group.scores$mDepression[2] + group.scores$sdDepression[2]*group.scores$mDepression[1]) / sum(group.scores$sdDepression)) # 11
round((group.scores$sdSomatic[1]*group.scores$mSomatic[2] + group.scores$sdSomatic[2]*group.scores$mSomatic[1]) / sum(group.scores$sdSomatic)) # 11
round((group.scores$sdArousal[1]*group.scores$mArousal[2] + group.scores$sdArousal[2]*group.scores$mArousal[1]) / sum(group.scores$sdArousal)) # 8
round((group.scores$sdSocial[1]*group.scores$mSocial[2] + group.scores$sdSocial[2]*group.scores$mSocial[1]) / sum(group.scores$sdSocial)) # 16
round((group.scores$sdIntrusive[1]*group.scores$mIntrusive[2] + group.scores$sdIntrusive[2]*group.scores$mIntrusive[1]) / sum(group.scores$sdIntrusive)) # 15
round((group.scores$sdTraumatic[1]*group.scores$mTraumatic[2] + group.scores$sdTraumatic[2]*group.scores$mTraumatic[1]) / sum(group.scores$sdTraumatic)) # 9
round((group.scores$sdAvoidance[1]*group.scores$mAvoidance[2] + group.scores$sdAvoidance[2]*group.scores$mAvoidance[1]) / sum(group.scores$sdAvoidance)) # 20
#-------- Jacobson-Truax Plots --------
# report only cutoff scores instead of RCI plots
# NEUROTIC
JTRCI::JTRCI(data = data.frame(dat.pre.post),
pre = "neurotic_pre",
post = "neurotic_post",
reliability = .692,
dysfM = 23.30, dysfSD = 8.91,
normM = 8.91, normSD = 8.45)
# POSITIVE
JTRCI::JTRCI(data = data.frame(dat.pre.post),
pre = "positive_pre",
post = "positive_post",
reliability = .731, higherIsBetter = T,
dysfM = 20.06, dysfSD = 7.26,
normM = 25.56, normSD = 6.45)
# DEPRESSION
JTRCI::JTRCI(data = data.frame(dat.pre.post),
pre = "depression_pre",
post = "depression_post",
reliability = .648,
dysfM = 18.57, dysfSD = 9.59,
normM = 6.56, normSD = 6.54)
# AROUSAL
JTRCI::JTRCI(data = data.frame(dat.pre.post),
pre = "arousal_pre",
post = "arousal_post",
reliability = .632,
dysfM = 13.88, dysfSD = 10.21,
normM = 4.19, normSD = 5.32)
# SOMATIC
JTRCI::JTRCI(data = data.frame(dat.pre.post),
pre = "somatic_pre",
post = "somatic_post",
reliability = .786,
dysfM = 14.96, dysfSD = 9.02,
normM = 8.98, normSD = 6.46, JTcrit = "C")
# SOCIAL
JTRCI::JTRCI(data = data.frame(dat.pre.post),
pre = "social_pre",
post = "social_post",
reliability = .770,
dysfM = 18.49, dysfSD = 11.74,
normM = 13.29, normSD = 10.14, JTcrit = "C")
# INTRUSIVE
JTRCI::JTRCI(data = data.frame(dat.pre.post),
pre = "intrusive_pre",
post = "intrusive_post",
reliability = .635,
dysfM = 19.96, dysfSD = 11.02,
normM = 10.44, normSD = 8.94, JTcrit = "C")
# TRAUMATIC
JTRCI::JTRCI(data = data.frame(dat.pre.post),
pre = "intrusive_pre",
post = "intrusive_post",
reliability = .754,
dysfM = 12.91, dysfSD = 6.15,
normM = 10.80, normSD = 8.94, JTcrit = "C")
# AVOIDANCE
JTRCI::JTRCI(data = data.frame(dat.pre.post),
pre = "avoidance_pre",
post = "avoidance_post",
reliability = .683,
dysfM = 25.64, dysfSD = 15.18,
normM = 11.81, normSD = 9.17, JTcrit = "C")