-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathprepare.R
243 lines (203 loc) · 13.6 KB
/
prepare.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
library("SimNPH")
library("tidyverse")
# read data ---------------------------------------------------------------
design_vars_all <- c("random_withdrawal", "n_pat_design", "recruitment")
design_vars_delayed <- c("delay", "hr_after_onset", "median_survival_ctrl", design_vars_all)
design_vars_crossing <- c("crossing", "hr_before", "hr_after", "median_survival_ctrl", design_vars_all)
design_vars_progression <- c("hr_death_before_prog", "hr_after_prog_ctrl", "median_time_to_prog_ctrl", "hr_prog", "median_survival_before_prog", design_vars_all)
design_vars_subgroup <- c("hr_trt", "hr_subgroup_display", "prevalence", "median_survival_ctrl", design_vars_all)
shhr_varnames <- c("hazard_ctrl", "prog_rate_ctrl", "hazard_after_prog", "hazard_trt", "delay", "hazard_trt_after", "hazard_trt_before", "hazard_subgroup", "prevalence", "prog_rate_trt", "crossing")
names(shhr_varnames) <- shhr_varnames
data_dir <- "../Diplomarbeit/Simulations/data/"
delayed <- readRDS(paste0(data_dir, "simulation_delayed_effect_WTGP024_2023-10-24_160438/results.Rds"))
crossing <- readRDS(paste0(data_dir, "simulation_crossing_harzards_WTGP024_2023-11-06_101033/results.Rds"))
subgroup <- readRDS(paste0(data_dir, "simulation_subgroup_WTGP024_2023-12-05_095206/results.Rds"))
progression <- readRDS(paste0(data_dir, "simulation_disease_progression_WTGP024_2023-11-16_094021/results.Rds"))
# Metadata ----------------------------------------------------------------
# metadata for methods ----------------------------------------------------
method_metadata <- tibble::tribble(
# variable name direction (one sided test) name for output estimator/test/gs-test procedure category
~method, ~direction, ~method_name, ~type, ~category,
"ahr_6m", "lower", "AHR 6m", "estimator", "Average Hazard Ratio",
"ahr_12m", "lower", "AHR 12m", "estimator", "Average Hazard Ratio",
"gahr_6m", "lower", "gAHR 6m", "estimator", "Geometric Average Hazard Ratio",
"gahr_12m", "lower", "gAHR 12m", "estimator", "Geometric Average Hazard Ratio",
"median_surv", "higher", "diff. med. surv.", "estimator", "Diff. Median Survival",
"milestone_6m", "higher", "milestone surv. ratio 6m", "estimator", "Ratio of Milestone Survival",
"milestone_12m", "higher", "milestone surv. ratio 12m", "estimator", "Ratio of Milestone Survival",
"rmst_diff_6m", "higher", "RMST (6m) difference", "estimator", "Diff. RMST",
"rmst_diff_12m", "higher", "RMST (12m) difference", "estimator", "Diff. RMST",
"cox", "lower", "Cox regression", "estimator", "Cox regression",
"aft_weibull", "higher", "AFT Weibull", "estimator", "Accelerated Failure Time Model",
"aft_lognormal", "higher", "AFT log-normal", "estimator", "Accelerated Failure Time Model",
"diff_med_weibull", "higher", "diff. med. surv. Weibull", "estimator", "Diff. Median Survival",
"fh_0_0", NA, "FH 0,0", "test", "Fleming-Harrington Test",
"fh_0_1", NA, "FH 0,1", "test", "Fleming-Harrington Test",
"fh_1_0", NA, "FH 1,0", "test", "Fleming-Harrington Test",
"fh_1_1", NA, "FH 1,1", "test", "Fleming-Harrington Test",
"logrank", NA, "logrank test", "test", "Log-Rank Test",
"max_combo", NA, "max-combo test", "test", "Max-Combo Test",
"modest_6", NA, "modestly wtd. t*=6m", "test", "Modestly Weighted Test",
"modest_8", NA, "modestly wtd. t*=8m", "test", "Modestly Weighted Test",
"fh_gs_0_0", NA, "FH 0,0, grp. seq.", "group sequential test", "Fleming-Harrington Test",
"fh_gs_0_1", NA, "FH 0,1, grp. seq.", "group sequential test", "Fleming-Harrington Test",
"fh_gs_1_0", NA, "FH 1,0, grp. seq.", "group sequential test", "Fleming-Harrington Test",
"fh_gs_1_1", NA, "FH 1,1, grp. seq.", "group sequential test", "Fleming-Harrington Test",
"logrank_gs", NA, "logrank test, grp. seq.", "group sequential test", "Log-Rank Test",
"max_combo_gs", NA, "max-combo test, grp. seq.", "group sequential test", "Max-Combo Test",
"modest_gs_6", NA, "modestly wtd. t*=6m, grp. seq.", "group sequential test", "Modestly Weighted Test",
"modest_gs_8", NA, "modestly wtd. t*=8m, grp. seq.", "group sequential test", "Modestly Weighted Test"
)
# metadata for output columns ---------------------------------------------
rename_cols <- tibble::tribble(
~colname_old, ~colname_new,
"random_withdrawal", "rate of random withdrawal",
"n_pat_design", "number of patients",
"recruitment", "recruitment time",
"delay", "delay of onset of treatment effect",
"hr_after_onset", "HR after onset of treatment effect",
"median_survival_ctrl", "median survival in the control arm",
"crossing", "time of crossing of the hazards",
"hr_before", "HR before crossing of the hazards",
"hr_after", "HR after crossing of the hazards",
"hr_trt", "HR treatment vs. control",
"hr_subgroup_display", "HR subgroup vs. control",
"prevalence", "prevalence",
"hr_death_before_prog", "HR death before prog. treatment vs. control",
"hr_after_prog_ctrl", "HR death after prog. vs. control before prog.",
"median_time_to_prog_ctrl", "median time to progression, control",
"hr_prog", "HR progression treatment vs. control",
"median_survival_before_prog", "median time to death without progression, control",
"median_survival_trt", "median survival in the treatment arm",
"rmst_trt_6m", "RMST (6m) in the treatment arm",
"rmst_ctrl_6m", "RMST (6m) in the control arm",
"gAHR_6m", "geometric average hazard ratio (6m)",
"AHR_6m", "average hazard ratio (6m)",
"rmst_trt_12m", "RMST (12m) in the treatment arm",
"rmst_ctrl_12m", "RMST (12m) in the control arm",
"gAHR_12m", "geometric average hazard ratio (12m)",
"AHR_12m", "average hazard ratio (12m)",
"milestone_survival_trt_6m", "milestone survival in the treatment arm (6m)",
"milestone_survival_ctrl_6m", "milestone survival in the control arm (6m)",
"milestone_survival_trt_12m", "milestone survival in the treatment arm (12m)",
"milestone_survival_ctrl_12m", "milestone survival in the control arm (12m)",
"descriptive.n_pat", "average number of patients",
"descriptive.evt", "average number of events",
"descriptive.evt_ctrl", "average number of events in the control arm",
"descriptive.evt_trt", "average number of events in the treatment arm",
"method", "method",
"mean_est", "mean point estimate",
"median_est", "median point estimate",
"sd_est", "standard deviation of point estimate",
"bias", "bias",
"mse", "mean squared error",
"mae", "mean absolute error",
"coverage", "CI coverage",
"null_cover", "proportions of CIs that contain the null value",
"width", "average CI width",
"mean_n_pat", "average number of patients (group sequential)",
"mean_n_evt", "average number of events (group sequential)",
"rejection", "rejection rate, one sided alpha=0.025",
"study_time", "average study time (group sequential)",
"followup", "average max followup (group sequential)",
)
# columns to tansform from days to months ---------------------------------
stats_scale <- c("mean_est", "median_est", "sd_est", "bias", "sd_bias", "mse", "sd_mse", "mae", "sd_mae", "width", "sd_width", "mean_sd", "sd_sd")
methods_scale <- c("diff_med_weibull", "median_surv", "rmst_diff_12m", "rmst_diff_6m")
time_varnames <- c(
"delay", "crossing", "recruitment", "median_survival_trt",
"median_survival_ctrl", "rmst_trt_6m", "rmst_ctrl_6m", "rmst_trt_12m",
"rmst_ctrl_12m", "descriptive.max_followup", "descriptive.study_time",
"descriptive.sd_max_followup", "descriptive.sd_study_time",
as.character(outer(methods_scale, stats_scale, \(x,y){str_c(x, ".", y)}))
)
time_trafo <- SimNPH::d2m
# custom transformations --------------------------------------------------
progression <- progression |>
mutate(
median_survival_before_prog = SimNPH::r2m(hazard_ctrl),
median_time_to_prog_ctrl = SimNPH::r2m(prog_rate_ctrl)
)
subgroup <- subgroup |>
mutate(
hr_subgroup_display = ifelse(
hr_subgroup > 1,
paste0("1/", 1/hr_subgroup),
as.character(hr_subgroup)
) |>
factor(levels=c("0.7", "0.8", "0.9", "1/0.9", "1/0.8", "1/0.7"), ordered = TRUE)
)
# calculate rejection -----------------------------------------------------
# prepare data for app ----------------------------------------------------
prepare_data <- function(dataset, design_varnames){
# transform time vars
dataset <- dataset |>
mutate(
across(any_of(time_varnames), time_trafo)
)
# pivot longer
dataset <- results_pivot_longer(dataset) |>
left_join(method_metadata, by = "method") |> # add metadata
mutate(
# one sided test based on CI
ci_based_one_sided_rejection = case_when(
direction == "lower" ~ 1 - null_upper,
direction == "higher" ~ 1 - null_lower,
TRUE ~ NA_real_
),
# combine columsn with different names acros estimators/tests/gs-tests
# coalesce: take first value, if it is missing take second value
mean_n_pat = coalesce(mean_n_pat, n_pat),
mean_n_evt = coalesce(mean_n_evt, n_evt),
sd_n_pat = coalesce(sd_n_pat, sd_npat),
sd_n_evt = coalesce(sd_n_evt, sd_nevt),
study_time = coalesce(study_time, descriptive.study_time),
followup = coalesce(followup, descriptive.max_followup),
rejection = coalesce(rejection, rejection_0.025),
rejection = coalesce(rejection, ci_based_one_sided_rejection)
)
# rename methods
stopifnot(all(dataset$method %in% method_metadata$method))
tmp_method_names <- method_metadata$method_name |>
setNames(method_metadata$method)
dataset$method <- tmp_method_names[dataset$method]
# rename columns
tmp_selection <- rename_cols$colname_old |>
setNames(rename_cols$colname_new) |>
c(shhr_varnames) |>
keep(\(x){x %in% names(dataset)})
dataset <- dataset |>
select(tmp_selection)
tmp_selection2 <- rename_cols$colname_new |>
setNames(rename_cols$colname_old)
design_varnames <- tmp_selection2[design_varnames] |>
unname()
list(
data = dataset,
design_variables = design_varnames,
methods = unique(dataset$method),
filter_values = lapply(
design_varnames,
\(var){
unique(dataset[[var]]) |>
sort()
}
) |>
setNames(design_varnames)
)
}
datasets <- list(
delayed = prepare_data(delayed, design_vars_delayed ),
crossing = prepare_data(crossing, design_vars_crossing ),
subgroup = prepare_data(subgroup, design_vars_subgroup ),
progression = prepare_data(progression, design_vars_progression )
)
saveRDS(datasets, "datasets.Rds")
# render markdown ---------------------------------------------------------
rmarkdown::render(
"description.md",
rmarkdown::html_document(
template = "pandoc_template.html"
),
output_file="app/description.html"
)