generated from opensafely/research-template
/
process_data.R
432 lines (353 loc) · 20.3 KB
/
process_data.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
################################################################################
#
# Description: This script imports data extracted by the cohort extractor and
# calculates additional variables needed for subsequent anlyses
# (i.e., eligibility criteria window)
#
# Input: /output/data/input.csv.gz
#
# Output: /output/data/data_processed.csv
# /output/data/data_processed_clean.csv
#
# Author(s): M Green
# Date last updated: 11/02/2022
#
################################################################################
# Preliminaries ----
## Import libraries
library('tidyverse')
library('gt')
library('gtsummary')
library('here')
library('tidyverse')
library('lubridate')
library('arrow')
library('reshape2')
library('dplyr')
library('readr')
library('survival')
library('survminer')
## Custom functions
source(here("analysis", "lib", "custom_functions.R"))
# Process data ----
cat("#### process data ####\n")
## Read in data (don't rely on defaults)
data_extract0 <- read_csv(
here::here("output", "data", "input.csv.gz"),
col_types = cols_only(
# PATIENT ID ----
patient_id = col_integer(),
# CENSORING ----
death_date = col_date(format = "%Y-%m-%d"),
has_died = col_logical(),
dereg_date = col_date(format = "%Y-%m-%d"),
registered_eligible = col_logical(),
registered_treated = col_logical(),
# NEUTRALISING MONOCLONAL ANTIBODIES OR ANTIVIRALS ----
sotrovimab_covid_therapeutics = col_date(format = "%Y-%m-%d"),
molnupiravir_covid_therapeutics = col_date(format = "%Y-%m-%d"),
casirivimab_covid_therapeutics = col_date(format = "%Y-%m-%d"),
# ELIGIBILITY CRITERIA VARIABLES ----
covid_test_positive = col_logical(),
covid_test_positive_date = col_date(format = "%Y-%m-%d"),
#covid_positive_test_type = col_character(),
covid_positive_previous_30_days = col_logical(),
symptomatic_covid_test = col_character(),
covid_symptoms_snomed = col_date(format = "%Y-%m-%d"),
primary_covid_hospital_discharge_date = col_date(format = "%Y-%m-%d"),
any_covid_hospital_discharge_date = col_date(format = "%Y-%m-%d"),
age = col_integer(),
# SOLID ORGAN TRANSPLANT - FOR INVESTIGATION ----
solid_organ_transplant_nhsd = col_date(format = "%Y-%m-%d"),
# HIGH RISK GROUPS ----
high_risk_cohort_covid_therapeutics = col_character(),
downs_syndrome_nhsd = col_date(format = "%Y-%m-%d"),
sickle_cell_disease_nhsd = col_date(format = "%Y-%m-%d"),
cancer_opensafely_snomed = col_date(format = "%Y-%m-%d"),
haematological_disease_nhsd = col_date(format = "%Y-%m-%d"),
ckd_stage_5_nhsd = col_date(format = "%Y-%m-%d"),
liver_disease_nhsd = col_date(format = "%Y-%m-%d"),
imid_nhsd = col_date(format = "%Y-%m-%d"),
immunosupression_nhsd = col_date(format = "%Y-%m-%d"),
hiv_aids_nhsd = col_date(format = "%Y-%m-%d"),
solid_organ_transplant_nhsd = col_date(format = "%Y-%m-%d"),
multiple_sclerosis_nhsd = col_date(format = "%Y-%m-%d"),
motor_neurone_disease_nhsd = col_date(format = "%Y-%m-%d"),
myasthenia_gravis_nhsd = col_date(format = "%Y-%m-%d"),
huntingtons_disease_nhsd = col_date(format = "%Y-%m-%d"),
# CLINICAL/DEMOGRAPHIC COVARIATES ----
sex = col_character(),
ethnicity_primis = col_character(),
ethnicity_sus = col_character(),
imd = col_character(),
region_nhs = col_character(),
region_covid_therapeutics = col_character(),
stp = col_character(),
# CLINICAL GROUPS ----
autism_nhsd = col_logical(),
care_home_primis = col_logical(),
dementia_nhsd = col_logical(),
housebound_opensafely = col_logical(),
learning_disability_primis = col_logical(),
shielded_primis = col_logical(),
serious_mental_illness_nhsd = col_logical(),
vaccination_status = col_character(),
# OUTCOMES ----
covid_positive_test_30_days_post_elig_or_treat = col_date(format = "%Y-%m-%d"),
covid_hospitalisation_outcome_date = col_date(format = "%Y-%m-%d"),
covid_hospitalisation_critical_care = col_integer(),
death_with_covid_on_the_death_certificate_date = col_date(format = "%Y-%m-%d"),
death_with_28_days_of_covid_positive_test = col_logical()
),
)
## Fix bad dummy data
if(Sys.getenv("OPENSAFELY_BACKEND") %in% c("", "expectations")){
data_extract0 <- data_extract0 %>%
mutate(date = sample(seq(as.Date('2021/11/01'), as.Date('2022/02/01'), by="day"), nrow(data_extract0), replace = TRUE),
covid_test_positive_date = as.Date(covid_test_positive_date, format = "%Y-%m-%d"),
covid_test_positive_date = as.Date(ifelse(!is.na(covid_test_positive_date), date, NA), origin = "1970-01-01"),
date2 = as.Date(covid_test_positive_date + sample(-1:10, dim(data_extract0)[1], replace=TRUE), origin = "1970-01-01"),
sotrovimab_covid_therapeutics = ifelse(!is.na(sotrovimab_covid_therapeutics), date2, NA),
molnupiravir_covid_therapeutics = ifelse(!is.na(molnupiravir_covid_therapeutics), date2, NA),
casirivimab_covid_therapeutics = ifelse(!is.na(casirivimab_covid_therapeutics), date2, NA))
}
## Parse NAs
data_extract <- data_extract0 %>%
mutate(across(
.cols = where(is.character),
.fns = ~na_if(.x, "")
)) %>%
mutate(across(
.cols = c(where(is.numeric), -ends_with("_id")), #convert numeric+integer but not id variables
.fns = ~na_if(.x, 0)
)) %>%
arrange(patient_id) %>%
select(all_of((names(data_extract0))))
## Format columns (i.e, set factor levels)
data_processed <- data_extract %>%
mutate(
across(
where(is.logical),
~.x*1L
)) %>%
mutate(
# NEUTRALISING MONOCLONAL ANTIBODIES OR ANTIVIRALS ----
treatment_date = as.Date(pmin(sotrovimab_covid_therapeutics, molnupiravir_covid_therapeutics, casirivimab_covid_therapeutics, na.rm = TRUE), origin = "1970-01-01"),
treatment_type = ifelse(!is.na(sotrovimab_covid_therapeutics), "Sotrovimab",
ifelse(!is.na(molnupiravir_covid_therapeutics), "Molnupiravir",
ifelse(!is.na(casirivimab_covid_therapeutics), "Casirivimab", NA))),
# ELIGIBILITY VARIABLES ----
## Time between positive test and treatment
tb_postest_treat = ifelse(covid_test_positive == 1, as.numeric(treatment_date - covid_test_positive_date), NA),
## Eligibility window
rare_neurological_conditions_nhsd = pmax(multiple_sclerosis_nhsd, motor_neurone_disease_nhsd, myasthenia_gravis_nhsd,
huntingtons_disease_nhsd, na.rm = T),
high_risk_group_nhsd_date = pmax(downs_syndrome_nhsd, sickle_cell_disease_nhsd, cancer_opensafely_snomed,
haematological_disease_nhsd, ckd_stage_5_nhsd, liver_disease_nhsd, imid_nhsd,
immunosupression_nhsd, hiv_aids_nhsd, solid_organ_transplant_nhsd, rare_neurological_conditions_nhsd,
na.rm = TRUE),
elig_start = as.Date(ifelse(covid_test_positive == 1 & (covid_test_positive_date >= high_risk_group_nhsd_date), covid_test_positive_date, NA), origin = "1970-01-01"),
elig_end = as.Date(elig_start + 5, origin = "1970-01-01"),
# HIGH RISK GROUPS ----
downs_syndrome_nhsd_name = ifelse(!is.na(downs_syndrome_nhsd), "Down's syndrome", NA),
sickle_cell_disease_nhsd_name = ifelse(!is.na(sickle_cell_disease_nhsd), "sickle cell disease", NA),
cancer_opensafely_name = ifelse(!is.na(cancer_opensafely_snomed), "solid cancer", NA),
haematological_disease_nhsd_name = ifelse(!is.na(haematological_disease_nhsd), "haematological diseases and stem cell transplant recipients", NA),
ckd_stage_5_nhsd_name = ifelse(!is.na(ckd_stage_5_nhsd), "renal disease", NA),
liver_disease_nhsd_name = ifelse(!is.na(liver_disease_nhsd), "liver disease", NA),
imid_nhsd_name = ifelse(!is.na(imid_nhsd), "IMID", NA),
immunosupression_nhsd_name = ifelse(!is.na(immunosupression_nhsd), "primary immune deficiencies", NA),
hiv_aids_nhsd_name = ifelse(!is.na(hiv_aids_nhsd), "HIV or AIDS immunosupression", NA),
solid_organ_transplant_nhsd_name = ifelse(!is.na(solid_organ_transplant_nhsd), "solid organ recipients", NA),
rare_neurological_conditions_nhsd_name = ifelse(!is.na(rare_neurological_conditions_nhsd), "rare neurological conditions", NA),
downs_syndrome_nhsd = ifelse(!is.na(downs_syndrome_nhsd), 1, NA),
sickle_cell_disease_nhsd = ifelse(!is.na(sickle_cell_disease_nhsd), 1, NA),
cancer_opensafely = ifelse(!is.na(cancer_opensafely_snomed), 1, NA),
haematological_disease_nhsd = ifelse(!is.na(haematological_disease_nhsd), 1, NA),
ckd_stage_5_nhsd = ifelse(!is.na(ckd_stage_5_nhsd), 1, NA),
liver_disease_nhsd = ifelse(!is.na(liver_disease_nhsd), 1, NA),
imid_nhsd = ifelse(!is.na(imid_nhsd), 1, NA),
immunosupression_nhsd = ifelse(!is.na(immunosupression_nhsd), 1, NA),
hiv_aids_nhsd = ifelse(!is.na(hiv_aids_nhsd), 1, NA),
solid_organ_transplant_nhsd = ifelse(!is.na(solid_organ_transplant_nhsd), 1, NA),
rare_neurological_conditions_nhsd = ifelse(!is.na(rare_neurological_conditions_nhsd), 1, NA),
downs_syndrome_therapeutics = ifelse(str_detect(high_risk_cohort_covid_therapeutics, "Downs syndrome") == TRUE, 1, NA),
sickle_cell_disease_therapeutics = ifelse(str_detect(high_risk_cohort_covid_therapeutics, "sickle cell disease") == TRUE, 1, NA),
cancer_therapeutics = ifelse(str_detect(high_risk_cohort_covid_therapeutics, "solid cancer") == TRUE, 1, NA),
haematological_disease_therapeutics = ifelse(str_detect(high_risk_cohort_covid_therapeutics, "haematological diseases and stem cell transplant recipients") == TRUE, 1, NA),
haematological_disease_therapeutics = ifelse(str_detect(high_risk_cohort_covid_therapeutics, "haematological diseases") == TRUE, 1, haematological_disease_therapeutics),
haematological_disease_therapeutics = ifelse(str_detect(high_risk_cohort_covid_therapeutics, "haematologic malignancy") == TRUE, 1, haematological_disease_therapeutics),
haematological_disease_therapeutics = ifelse(str_detect(high_risk_cohort_covid_therapeutics, "stem cell transplant recipients") == TRUE, 1, haematological_disease_therapeutics),
ckd_stage_5_therapeutics = ifelse(str_detect(high_risk_cohort_covid_therapeutics, "renal disease") == TRUE, 1, NA),
liver_disease_therapeutics = ifelse(str_detect(high_risk_cohort_covid_therapeutics, "liver disease") == TRUE, 1, NA),
imid_therapeutics = ifelse(str_detect(high_risk_cohort_covid_therapeutics, "IMID") == TRUE, 1, NA),
immunosupression_therapeutics = ifelse(str_detect(high_risk_cohort_covid_therapeutics, "primary immune deficiencies") == TRUE, 1, NA),
hiv_aids_therapeutics = ifelse(str_detect(high_risk_cohort_covid_therapeutics, "HIV or AIDS") == TRUE, 1, NA),
solid_organ_transplant_therapeutics = ifelse(str_detect(high_risk_cohort_covid_therapeutics, "solid organ recipients") == TRUE, 1, NA),
rare_neurological_conditions_therapeutics = ifelse(str_detect(high_risk_cohort_covid_therapeutics, "rare neurological conditions") == TRUE, 1, NA),
downs_syndrome = ifelse(downs_syndrome_nhsd == 1 | downs_syndrome_therapeutics == 1, 1, NA),
sickle_cell_disease = ifelse(sickle_cell_disease_nhsd == 1 | sickle_cell_disease_therapeutics== 1, 1, NA),
solid_cancer = ifelse(cancer_opensafely == 1 | cancer_therapeutics == 1, 1, NA),
haematological_disease = ifelse(haematological_disease_nhsd == 1 | haematological_disease_therapeutics == 1, 1, NA),
renal_disease = ifelse(ckd_stage_5_nhsd == 1 | ckd_stage_5_therapeutics == 1, 1, NA),
liver_disease = ifelse(liver_disease_nhsd == 1 | liver_disease_therapeutics == 1, 1, NA),
imid = ifelse(imid_nhsd == 1 | imid_therapeutics == 1, 1, NA),
immunosupression = ifelse(immunosupression_nhsd == 1 | immunosupression_therapeutics == 1, 1, NA),
hiv_aids = ifelse(hiv_aids_nhsd == 1 | hiv_aids_therapeutics == 1, 1, NA),
solid_organ_transplant = ifelse(solid_organ_transplant_nhsd == 1 | solid_organ_transplant_therapeutics == 1, 1, NA),
rare_neurological_conditions = ifelse(rare_neurological_conditions_nhsd == 1 | rare_neurological_conditions_therapeutics == 1, 1, NA)
) %>%
unite("high_risk_group_nhsd_combined", downs_syndrome_nhsd_name, sickle_cell_disease_nhsd_name, cancer_opensafely_name,
haematological_disease_nhsd_name, ckd_stage_5_nhsd_name, liver_disease_nhsd_name, imid_nhsd_name, immunosupression_nhsd_name,
hiv_aids_nhsd_name, solid_organ_transplant_nhsd_name, rare_neurological_conditions_nhsd_name, sep = ",", na.rm = T) %>%
mutate(
## Find matches between nhsd high risk cohorts and therapeutics high risk cohorts
ind_therapeutic_groups = map_chr(strsplit(high_risk_cohort_covid_therapeutics, ","), paste,collapse="|"),
match = str_detect(high_risk_group_nhsd_combined, ind_therapeutic_groups)) %>%
rowwise() %>%
mutate(
## Combine nhsd cohorts with theraputics cohorts to get list of all cohorts
high_risk_group_combined = as.character(ifelse(match == TRUE,
paste(high_risk_group_nhsd_combined, high_risk_cohort_covid_therapeutics, sep = ","), "")),
high_risk_group_combined = ifelse(high_risk_group_combined == "NA", "", high_risk_group_combined),
high_risk_group_combined = as.character(paste(unique(unlist(strsplit(high_risk_group_combined, ","))), collapse = ",")),
high_risk_group_combined_count = ifelse(high_risk_group_combined != "", str_count(high_risk_group_combined,",") + 1, NA)) %>%
ungroup() %>%
mutate(
# Cinic/demo variables
sex = fct_case_when(
sex == "F" ~ "Female",
sex == "M" ~ "Male",
#sex == "I" ~ "Inter-sex",
#sex == "U" ~ "Unknown",
TRUE ~ NA_character_
),
ethnicity_filled = ifelse(is.na(ethnicity_primis), ethnicity_sus, ethnicity_primis),
ethnicity = ifelse(is.na(ethnicity_filled), 6, ethnicity_filled),
ethnicity = fct_case_when(
ethnicity == "1" ~ "White",
ethnicity == "2" ~ "Mixed",
ethnicity == "3" ~ "Asian or Asian British",
ethnicity == "4" ~ "Black or Black British",
ethnicity == "5" ~ "Other ethnic groups",
ethnicity == "6" ~ "Unknown",
#TRUE ~ "Unknown"
TRUE ~ NA_character_),
imd = na_if(imd, "0"),
imd = fct_case_when(
imd == 1 ~ "1 most deprived",
imd == 2 ~ "2",
imd == 3 ~ "3",
imd == 4 ~ "4",
imd == 5 ~ "5 least deprived",
#TRUE ~ "Unknown",
TRUE ~ NA_character_
),
region_nhs = fct_case_when(
region_nhs == "London" ~ "London",
region_nhs == "East" ~ "East of England",
region_nhs == "East Midlands" ~ "East Midlands",
region_nhs == "North East" ~ "North East",
region_nhs == "North West" ~ "North West",
region_nhs == "South East" ~ "South East",
region_nhs == "South West" ~ "South West",
region_nhs == "West Midlands" ~ "West Midlands",
region_nhs == "Yorkshire and The Humber" ~ "Yorkshire and the Humber",
#TRUE ~ "Unknown",
TRUE ~ NA_character_),
# OUTCOMES ----
covid_positive_test_30_days_post_elig_or_treat_date = covid_positive_test_30_days_post_elig_or_treat,
covid_positive_test_30_days_post_elig_or_treat = ifelse(!is.na(covid_positive_test_30_days_post_elig_or_treat_date), 1, 0),
start_date = pmin(covid_test_positive_date, treatment_date, na.rm = T),
covid_hospital_admission = ifelse(covid_hospitalisation_outcome_date > start_date, 1, 0),
covid_hospitalisation_critical_care = ifelse(covid_hospitalisation_critical_care > 0 & covid_hospital_admission == 1, 1, 0),
covid_death = ifelse(!is.na(death_with_covid_on_the_death_certificate_date) |
death_with_28_days_of_covid_positive_test == 1, 1, 0)
)
# Save dataset(s) ----
write_rds(data_processed, here::here("output", "data", "data_processed.rds"), compress = "gz")
# Process clean data ----
cat("#### process clean data ####\n")
## Apply eligibility and exclusion criteria
data_processed_eligible <- data_processed %>%
filter(
# Alive and registered
has_died == 0,
registered_eligible == 1 | registered_treated == 1,
# Apply eligibility criteria
covid_test_positive == 1,
covid_positive_previous_30_days != 1,
(tb_postest_treat <= 5 & tb_postest_treat >= 0) | is.na(tb_postest_treat),
!is.na(high_risk_group_nhsd_date),
# Apply exclusion criteria
is.na(primary_covid_hospital_discharge_date) | (primary_covid_hospital_discharge_date < (elig_start - 30) & primary_covid_hospital_discharge_date > (elig_start)),
age >= 12,
# Only eligible patients
!is.na(elig_start),
) %>%
mutate(eligibility_status = "Eligible")
cat("#### eligible patients ####\n")
print(dim(data_processed_eligible))
print(table(data_processed_eligible$match))
## Include treated patients not flagged as eligible
data_processed_treated <- data_processed %>%
filter(
# Treated but non-eligible patients
!is.na(treatment_date),
!(patient_id %in% unique(data_processed_eligible$patient_id)),
) %>%
mutate(eligibility_status = "Treated")
cat("#### treated patients ####\n")
print(dim(data_processed_treated))
print(table(data_processed_treated$match))
## Free up space and combine
rm(data_processed)
data_processed_combined <- rbind(data_processed_eligible, data_processed_treated)
rm(data_processed_eligible)
rm(data_processed_treated)
print(dim(data_processed_combined))
print(table(data_processed_combined$eligibility_status))
print(table(data_processed_combined$eligibility_status, data_processed_combined$match))
## Exclude patients issued more than one treatment within two weeks
dup_ids <- data_processed_combined %>%
select(patient_id, treatment_date, sotrovimab_covid_therapeutics, molnupiravir_covid_therapeutics, casirivimab_covid_therapeutics) %>%
filter(!is.na(treatment_date)) %>%
mutate(sotrovimab_covid_therapeutics = as.Date(sotrovimab_covid_therapeutics, origin="1970-01-01"),
molnupiravir_covid_therapeutics = as.Date(molnupiravir_covid_therapeutics, origin="1970-01-01"),
casirivimab_covid_therapeutics = as.Date(casirivimab_covid_therapeutics, origin="1970-01-01"),
sot_mol_diff = as.numeric(sotrovimab_covid_therapeutics - molnupiravir_covid_therapeutics),
sot_cas_diff = as.numeric(sotrovimab_covid_therapeutics - casirivimab_covid_therapeutics),
mol_cas_diff = as.numeric(molnupiravir_covid_therapeutics - casirivimab_covid_therapeutics)) %>%
melt(id.var = "patient_id", measure.vars = c("sot_mol_diff", "sot_cas_diff", "mol_cas_diff")) %>%
filter(!is.na(value),
value <= 14 | value >= -14) %>%
group_by(patient_id) %>%
arrange(patient_id)
cat("#### patients with more than one treatment ####\n")
print(dim(dup_ids))
data_processed_clean <- data_processed_combined %>%
subset(!(patient_id %in% unique(dup_ids$patient_id))) %>%
select(
# ID
patient_id, eligibility_status,
# Censoring
has_died, death_date, dereg_date, registered_eligible, registered_treated,
# Eligibility
covid_test_positive, covid_test_positive_date, covid_positive_previous_30_days, tb_postest_treat, elig_start, elig_end, primary_covid_hospital_discharge_date,
any_covid_hospital_discharge_date,
# Treatment
sotrovimab_covid_therapeutics, molnupiravir_covid_therapeutics, casirivimab_covid_therapeutics, treatment_date, treatment_type,
# High risk cohort
downs_syndrome, sickle_cell_disease, solid_cancer, haematological_disease, renal_disease, liver_disease, imid, immunosupression,
hiv_aids, solid_organ_transplant, rare_neurological_conditions, high_risk_group_nhsd_combined, high_risk_cohort_covid_therapeutics,
match, high_risk_group_combined, high_risk_group_combined_count,
# Clinical and demographic variables
age, sex, ethnicity, imd, region_nhs, region_covid_therapeutics, stp,
# Clinical groups
autism_nhsd, care_home_primis, dementia_nhsd, housebound_opensafely, learning_disability_primis, shielded_primis,
serious_mental_illness_nhsd, vaccination_status,
# Outcomes
covid_positive_test_30_days_post_elig_or_treat, covid_hospitalisation_outcome_date, covid_hospitalisation_critical_care, covid_death)
rm(data_processed_combined)
# Save dataset(s) ----
write_rds(data_processed_clean, here::here("output", "data", "data_processed_clean.rds"), compress = "gz")