generated from opensafely/research-template
/
process_unmatched.R
282 lines (211 loc) · 13.6 KB
/
process_unmatched.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
# # # # # # # # # # # # # # # # # # # # #
# This script:
# merge case and control groups & sort variables for analysis
#
#
# # # # # # # # # # # # # # # # # # # # #
library('tidyverse')
library("ggplot2")
library('dplyr')
library('lubridate')
setwd(here::here("output"))
#setwd("/Users/yayang/Documents/GitHub/amr-uom-brit/output")
# extracted dataset after matching
df=read_csv("input_unmatched.csv")
######## time ##########
## wave
df$wave=ifelse(df$patient_index_date > as.Date("2020-11-30"),"3", # wave3(national vaccination programme):Dec2020--Dec,2021
ifelse(df$patient_index_date > as.Date("2020-08-31"),"2", # wave2(test for wider population): Sep-Dec,2020
ifelse( df$patient_index_date > as.Date("2020-01-31"),"1","0")))# wave1(test for health workers):Feb-Aug,2020
#1 die before patient_index_date/#2 die after patient_index_date/# 0 never die
df$died_ever=ifelse(df$ons_died_date < df$patient_index_date, 1, 2)
df$died_ever=ifelse(is.na(df$ons_died_date),0,df$died_ever)
df=df%>%filter(died_ever != 1) # exclude died before index
#1 deregister before patient_index_date/#2 deregister after patient_index_date/# 0 never de-register
df$dereg_ever=ifelse(df$dereg_date < df$patient_index_date, 1, 2)
df$dereg_ever=ifelse(is.na(df$dereg_date),0,df$dereg_ever)
df=df%>%filter(died_ever != 1) # exclude de-register before index
####### matching variables ########
## age
df=df%>%filter(df$age_cat != "0")
df$age_cat <- factor(df$age_cat, levels=c("0-4", "18-29","30-39","40-49","50-59","60-69","70-79","80+"))
######## antibiotics exposure ##########
##antibiotic type
# select ab types columns
col=c("Rx_Amikacin", "Rx_Amoxicillin", "Rx_Ampicillin", "Rx_Azithromycin", "Rx_Aztreonam", "Rx_Benzylpenicillin", "Rx_Cefaclor", "Rx_Cefadroxil", "Rx_Cefalexin", "Rx_Cefamandole", "Rx_Cefazolin", "Rx_Cefepime", "Rx_Cefixime", "Rx_Cefotaxime", "Rx_Cefoxitin", "Rx_Cefpirome", "Rx_Cefpodoxime", "Rx_Cefprozil", "Rx_Cefradine", "Rx_Ceftazidime", "Rx_Ceftriaxone", "Rx_Cefuroxime", "Rx_Chloramphenicol", "Rx_Cilastatin", "Rx_Ciprofloxacin", "Rx_Clarithromycin", "Rx_Clindamycin", "Rx_Co_amoxiclav", "Rx_Co_fluampicil", "Rx_Colistimethate", "Rx_Dalbavancin", "Rx_Dalfopristin", "Rx_Daptomycin", "Rx_Demeclocycline", "Rx_Doripenem", "Rx_Doxycycline", "Rx_Ertapenem", "Rx_Erythromycin", "Rx_Fidaxomicin", "Rx_Flucloxacillin", "Rx_Fosfomycin", "Rx_Fusidate", "Rx_Gentamicin", "Rx_Levofloxacin", "Rx_Linezolid", "Rx_Lymecycline", "Rx_Meropenem", "Rx_Methenamine", "Rx_Metronidazole", "Rx_Minocycline", "Rx_Moxifloxacin", "Rx_Nalidixic_acid", "Rx_Neomycin", "Rx_Netilmicin", "Rx_Nitazoxanid", "Rx_Nitrofurantoin", "Rx_Norfloxacin", "Rx_Ofloxacin", "Rx_Oxytetracycline", "Rx_Phenoxymethylpenicillin", "Rx_Piperacillin", "Rx_Pivmecillinam", "Rx_Pristinamycin", "Rx_Rifaximin", "Rx_Sulfadiazine", "Rx_Sulfamethoxazole", "Rx_Sulfapyridine", "Rx_Taurolidin", "Rx_Tedizolid", "Rx_Teicoplanin", "Rx_Telithromycin", "Rx_Temocillin", "Rx_Tetracycline", "Rx_Ticarcillin", "Rx_Tigecycline", "Rx_Tinidazole", "Rx_Tobramycin", "Rx_Trimethoprim", "Rx_Vancomycin")
df[col]=df[col]%>%mutate_all(~replace(., is.na(.), 0)) # recode NA -> 0
df$total_ab=rowSums(df[col])# total types number -> total ab prescription
df[col]=ifelse(df[col]>0,1,0) # number of matches-> binary flag(1,0)
df$ab_types=rowSums(df[col]>0)# count number of types
write_rds(df[col], here::here("output", "abtype79.rds"))
df=df[ ! names(df) %in% col]
df$ab_types=ifelse(is.na(df$ab_types),0,df$ab_types) # no ab
##antibiotic prescribing frequency
df$ab_last_date=as.Date(df$ab_last_date)
df$ab_first_date=as.Date(df$ab_first_date)
# use AB more than once
df$interval=as.integer(difftime(df$ab_last_date,df$ab_first_date,unit="day"))
df$interval=ifelse(df$interval==0,1,df$interval)#less than 1 day (first=last) ~ record to 1
df$lastABtime=as.integer(difftime(df$patient_index_date,df$ab_last_date,unit="day"))
df$lastABtime=ifelse(is.na(df$lastABtime),0,df$lastABtime)
## quintile category
#quintile<-function(x){
# ifelse(is.na(x)|x==0,"0",
# ifelse(x>quantile(x,.8),"5",
# ifelse(x>quantile(x,.6),"4",
# ifelse(x>quantile(x,.4),"3",
# ifelse(x>quantile(x,.2),"2","1")))))}
# df$ab_qn=quintile(df$ab_prescriptions)
# df$br_ab_qn=quintile(df$broad_ab_prescriptions)
# set ab quintile category - compare two ways #
##### I. ab_prescription
### 1.1-ab quintile = according to unique ab prescription numbers
df$ab_prescriptions=ifelse(df$ab_prescriptions==0,NA,df$ab_prescriptions) # filter no ab
qn_num=unique(df$ab_prescriptions)
qn_cat1=quantile(qn_num,0.2,na.rm=T)
qn_cat2=quantile(qn_num,0.4,na.rm=T)
qn_cat3=quantile(qn_num,0.6,na.rm=T)
qn_cat4=quantile(qn_num,0.8,na.rm=T)
df$ab_qn_5=ifelse(is.na(df$ab_prescriptions),0,
ifelse(df$ab_prescriptions<=qn_cat1,1,
ifelse(df$ab_prescriptions<=qn_cat2,2,
ifelse(df$ab_prescriptions<=qn_cat3,3,
ifelse(df$ab_prescriptions<=qn_cat4,4,5)
))))
### 1.2-broad ab quintile = according to unique broad ab prescription numbers
df$broad_ab_prescriptions=ifelse(df$broad_ab_prescriptions==0,NA,df$broad_ab_prescriptions) # filter no ab+no br ab
br_qn_num=unique(df$broad_ab_prescriptions)
br_qn_cat1=quantile(br_qn_num,0.2,na.rm=T)
br_qn_cat2=quantile(br_qn_num,0.4,na.rm=T)
br_qn_cat3=quantile(br_qn_num,0.6,na.rm=T)
br_qn_cat4=quantile(br_qn_num,0.8,na.rm=T)
df$br_ab_qn_5=ifelse(is.na(df$broad_ab_prescriptions),0,
ifelse(df$broad_ab_prescriptions<=br_qn_cat1,1,
ifelse(df$broad_ab_prescriptions<=br_qn_cat2,2,
ifelse(df$broad_ab_prescriptions<=br_qn_cat3,3,
ifelse(df$broad_ab_prescriptions<=br_qn_cat4,4,5)
))))
df$br_ab_qn_5=ifelse(is.na(df$ab_prescriptions),"without any ab",
ifelse(is.na(df$broad_ab_prescriptions)|df$broad_ab_prescriptions==0,"without broad ab",
df$br_ab_qn_5))
### 2.1-ab quintile = according to ab prescription numbers
df$ab_prescriptions=ifelse(df$ab_prescriptions==0,NA,df$ab_prescriptions) # filter no ab
df=df%>%mutate(ab_qn=ntile(ab_prescriptions,5))
df$ab_qn=ifelse(is.na(df$ab_qn),0,df$ab_qn)# no ab ->0; ab exp. ->1~5
df$ab_qn=as.factor(df$ab_qn)
### 2.2-broad ab quintile = according to broad ab prescription numbers
df$broad_ab=ifelse(is.na(df$ab_prescriptions)| # without any ab
is.na(df$broad_ab_prescriptions)|df$broad_ab_prescriptions==0,NA, # without broad ab
df$broad_ab_prescriptions) # with broad ab
df=df%>%mutate(br_ab_qn=ntile(broad_ab,5))
df$br_ab_qn=ifelse(is.na(df$ab_prescriptions),"without any ab",
ifelse(is.na(df$broad_ab_prescriptions)|df$broad_ab_prescriptions==0,"without broad ab",
df$br_ab_qn))
df$br_ab_qn=ifelse(is.na(df$br_ab_qn),0,df$br_ab_qn)
df$br_ab_qn=as.factor(df$br_ab_qn)
# ab_continuous
df$ab_prescriptions=ifelse(is.na(df$ab_prescriptions),0,df$ab_prescriptions) # recode NA to 0
df$broad_ab_prescriptions=ifelse(is.na(df$broad_ab_prescriptions),0,df$broad_ab_prescriptions) # recode NA to 0
#### II. total ab- calculated from 79 knid ab
### 1.1-ab quintile = according to unique ab prescription numbers
df$total_ab=ifelse(df$total_ab==0,NA,df$total_ab) # filter no ab
qn_num=unique(df$total_ab)
qn_cat1=quantile(qn_num,0.2,na.rm=T)
qn_cat2=quantile(qn_num,0.4,na.rm=T)
qn_cat3=quantile(qn_num,0.6,na.rm=T)
qn_cat4=quantile(qn_num,0.8,na.rm=T)
df$total_ab_qn_5=ifelse(is.na(df$total_ab),0,
ifelse(df$total_ab<=qn_cat1,1,
ifelse(df$total_ab<=qn_cat2,2,
ifelse(df$total_ab<=qn_cat3,3,
ifelse(df$total_ab<=qn_cat4,4,5)
))))
### 1.2-broad ab quintile = according to unique broad ab prescription numbers
df$broad_ab_prescriptions=ifelse(df$broad_ab_prescriptions==0,NA,df$broad_ab_prescriptions) # filter no ab+no br ab
br_qn_num=unique(df$broad_ab_prescriptions)
br_qn_cat1=quantile(br_qn_num,0.2,na.rm=T)
br_qn_cat2=quantile(br_qn_num,0.4,na.rm=T)
br_qn_cat3=quantile(br_qn_num,0.6,na.rm=T)
br_qn_cat4=quantile(br_qn_num,0.8,na.rm=T)
df$br_total_ab_qn_5=ifelse(is.na(df$total_ab),0,
ifelse(df$broad_ab_prescriptions<=br_qn_cat1,1,
ifelse(df$broad_ab_prescriptions<=br_qn_cat2,2,
ifelse(df$broad_ab_prescriptions<=br_qn_cat3,3,
ifelse(df$broad_ab_prescriptions<=br_qn_cat4,4,5)
))))
df$br_total_ab_qn_5=ifelse(is.na(df$total_ab),"without any ab",
ifelse(is.na(df$broad_ab_prescriptions)|df$broad_ab_prescriptions==0,"without broad ab",
df$br_total_ab_qn_5))
### 2.1-ab quintile = according to total ab numbers
df$total_ab=ifelse(df$total_ab==0,NA,df$total_ab) # filter no ab
df=df%>%mutate(total_ab_qn=ntile(total_ab,5))
df$total_ab_qn=ifelse(is.na(df$total_ab_qn),0,df$total_ab_qn)# no ab ->0; ab exp. ->1~5
df$total_ab_qn=as.factor(df$total_ab_qn)
# ab_continuous
df$total_ab=ifelse(is.na(df$total_ab),0,df$total_ab) # recode NA to 0
######## confounding variables #########
## ethnicity
df$ethnicity=ifelse(is.na(df$ethnicity),"6",df$ethnicity)
df=df%>%mutate(ethnicity_6 = case_when(ethnicity == 1 ~ "White",
ethnicity == 2 ~ "Mixed",
ethnicity == 3 ~ "South Asian",
ethnicity == 4 ~ "Black",
ethnicity == 5 ~ "Other",
ethnicity == 6 ~ "Unknown"))
df$ethnicity_6=as.factor(df$ethnicity_6)
## BMI category
#bmi
#remove very low observations
df$bmi=ifelse(df$bmi<15 | df$bmi>50, NA, df$bmi)
# bmi categories
df<- df %>%
mutate(bmi_cat = case_when(is.na(bmi) ~ "Unknown",
bmi<18.5 ~ "Underweight",
bmi<25 ~ "Healthy weight",
bmi<30 ~ "Overweight",
bmi>=30 ~"Obese"))
df$bmi_cat<- as.factor(df$bmi_cat)
#summary(df_one_pat$bmi_cat)
##smoking status
df=df%>%mutate(smoking_cat_3= case_when(smoking_status=="S" ~ "Current",
smoking_status=="E" ~ "Former",
smoking_status=="N" ~ "Never",
smoking_status=="M" ~ "Unknown",
is.na(smoking_status) ~ "Unknown"))
### CCI
df$cancer<- ifelse(df$cancer_comor == 1, 2, 0)
df$cvd <- ifelse(df$cardiovascular_comor == 1, 1, 0)
df$copd <- ifelse(df$chronic_obstructive_pulmonary_comor == 1, 1, 0)
df$heart_failure <- ifelse(df$heart_failure_comor == 1, 1, 0)
df$connective_tissue <- ifelse(df$connective_tissue_comor == 1, 1, 0)
df$dementia <- ifelse(df$dementia_comor == 1, 1, 0)
df$diabetes <- ifelse(df$diabetes_comor == 1, 1, 0)
df$diabetes_complications <- ifelse(df$diabetes_complications_comor == 1, 2, 0)
df$hemiplegia <- ifelse(df$hemiplegia_comor == 1, 2, 0)
df$hiv <- ifelse(df$hiv_comor == 1, 6, 0)
df$metastatic_cancer <- ifelse(df$metastatic_cancer_comor == 1, 6, 0)
df$mild_liver <- ifelse(df$mild_liver_comor == 1, 1, 0)
df$mod_severe_liver <- ifelse(df$mod_severe_liver_comor == 1, 3, 0)
df$mod_severe_renal <- ifelse(df$mod_severe_renal_comor == 1, 2, 0)
df$mi <- ifelse(df$mi_comor == 1, 1, 0)
df$peptic_ulcer <- ifelse(df$peptic_ulcer_comor == 1, 1, 0)
df$peripheral_vascular <- ifelse(df$peripheral_vascular_comor == 1, 1, 0)
comor=c("cancer_comor","cardiovascular_comor", "chronic_obstructive_pulmonary_comor", "heart_failure_comor", "connective_tissue_comor", "dementia_comor", "diabetes_comor", "diabetes_complications_comor", "hemiplegia_comor", "hiv_comor", "metastatic_cancer_comor", "mild_liver_comor", "mod_severe_liver_comor", "mod_severe_renal_comor", "mi_comor", "peptic_ulcer_comor", "peripheral_vascular_comor")
df$Charlson=rowSums(df[comor])
write_rds(df[comor], here::here("output", "comor17.rds"))
df= df[!names(df)%in%comor]
df=df%>%mutate(CCI=case_when(Charlson<1 ~ "Very low",
Charlson<3 ~ "Low",
Charlson<5 ~ "Medium",
Charlson<7 ~ "High",
Charlson>=7 ~ "Very high"))
CCI_comor=c("cancer","cvd","copd","heart_failure","connective_tissue","dementia","diabetes","diabetes_complications","hemiplegia","hiv","metastatic_cancer","mild_liver","mod_severe_liver","mod_severe_renal","mi","peptic_ulcer","peripheral_vascular")
#df= df[!names(df)%in%CCI_comor]
# covid vaccine
df$covrx1_ever=ifelse(is.na(df$covrx1_dat),0,1)
df$covrx2_ever=ifelse(is.na(df$covrx2_dat),0,1)
df$covrx_ever=ifelse(df$covrx1_ever>0|df$covrx2_ever>0,1,0)
# variables for analysis
df=subset(df,select=c("wave","patient_index_date","patient_id","sex","age","age_cat","stp","region","ethnicity_6","bmi","bmi_cat","CCI","Charlson","smoking_cat_3","imd","care_home","covrx_ever","flu_vaccine","ab_types","interval", "lastABtime","ab_prescriptions","ab_qn_5","ab_qn","total_ab","total_ab_qn_5","total_ab_qn","broad_ab_prescriptions", "br_ab_qn","br_ab_qn_5","br_total_ab_qn_5","cancer","cvd","copd","heart_failure","connective_tissue","dementia","diabetes","diabetes_complications","hemiplegia","hiv","metastatic_cancer","mild_liver","mod_severe_liver","mod_severe_renal","mi","peptic_ulcer","peripheral_vascular"))
write_rds(df, here::here("output", "unmatched_outcome.rds"))
rm(list=ls())