-
Notifications
You must be signed in to change notification settings - Fork 3
/
data-processing.R
190 lines (135 loc) · 8.21 KB
/
data-processing.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
rm (list = ls())
library(tidyverse, warn.conflicts = FALSE)
library(plotly)
# Suppress summarise info
options(dplyr.summarise.inform = FALSE)
raw_data <- NULL
# Read the data from the meta-analysis repo
if(grepl('^/var/shiny/meta-analysis-shiny', getwd()) || grepl('/srv/shiny-server/meta-analyses-physical-activity', getwd())){
# Set encoding as found at: https://stackoverflow.com/a/14363274
# only for server
raw_data <- read.csv("../meta-analysis/data/R_variables_March23.csv", fileEncoding="latin1",
header = T, stringsAsFactors = F, skipNul = TRUE)
}else{
raw_data <- read.csv("../meta-analysis/data/R_variables_March23.csv", header = T,
stringsAsFactors = F, skipNul = TRUE)
}
local_pa_domain_subgroup <- "LTPA"
local_last_knot <- 0.75
ALT <- FALSE
NO_BMI_EFFECT <- FALSE
raw_data$mean_followup <- as.numeric(raw_data$mean_followup)
raw_data$n_baseline <- as.numeric(raw_data$n_baseline)
#raw_data[is.na(raw_data$n_baseline),]$n_baseline <- 0
raw_data[(is.na(raw_data$tot_personyrs)),]$tot_personyrs <-
raw_data[(is.na(raw_data$tot_personyrs)),]$mean_followup * raw_data[(is.na(raw_data$tot_personyrs)),]$n_baseline
raw_data[(is.na(raw_data$mean_followup)),]$mean_followup <-
raw_data[(is.na(raw_data$mean_followup)),]$tot_personyrs / raw_data[(is.na(raw_data$mean_followup)),]$n_baseline
raw_data$outcome <- trimws(raw_data$outcome)
raw_data$pa_domain_subgroup <- trimws(raw_data$pa_domain_subgroup)
raw_data$overall <- trimws(raw_data$overall)
raw_data$sex_subgroups <- trimws(raw_data$sex_subgroups)
raw_data$outcome_type <- trimws(raw_data$outcome_type)
raw_data$totalpersons <- as.numeric(raw_data$n_per_category)
raw_data$personyrs <- as.numeric(raw_data$person_years_per_category)
raw_data[raw_data$effect_measure == "RR",]$effect_measure <- "rr"
raw_data[raw_data$effect_measure == "HR",]$effect_measure <- "hr"
raw_data[raw_data$effect_measure == "OR",]$effect_measure <- "or"
raw_data$type <- ""
raw_data[raw_data$effect_measure == "or",]$type <- "ir"
raw_data[raw_data$effect_measure == "rr",]$type <- "ir"
raw_data[raw_data$effect_measure == "hr",]$type <- "ci"
raw_data$type <- as.character(raw_data$type)
raw_data[raw_data$outcome_type == "Both",]$outcome_type <- "Fatal-and-non-fatal"
raw_data$outcome_type <- as.character(raw_data$outcome_type)
## RENAME columns
raw_data <- raw_data %>% rename(effect = most_adj_effect, cases = cases_per_category,
uci_effect = most_adj_uci, lci_effect = most_adj_lci) %>%
mutate(cases = as.numeric(cases), uci_effect = as.numeric(uci_effect), lci_effect = as.numeric(lci_effect))
# ## FOR SENSITIVITY ANALYSIS
#raw_data$dose <- raw_data$Final.Harmonised.exposure..MMET.hrs.wk...FOR.SENSITIVITY.ANALYSIS
# MAIN <- TRUE
## FOR THE CURRENT ASSUMPTIONS
# raw_data$dose <- round(ifelse(ALT, raw_data$m_met_h_wk_alt, raw_data$m_met_h_wk), 2)
if (ALT){
raw_data$dose <- round(raw_data$m_met_h_wk_alt, 2)
}else{
raw_data$dose <- round(raw_data$m_met_h_wk, 2)
}
if (NO_BMI_EFFECT){
raw_data$effect <- as.numeric(raw_data$no_bmi_effect)
raw_data$lci_effect <- as.numeric(raw_data$no_bmi_lci)
raw_data$uci_effect <- as.numeric(raw_data$no_bmi_uci)
}
#raw_data$Final.Harmonised.exposure..MMET.hrs.wk. <- NULL
raw_data$RR <- raw_data$effect
raw_data <- subset(raw_data, select = c(ref_number, first_author, outcome, outcome_type, sex_subgroups, type, n_baseline, totalpersons, tot_personyrs, personyrs,
mean_followup, dose, RR, effect, effect_measure, lci_effect, uci_effect, cases, overall, pa_domain_subgroup))
## Populate missing totalpersons and personyrs
for (i in unique(raw_data$ref_number)){
raw_data[!is.na(raw_data$n_baseline) & raw_data$ref_number == i & (is.na(raw_data$totalpersons)) & !(is.na(raw_data$personyrs)) & !(is.na(raw_data$tot_personyrs)),]$totalpersons <-
round(raw_data[!is.na(raw_data$n_baseline) & raw_data$ref_number == i & (is.na(raw_data$totalpersons)) & !(is.na(raw_data$personyrs)) & !(is.na(raw_data$tot_personyrs)),]$personyrs /
sum(raw_data[!is.na(raw_data$n_baseline) & raw_data$ref_number == i & (is.na(raw_data$totalpersons)) & !(is.na(raw_data$personyrs)) & !(is.na(raw_data$tot_personyrs)),]$tot_personyrs) *
raw_data[!is.na(raw_data$n_baseline) & raw_data$ref_number == i & (is.na(raw_data$totalpersons)) & !(is.na(raw_data$personyrs)) & !(is.na(raw_data$tot_personyrs)),]$n_baseline)
raw_data[!is.na(raw_data$n_baseline) & raw_data$ref_number == i & (!is.na(raw_data$totalpersons)) & (is.na(raw_data$personyrs)) & !(is.na(raw_data$cases)),]$personyrs <-
round(raw_data[!is.na(raw_data$n_baseline) & raw_data$ref_number == i & (!is.na(raw_data$totalpersons)) & (is.na(raw_data$personyrs)) & !(is.na(raw_data$cases)),]$cases /
sum(raw_data[!is.na(raw_data$n_baseline) & raw_data$ref_number == i & (!is.na(raw_data$totalpersons)) & (is.na(raw_data$personyrs)) & !(is.na(raw_data$cases)),]$cases) *
raw_data[!is.na(raw_data$n_baseline) & raw_data$ref_number == i & (!is.na(raw_data$totalpersons)) & (is.na(raw_data$personyrs)) & !(is.na(raw_data$cases)),]$tot_personyrs)
}
raw_data$outcome <- stringi::stri_trans_totitle(raw_data$outcome, opts_brkiter = stringi::stri_opts_brkiter(type = "sentence"))
raw_data[raw_data$outcome == 'All-cause cvd',]$outcome <- 'All-cause CVD'
# Identify unique outcomes
uoutcome <- data.frame(outcome = as.character(unique(raw_data$outcome)))
uoutcome$outcome <- as.character(uoutcome$outcome)
# Sort
uoutcome$outcome <- sort(uoutcome$outcome)
# Remove the blank outcome
uoutcome <- dplyr::filter(uoutcome, outcome != "")
# All-cause mortality
# Cardiovascular diseases
# Coronary heart disease
# Stroke
# Breast cancer
# Colon cancer
# Endometrial cancer
# Lung cancer
# Total cancer
#uoutcome$outcome <- uoutcome[c(1, 3, 5, 9, 2, 4, 6, 8, 10, 7),]
overall_pop_tbles <- read_csv("../meta-analysis/data/csv/MA-DR/combined_tables.csv")
gender_pop_tbles <- read_csv("../meta-analysis/data/csv/MA-DR/combined_tables_by_gender.csv")
pa_exposure <- "LTPA"
#data$outcome <- stringi::stri_trans_totitle(data$outcome)
# Read the functions from the meta-analysis repo
source("../meta-analysis/script/all-functions.R")
# Filter by overall 1 for total population
raw_data_tp_ltpa <- subset(raw_data, pa_domain_subgroup == "LTPA" & overall == 1)
# Filter by sex subgroups for both men and female
raw_data_gsp_ltpa <- subset(raw_data, pa_domain_subgroup == "LTPA" & (sex_subgroups %in% c(1,2)))
for (i in 1:nrow(uoutcome)){
if (!uoutcome$outcome[i] %in% c('Breast cancer','Endometrial cancer')){
dat <- subset(raw_data_gsp_ltpa, outcome == uoutcome$outcome[i])
uid <- unique(dat$ref_number)
for (j in 1:length(uid)){
dat1 <- subset(dat, ref_number == uid[j])
usexgroups <- unique(dat1$sex_subgroups)
if (length(usexgroups) == 1){
# Remove single gender specific studies
raw_data_gsp_ltpa <- subset(raw_data_gsp_ltpa, (ref_number != uid[j]))
}
}
}
}
## Create ref_number for men and women subgroups
## for total population
if (nrow(raw_data_tp_ltpa[raw_data_tp_ltpa$overall != 1 & raw_data_tp_ltpa$sex_subgroups == 1,]) > 0)
raw_data_tp_ltpa[raw_data_tp_ltpa$overall != 1 & raw_data_tp_ltpa$sex_subgroups == 1,]$ref_number <- paste(raw_data_tp_ltpa[raw_data_tp_ltpa$overall != 1 & raw_data_tp_ltpa$sex_subgroups == 1,]$ref_number, "-1")
if (nrow(raw_data_tp_ltpa[raw_data_tp_ltpa$overall != 1 & raw_data_tp_ltpa$sex_subgroups == 2,]) > 0)
raw_data_tp_ltpa[raw_data_tp_ltpa$overall != 1 & raw_data_tp_ltpa$sex_subgroups == 2,]$ref_number <- paste(raw_data_tp_ltpa[raw_data_tp_ltpa$overall != 1 & raw_data_tp_ltpa$sex_subgroups == 2,]$ref_number, "-2")
# Create ID column
raw_data_tp_ltpa$id <- as.integer(as.factor(raw_data_tp_ltpa$ref_number))
raw_data_tp_ltpa <- plyr::arrange(raw_data_tp_ltpa, outcome)
raw_data_gsp_ltpa[raw_data_gsp_ltpa$sex_subgroups == 1,]$ref_number <- paste0(raw_data_gsp_ltpa[raw_data_gsp_ltpa$sex_subgroups == 1,]$ref_number, "-1")
raw_data_gsp_ltpa[raw_data_gsp_ltpa$sex_subgroups == 2,]$ref_number <- paste0(raw_data_gsp_ltpa[raw_data_gsp_ltpa$sex_subgroups == 2,]$ref_number, "-2")
# Create ID column
raw_data_gsp_ltpa$id <- as.integer(as.factor(raw_data_gsp_ltpa$ref_number))
raw_data_gsp_ltpa <- plyr::arrange(raw_data_gsp_ltpa, outcome)