generated from opensafely/research-template
/
model.R
106 lines (91 loc) · 4 KB
/
model.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
# qFIT analysis for Gastro COVID-19 recovery project
# Authors: Nick Kennedy (NikNakk), Gareth Walker
library(tidyverse)
library(lubridate)
source("analysis/sro_gastro_functions.R")
# Read data output from cohortextractor
input <- read_csv("output/input.csv", col_types = cols(.default = col_character())) %>%
mutate(
across(matches("^(qfit|ft.*|crc.*)_date$"), lubridate::ymd),
across(c(ft_referral, ft_clinic, crc_diagnosis), ~as.logical(as.integer(.x))),
across(c(qfit, age, fobt_num), as.numeric),
qfit_pos = qfit > 10
)
# Counts of qFIT by STP
input %>%
group_by(stp) %>%
summarise(
n_people = n(),
age_overall = med_iqr(age),
female = sum(sex == "F"),
had_qfit = sum(!is.na(qfit_date)),
qfit_result = med_iqr(qfit[!is.na(qfit_date)]),
qfit_gt_10 = sum(qfit > 10, na.rm = TRUE),
had_ft_colorectal_referral = sum(ft_referral, na.rm = TRUE),
had_ft_colorectal_clinic = sum(ft_clinic, na.rm = TRUE),
had_colonoscopy = sum(!is.na(colonoscopy_sus)),
had_flexi_sig = sum(!is.na(flexi_sig_sus)),
had_crc_diagnosis = sum(crc_diagnosis, na.rm = TRUE),
referral_if_qfit_gt10 = sum(!is.na(ft_referral_date[qfit > 10 & !is.na(qfit_date)])),
referral_if_qfit_le10 = sum(!is.na(ft_referral_date[qfit <= 10 & !is.na(qfit_date)])),
referral_if_qfit_not_done = sum(!is.na(ft_referral_date[is.na(qfit_date)])),
qfit_referral_timing = med_iqr((qfit_date %--% ft_referral_date) / ddays(1), na.rm = TRUE),
qfit_clinic_timing = med_iqr((qfit_date %--% ft_clinic_date) / ddays(1), na.rm = TRUE),
qfit_colonoscopy_timing = med_iqr((qfit_date %--% colonoscopy_sus) / ddays(1), na.rm = TRUE),
qfit_flexi_sig_timing = med_iqr((qfit_date %--% flexi_sig_sus) / ddays(1), na.rm = TRUE),
qfit_diagnosis_timing = med_iqr((qfit_date %--% crc_diagnosis_date) / ddays(1), na.rm = TRUE),
qfit_age = med_iqr(age[!is.na(qfit_date)]),
qfit_female = sum(sex[!is.na(qfit_date)] == "F"),
top_four_fit_values = table(round(qfit[!is.na(qfit_date)], 0)) %>%
sort(decreasing = TRUE) %>%
head(4) %>%
sprintf("%s:%d", names(.), .) %>%
paste(collapse = "; ")
) %>%
write_csv("output/qfit_by_stp.csv")
qfits_by_stp <- input %>%
filter(!is.na(qfit_date)) %>%
count(stp, qfit = round(qfit)) %>%
ggplot(aes(qfit, n)) +
geom_col(fill = "dark blue") +
facet_wrap(quos(stp), scales = "free")
ggsave("output/qfits_by_stp.png", qfits_by_stp, width = 24, height = 24)
qfits_by_stp_near_10 <- input %>%
filter(!is.na(qfit_date)) %>%
count(stp, qfit = round(qfit)) %>%
ggplot(aes(qfit, n)) +
geom_col(fill = "dark blue") +
xlim(c(5, 15)) +
facet_wrap(quos(stp), scales = "free")
ggsave("output/qfits_by_stp_near_10.png", qfits_by_stp_near_10, width = 24, height = 24)
qfit_months_by_stp <- input %>%
filter(!is.na(qfit_date)) %>%
count(stp, qfit_month = floor_date(qfit_date, "months")) %>%
ggplot(aes(qfit_month, n)) +
geom_col(fill = "dark blue") +
facet_wrap(quos(stp), scales = "free")
ggsave("output/qfit_months_by_stp.png", qfit_months_by_stp, width = 24, height = 24)
fobts_by_stp <- input %>%
filter(!is.na(fobt_num_date)) %>%
count(stp, fobt_num = round(fobt_num)) %>%
ggplot(aes(fobt_num, n)) +
geom_col(fill = "dark blue") +
facet_wrap(quos(stp), scales = "free")
ggsave("output/fobts_by_stp.png", fobts_by_stp, width = 24, height = 24)
fobt_cat <- input %>%
filter(!is.na(fobt_date)) %>%
count(stp, fobt) %>%
ggplot(aes(fobt, n)) +
geom_col(fill = "dark blue") +
facet_wrap(quos(stp), scales = "free") +
coord_flip()
ggsave("output/fobts_by_stp_cat.png", fobt_cat, width = 24, height = 24)
fobt_fit_cat <- input %>%
filter(!is.na(fobt_date)) %>%
count(stp, has_fit = coalesce(fobt_date == qfit_date, FALSE), fobt) %>%
ggplot(aes(fobt, n, fill = has_fit)) +
geom_col(position = "stack") +
facet_wrap(quos(stp), scales = "free") +
coord_flip()
ggsave("output/fobts_by_stp_cat_fit.png", fobt_fit_cat, width = 24, height = 24)
# write_csv(as.data.frame(table(input$qfit)), "output/summary_qfit_values.csv")