library(readxl)
library(tidyverse)
library(ggsci)
library(gridExtra)
library(scales)
library(tableone)
library(knitr)
library(kableExtra)
library(gridExtra)
library(lemon)
library(extrafont)
#font_import()
loadfonts(device = "win")
windowsFonts(OpenSans = windowsFont("Open Sans"))
fct_case_when <- function(...) {
args <- as.list(match.call())
levels <- sapply(args[-1], function(f) f[[3]]) # extract RHS of formula
levels <- levels[!is.na(levels)]
factor(dplyr::case_when(...), levels=levels)
}
ds_raw <- read_excel("../data/data_ISP.xlsx", range = "A5:BW160", col_names = F)
ds_preproc_names <- ds_raw %>%
rename_all(~ c("id", "dob", "sex", "date_event", "date_adm", "etiology", "isp_icu", "diagnosis_t0", "crs_r_t0", "lcf_t0", "therapy_adm", "tracheo_t0", "breath_t0", "feed_t0", "cateth_t0", "bedsore_t0", "fractures_t0", "asworth_t0", "poa_t0", "craniolac_t0","hydrocef_t0", "erbi_sub_a_t0", "erbi_sub_b_t0", "psh_cfs_t0", "psh_dlt_t0", "psh_tot_t0", "nr_crisis_t0", "isp_treat_t0", "psh_cfs_t1", "psh_dlt_t1", "psh_tot_t1", "nr_crisis_t1", "isp_treat_t1", "diagnosis_t2", "crs_r_t2", "lcf_t2", "tracheo_t2", "breath_t2", "feed_t2", "cateth_t2", "bedsore_t2", "fractures_t2", "asworth_t2", "poa_t2", "craniolac_t2","hydrocef_t2", "erbi_sub_a_t2", "erbi_sub_b_t2", "psh_cfs_t2", "psh_dlt_t2", "psh_tot_t2", "nr_crisis_t2", "isp_treat_t2", "date_disch", "diagnosis_t3", "crs_r_t3", "lcf_t3", "tracheo_t3", "breath_t3", "feed_t3", "cateth_t3", "bedsore_t3", "asworth_t3", "poa_t3", "craniolac_t3","hydrocef_t3", "erbi_sub_a_t3", "erbi_sub_b_t3", "psh_cfs_t3", "psh_dlt_t3", "psh_tot_t3", "nr_crisis_t3", "isp_treat_t3", "therapy_disch", "gos")) %>%
mutate(id = str_remove(id, " "))
ds_preproc <- ds_preproc_names %>%
filter(!(is.na(id))) %>%
mutate(sex = factor(sex),
#diagnosis = factor(recode(diagnosis, "EMERSIONE" = "Emersion", "SMC" = "MCS", "SCM" = "MCS", "SV" = "VS")),
date_event = replace(date_event, id == "A43", as.Date("2018-11-05")),
date_event = replace(date_event, id == "B28", as.Date("2018-12-29")),
date_adm = as.Date(as.numeric(replace(date_adm, id == "A48", 43493)), origin = "1899-12-30"),
date_adm = replace(date_adm, id == "A43", as.Date("2019-01-08")),
date_disch = replace(date_disch, id == "N14", as.Date("2018-10-18")),
date_disch = replace(date_disch, id == "O17", as.Date("2019-04-18")),
date_disch = replace(date_disch, id == "F5", as.Date("2019-06-19")),
date_disch = replace(date_disch, id == "M3", as.Date("2019-05-05")),
etiology = fct_case_when(
etiology == 1 ~ "TBI",
etiology == 2 ~ "Vascular",
etiology == 3 ~ "Anoxic",
etiology == 4 ~ "Infective",
etiology == 5 ~ "Neoplastic",
etiology == 6 ~ "Mixed"),
psh_tot_t3 = as.numeric(recode(psh_tot_t3, "<8" = "1"))) %>%
mutate_at(.vars = vars(contains("diagnosis")),
.funs = list(~factor(recode(., "EMERSIONE" = "Emersion", "SMC" = "MCS", "SCM" = "MCS", "SV" = "VS", "DIMESSO" = "Discharged", "DIMESSA" = "Discharged", "DECESSO" = "Dead")))) %>%
mutate_at(.vars = vars(contains("crs_r")),
.funs = list(as.numeric)) %>%
mutate_at(.vars = vars(contains("erbi_sub_a")),
.funs = list(~case_when(
. > 0 ~ .*-1,
TRUE ~ .))) %>%
mutate_at(.vars = vars(contains("breath")),
.funs = list(~factor(case_when(
. == 1 ~ "Autonomous",
. == 2 ~ "Autonomous + O2",
. == 3 ~ "Mechanical"), levels=c("Autonomous", "Autonomous + O2", "Mechanical")))) %>%
mutate_at(.vars = vars(contains("feed")),
.funs = list(~fct_case_when(
. == 4 ~ "Oral",
. == 1 ~ "Parenteral",
. == 2 ~ "NG-tube",
. == 3 ~ "PEG"))) %>%
mutate_at(.vars = vars(therapy_adm, isp_treat_t0, isp_treat_t1, isp_treat_t2, isp_treat_t3, therapy_disch),
.funs = list(~str_replace_all(., c("[\\.]|[;]" = ",", "O" = "0", ",$" = "", "4,5999999999999996" = "4,6", "4.5999999999999996" = "4.6")))) %>%
mutate_at(.vars = vars(dob, date_event, date_disch),
.funs = list(as.Date.POSIXct)) %>%
mutate_at(.vars = vars(contains("nr_crisis")),
.funs = list(~fct_case_when(
. == 0 ~ "none",
. == 1 ~ "<3",
. == 2 ~ "3-5",
. == 3 ~ ">5"))) %>%
mutate_at(.vars = vars(isp_icu, tracheo_t0, cateth_t0:hydrocef_t0, tracheo_t2, cateth_t2:hydrocef_t2, tracheo_t3, cateth_t3:hydrocef_t3),
.funs = list(~factor(recode(., "N0" = "no", "NO" = "no", "SI," = "yes", "SI" = "yes", "si" = "yes", "SI, NO" = "no", "BO" = "no", "NO DVP" = "no", "SI DVP" = "yes"))))
ds_therapy <- ds_preproc %>%
select(id, therapy_adm, isp_treat_t0, isp_treat_t1, isp_treat_t2, isp_treat_t3, therapy_disch) %>%
separate(therapy_adm, into = paste("adm", 1:5, sep = "_")) %>%
separate(isp_treat_t0, into = paste("t0", 1:5, sep = "_")) %>%
separate(isp_treat_t1, into = paste("t1", 1:5, sep = "_")) %>%
separate(isp_treat_t2, into = paste("t2", 1:5, sep = "_")) %>%
separate(isp_treat_t3, into = paste("t3", 1:5, sep = "_")) %>%
separate(therapy_disch, into = paste("disch", 1:5, sep = "_")) %>%
gather(treatment, n, adm_1:disch_5, na.rm = T) %>%
mutate(n = fct_case_when(
n == 0 ~ "none",
n == 1 ~ "opioids",
n == 2 ~ "benzos",
n == 3 ~ "barbiturate",
n == 4 ~ "beta_blockers",
n == 5 ~ "alpha_blockers",
n == 6 ~ "dopa_antagonist",
n == 7 ~ "baclofen",
n == 8 ~ "other"),
treatment = str_replace(treatment, pattern = '_.*', replacement = ''),
val = 1)
ds_treatments <- ds_therapy %>%
expand(id, treatment, n) %>%
unite(treat, treatment, n, sep = "_") %>%
anti_join(ds_therapy %>%
select(-val) %>%
unite(treat, treatment, n, sep = "_")) %>%
mutate(val = 0)
ds_therapy_wide <- ds_therapy %>%
unite(treat, treatment, n, sep = "_") %>%
bind_rows(ds_treatments) %>%
spread(key = treat, value = val, fill = 0) %>%
select(id:adm_other, t0_alpha_blockers:t3_other, everything())
ds_proc <- ds_preproc %>%
select(-c(therapy_adm, isp_treat_t0, isp_treat_t1, isp_treat_t2, isp_treat_t3, therapy_disch)) %>%
left_join(ds_therapy_wide) %>%
mutate_at(vars(contains("psh_tot")),
.funs = list(cat = ~cut(., breaks = c(0,7,16,100), labels = c("Unlikely", "Possible", "Probable")))) %>%
mutate(etiology = factor(etiology),
days_icu = as.numeric(date_adm - date_event),
days_rehab = as.numeric(date_disch - date_adm),
days_event = as.numeric(date_disch - date_event),
rehab_center = factor(str_sub(id, end = 1)),
age_event = round(as.numeric(date_event - dob)/365.25, 1),
gos_cat = fct_case_when(
gos == 1 ~ "Death",
gos == 2 ~ "Vegetative State",
gos == 3 ~ "Severe Disability",
gos == 4 ~ "Moderate Disability",
gos == 5 ~ "Low Disability"
)) %>%
filter(days_icu < 93)
ds_isp <- ds_proc %>%
select(id, rehab_center, age_event, sex, days_icu, etiology, isp_icu, adm_alpha_blockers:adm_other, diagnosis_t0:psh_tot_t0, psh_tot_t0_cat, nr_crisis_t0, t0_alpha_blockers:t0_other, psh_cfs_t1:psh_tot_t1, psh_tot_t1_cat, nr_crisis_t1, t1_alpha_blockers:t1_other, diagnosis_t2:psh_tot_t2, psh_tot_t2_cat, nr_crisis_t2, t2_alpha_blockers:t2_other, diagnosis_t3:psh_tot_t3, psh_tot_t3_cat, nr_crisis_t3, t3_alpha_blockers:t3_other, days_rehab, gos_cat, disch_alpha_blockers:disch_other)
Data description
format_table = function(data) {
data %>%
mutate(
var = trimws(var),
var = str_replace_all(var, "_", " "),
var = str_to_title(var),
var = str_remove(var, " T0"),
var = str_remove(var, " T2"),
var = str_remove(var, " T3"),
var = as.character(var),
var = case_when(
var == "Age Event (Mean (Sd))" ~ "Age Event (years)",
var == "Days Icu (Mean (Sd))" ~ "Length of stay ICU (days)",
var == "Days Rehab (Mean (Sd))" ~ "Length of stay IRU (days)",
var == "Lcf (Median [Iqr])" ~ "LCF Score",
str_detect(var,"Tracheo") ~ "Tracheostomy (%)",
var == "Breath (%)" ~ "Breathing (%)",
var == "Feed (%)" ~ "Feeding (%)",
str_detect(var,"Cateth") ~ "Catether (%)",
str_detect(var,"Asworth") ~ "Spasticity (%)",
str_detect(var,"Craniolac") ~ "DC (%)",
str_detect(var, "Mcs") ~ "MCS",
str_detect(var, "Vs") ~ "VS",
str_detect(var, "Tbi") ~ "TBI",
str_detect(var , "Ng-Tube") ~ "NG-Tube",
str_detect(var, "Peg") ~ "PEG",
str_detect(var, "Psh Tot Cat") ~ "PSH-AM",
TRUE ~ str_remove(var, "= Yes ")))
}
vars_tab1 <- c("age_event", "sex", "days_icu", "days_rehab", "etiology", "diagnosis_t0", "lcf_t0", "tracheo_t0", "breath_t0", "feed_t0", "cateth_t0", "bedsore_t0", "asworth_t0", "craniolac_t0", "psh_tot_t0_cat", "nr_crisis_t0")
ds_therapy_tab <- as.data.frame(
print(CreateTableOne(
vars = "treatment",
strata = "time",
data = ds_therapy %>%
rename(time = treatment, treatment = n) %>%
mutate(time = factor(time, labels = c("Admission", "Week 1", "Week 2", "Month 4", "Discharge-ISP", "Discharge")),
treatment = factor(treatment, labels = str_to_title(str_replace_all(levels(treatment), "_", " ")))),
factorVars = "treatment", test = F),
catDigits = 1, contDigits = 1, noSpaces = T, missing = F, printToggle = F)[-1,]) %>%
rename_at(1:6, ~ rep("Overall",6))
ds_tab1 <- as.data.frame(print(CreateTableOne(vars = vars_tab1, data = ds_isp), catDigits = 1, contDigits = 1, noSpaces = T, missing = F, printToggle = F, nonnormal = "lcf_t0")) %>%
#mutate_all(as.character) %>%
mutate(var = row.names(.),
Overall = case_when(
str_detect(var, "mean") ~ str_replace(str_replace(as.character(Overall), "\\(", "? "), "\\)", ""),
str_detect(var, "median") ~ str_replace(as.character(Overall), "\\,", " -"),
TRUE ~ as.character(Overall))) %>%
format_table() %>%
bind_rows(ds_therapy_tab[1] %>% mutate(var = str_to_title(row.names(.)))) %>%
select(var, everything()) %>%
rename(Admission = Overall)
# kable(ds_tab1, caption = "Table 1. Charasteristics of patients at admission in rehab centers") %>%
# kable_styling(bootstrap_options = c("striped", "hover", "condensed")) %>%
# add_indent(c(6:11, 13:15, 19:21, 23:26, 32:34, 36:39, 41:49))
vars_tab2 <- c("diagnosis_t2", "lcf_t2", "tracheo_t2", "breath_t2", "feed_t2", "cateth_t2", "bedsore_t2", "asworth_t2", "craniolac_t2", "psh_tot_t2_cat", "nr_crisis_t2")
ds_tab2 <- as.data.frame(print(CreateTableOne(vars = vars_tab2, data = ds_isp %>% filter(!diagnosis_t2 %in% c("Dead", "Discharged")) %>% mutate(diagnosis_t2 = factor(diagnosis_t2))), catDigits = 1, contDigits = 1, noSpaces = T, missing = F, printToggle = F, nonnormal = "lcf_t2")) %>%
mutate(var = row.names(.),
Overall = case_when(
str_detect(var, "mean") ~ str_replace(str_replace(as.character(Overall), "\\(", "? "), "\\)", ""),
str_detect(var, "median") ~ str_replace(as.character(Overall), "\\,", " -"),
TRUE ~ as.character(Overall))) %>%
format_table() %>%
bind_rows(ds_therapy_tab[2] %>% mutate(var = str_to_title(row.names(.)))) %>%
select(var, everything()) %>%
rename(`4th Month` = Overall)
# kable(ds_tab2, caption = "Table 2. Charasteristics of patients after 4 months from rehab start") %>%
# kable_styling(bootstrap_options = c("striped", "hover", "condensed")) %>%
# add_indent(c(3:5,9:11,13:16, 22:24, 26:29, 31:39))
vars_tab3 <- c("diagnosis_t3", "lcf_t3", "tracheo_t3", "breath_t3", "feed_t3", "cateth_t3", "bedsore_t3", "asworth_t3", "craniolac_t3", "psh_tot_t3_cat", "nr_crisis_t3", "days_rehab", "gos_cat")
ds_tab3 <- as.data.frame(print(CreateTableOne(vars = vars_tab3, data = ds_isp %>% filter(!diagnosis_t3 %in% c("Dead", "Discharged")) %>% mutate(diagnosis_t3 = factor(diagnosis_t3))), catDigits = 1, contDigits = 1, noSpaces = T, missing = F, printToggle = F, nonnormal = "lcf_t3")) %>%
mutate(var = row.names(.),
Overall = case_when(
str_detect(var, "mean") ~ str_replace(str_replace(as.character(Overall), "\\(", "? "), "\\)", ""),
str_detect(var, "median") ~ str_replace(as.character(Overall), "\\,", " -"),
TRUE ~ as.character(Overall))) %>%
format_table() %>%
bind_rows(ds_therapy_tab[6] %>% mutate(var = str_to_title(row.names(.)))) %>%
select(var, everything()) %>%
rename(Discharge = Overall)
# kable(ds_tab3, caption = "Table 3. Charasteristics of patients at discharge") %>%
# kable_styling(bootstrap_options = c("striped", "hover", "condensed")) %>%
# add_indent(c(3:5,9:11,13:16, 22:24, 26:29, 32:36, 38:46))
ds_tab123 <- ds_tab1 %>%
left_join(ds_tab2) %>%
left_join(ds_tab3) %>%
replace_na(list(`4th Month` = "", Discharge = ""))
kable(ds_tab123, caption = "Table 1. Charasteristics of patients at admission in rehab centers") %>%
kable_styling(bootstrap_options = c("striped", "hover", "condensed")) %>%
add_indent(c(7:11, 13:15, 19:21, 23:26, 32:34, 36:39, 41:49))
Table 1. Charasteristics of patients at admission in rehab centers
|
var
|
Admission
|
4th Month
|
Discharge
|
|
N
|
152
|
138
|
148
|
|
Age Event (years)
|
52.3 ? 18.0
|
|
|
|
Sex = M (%)
|
95 (62.5)
|
|
|
|
Length of stay ICU (days)
|
45.4 ? 20.1
|
|
|
|
Length of stay IRU (days)
|
164.4 ? 75.9
|
|
165.6 ? 75.6
|
|
Etiology (%)
|
|
|
|
|
TBI
|
54 (35.5)
|
|
|
|
Vascular
|
77 (50.7)
|
|
|
|
Anoxic
|
15 (9.9)
|
|
|
|
Infective
|
2 (1.3)
|
|
|
|
Neoplastic
|
4 (2.6)
|
|
|
|
Diagnosis (%)
|
|
|
|
|
Emersion
|
39 (25.7)
|
76 (55.1)
|
95 (64.2)
|
|
MCS
|
54 (35.5)
|
32 (23.2)
|
34 (23.0)
|
|
VS
|
59 (38.8)
|
30 (21.7)
|
19 (12.8)
|
|
LCF Score
|
3.0 [2.0 - 4.0]
|
4.0 [3.0 - 6.0]
|
5.0 [3.0 - 7.0]
|
|
Tracheostomy (%)
|
135 (88.8)
|
78 (56.5)
|
54 (36.5)
|
|
Breathing (%)
|
|
|
|
|
Autonomous
|
58 (40.6)
|
112 (81.8)
|
126 (85.1)
|
|
Autonomous + O2
|
80 (55.9)
|
25 (18.2)
|
21 (14.2)
|
|
Mechanical
|
5 (3.5)
|
0 (0.0)
|
1 (0.7)
|
|
Feeding (%)
|
|
|
|
|
Oral
|
11 (7.3)
|
55 (39.9)
|
74 (50.7)
|
|
Parenteral
|
2 (1.3)
|
1 (0.7)
|
1 (0.7)
|
|
NG-Tube
|
63 (41.7)
|
11 (8.0)
|
3 (2.1)
|
|
PEG
|
75 (49.7)
|
71 (51.4)
|
68 (46.6)
|
|
Catether (%)
|
147 (96.7)
|
60 (43.5)
|
44 (29.7)
|
|
Bedsore (%)
|
55 (36.2)
|
38 (27.5)
|
21 (14.2)
|
|
Spasticity (%)
|
31 (20.4)
|
45 (32.6)
|
43 (29.1)
|
|
DC (%)
|
32 (21.1)
|
23 (16.7)
|
13 (8.8)
|
|
PSH-AM
|
|
|
|
|
Unlikely
|
105 (69.5)
|
114 (83.2)
|
137 (93.2)
|
|
Possible
|
30 (19.9)
|
13 (9.5)
|
5 (3.4)
|
|
Probable
|
16 (10.6)
|
10 (7.3)
|
5 (3.4)
|
|
Nr Crisis (%)
|
|
|
|
|
None
|
113 (79.6)
|
101 (79.5)
|
116 (83.5)
|
|
<3
|
14 (9.9)
|
16 (12.6)
|
15 (10.8)
|
|
3-5
|
14 (9.9)
|
4 (3.1)
|
5 (3.6)
|
|
>5
|
1 (0.7)
|
6 (4.7)
|
3 (2.2)
|
|
Treatment (%)
|
|
|
|
|
None
|
24 (10.8)
|
32 (12.7)
|
119 (77.8)
|
|
Opioids
|
10 (4.5)
|
2 (0.8)
|
0 (0.0)
|
|
Benzos
|
22 (9.9)
|
35 (13.9)
|
12 (7.8)
|
|
Barbiturate
|
6 (2.7)
|
3 (1.2)
|
0 (0.0)
|
|
Beta Blockers
|
69 (31.1)
|
45 (17.9)
|
11 (7.2)
|
|
Alpha Blockers
|
16 (7.2)
|
4 (1.6)
|
1 (0.7)
|
|
Dopa Antagonist
|
3 (1.4)
|
24 (9.6)
|
0 (0.0)
|
|
Baclofen
|
17 (7.7)
|
22 (8.8)
|
6 (3.9)
|
|
Other
|
55 (24.8)
|
84 (33.5)
|
4 (2.6)
|
Descriptive Plots (Captions are below the figures)
cbPalette <- c(pal_nejm("default", alpha = 1)(8), "#000000")
ds_plot <- ds_isp %>%
select(id, c(vars_tab1, vars_tab2, vars_tab3), psh_tot_t1_cat, nr_crisis_t1) %>%
select(-c(age_event, sex, days_icu, etiology, days_rehab, gos_cat)) %>%
rename(psh_t0 = psh_tot_t0_cat, psh_t1 = psh_tot_t1_cat, psh_t2 = psh_tot_t2_cat, psh_t3 = psh_tot_t3_cat) %>%
gather(var, value, -id, na.rm = T, factor_key = T) %>%
separate(var, c("var", "time"), sep = "_t") %>%
mutate(time = fct_case_when(
time == "0" ~ "Admission",
time == "1" ~ "2nd Week",
time == "2" ~ "4th month",
time == "3" ~ "Discharge"
))
g <- ggplot(data = ds_plot %>%
filter(var %in% c("psh", "nr_crisis")) %>%
spread(var,value) %>%
group_by(time, nr_crisis, psh) %>%
summarise(n = n()) %>%
mutate(freq = n/sum(n)) %>%
ungroup() %>%
mutate(nr_crisis = factor(nr_crisis, levels = c("none", "<3", "3-5", ">5")),
psh = factor(psh, levels = c("Unlikely", "Possible", "Probable"))) %>%
filter(!is.na(nr_crisis)) %>%
filter(!is.na(psh)) %>%
complete(nr_crisis, nesting(time, psh), fill = list(n = 0, freq = 0)),
aes(x = psh, y = n, fill = nr_crisis))
g + geom_bar(stat="identity", position=position_dodge(), colour="black") +
scale_fill_manual(name = "Nr. ISP Crisis", values = cbPalette) +
#scale_y_continuous(trans='log2') +
labs(x = "PSH Total Score", y = "nr. of Patients") +
facet_wrap(~ time) +
theme_bw(base_family="OpenSans") +
theme(text = element_text(size=16),
axis.text.x = element_text(size=12),
legend.position = "bottom")
ds_out <- ds_isp %>%
select(id, contains("nr_crisis"),contains("psh_tot"), contains("gos")) %>%
mutate(isp_crisis = ds_isp %>%
select(contains("nr_crisis")) %>%
mutate_all(.funs = list(~ as.numeric(.) - 1)) %>%
mutate(isp = rowSums(., na.rm = T),
isp = factor(isp>0, c(F,T), c("none", "yes"))) %>%
pull(isp),
psh_positive = ds_isp %>%
select(contains("psh_tot")) %>%
select(contains("cat")) %>%
mutate_all(.funs = list(~ as.numeric(.) - 1)) %>%
mutate(psh = rowSums(., na.rm = T),
psh = factor(psh>0, c(F,T), c("negative", "positive"))) %>%
pull(psh),
gos_cat = fct_rev(gos_cat),
gos = as.numeric(gos_cat))
library(descr)
ct_t0 <- crosstab(ds_out$psh_tot_t0_cat, ds_out$nr_crisis_t0,
dnn = c("PSH", "Nr.Crisis"),
drop.levels = F, expected = F, prop.c = F, prop.r = TRUE, fisher = T, plot = FALSE)
ds_xtab_t0 <- as.data.frame(ct_t0$tab) %>%
spread(Nr.Crisis,Freq) %>%
mutate_if(is.numeric, as.character) %>%
bind_rows(as.data.frame(ct_t0$prop.row) %>%
spread(Nr.Crisis,Freq) %>%
mutate_if(is.numeric, .funs = list(~paste(round(.*100,2),"%", sep="")))) %>%
arrange(PSH)
kable(ds_xtab_t0, caption = "Table 2. Numbers of ISP Crisis by PSH Score at Admission") %>%
kable_styling(bootstrap_options = c("striped", "hover", "condensed")) %>%
collapse_rows(columns = 1, valign = "top") %>%
add_header_above(c(" ", "Nr. Crisis" = 4))
Table 2. Numbers of ISP Crisis by PSH Score at Admission
|
|
Nr. Crisis
|
|
PSH
|
none
|
<3
|
3-5
|
>5
|
|
Unlikely
|
104
|
1
|
0
|
0
|
|
99.05%
|
0.95%
|
0%
|
0%
|
|
Possible
|
8
|
12
|
7
|
0
|
|
29.63%
|
44.44%
|
25.93%
|
0%
|
|
Probable
|
0
|
1
|
7
|
1
|
|
0%
|
11.11%
|
77.78%
|
11.11%
|
ct_t1 <- crosstab(ds_out$psh_tot_t1_cat, ds_out$nr_crisis_t1,
dnn = c("PSH", "Nr.Crisis"),
drop.levels = F, expected = F, prop.c = F, prop.r = TRUE, fisher = T, plot = FALSE)
ds_xtab_t1 <- as.data.frame(ct_t1$tab) %>%
spread(Nr.Crisis,Freq) %>%
mutate_if(is.numeric, as.character) %>%
bind_rows(as.data.frame(ct_t1$prop.row) %>%
spread(Nr.Crisis,Freq) %>%
mutate_if(is.numeric, .funs = list(~paste(round(.*100,2),"%", sep="")))) %>%
arrange(PSH)
kable(ds_xtab_t1, caption = "Table 2. Numbers of ISP Crisis by PSH Score at Admission") %>%
kable_styling(bootstrap_options = c("striped", "hover", "condensed")) %>%
collapse_rows(columns = 1, valign = "top") %>%
add_header_above(c(" ", "Nr. Crisis" = 4))
Table 2. Numbers of ISP Crisis by PSH Score at Admission
|
|
Nr. Crisis
|
|
PSH
|
none
|
<3
|
3-5
|
>5
|
|
Unlikely
|
115
|
1
|
0
|
0
|
|
99.14%
|
0.86%
|
0%
|
0%
|
|
Possible
|
7
|
10
|
3
|
0
|
|
35%
|
50%
|
15%
|
0%
|
|
Probable
|
0
|
2
|
5
|
0
|
|
0%
|
28.57%
|
71.43%
|
0%
|
ct_t2 <- crosstab(ds_out$psh_tot_t2_cat, ds_out$nr_crisis_t2,
dnn = c("PSH", "Nr.Crisis"),
drop.levels = F, expected = F, prop.c = F, prop.r = TRUE, fisher = T, plot = FALSE)
ds_xtab_t2 <- as.data.frame(ct_t2$tab) %>%
spread(Nr.Crisis,Freq) %>%
mutate_if(is.numeric, as.character) %>%
bind_rows(as.data.frame(ct_t2$prop.row) %>%
spread(Nr.Crisis,Freq) %>%
mutate_if(is.numeric, .funs = list(~paste(round(.*100,2),"%", sep="")))) %>%
arrange(PSH)
kable(ds_xtab_t2, caption = "Table 2. Numbers of ISP Crisis by PSH Score at Admission") %>%
kable_styling(bootstrap_options = c("striped", "hover", "condensed")) %>%
collapse_rows(columns = 1, valign = "top") %>%
add_header_above(c(" ", "Nr. Crisis" = 4))
Table 2. Numbers of ISP Crisis by PSH Score at Admission
|
|
Nr. Crisis
|
|
PSH
|
none
|
<3
|
3-5
|
>5
|
|
Unlikely
|
99
|
13
|
0
|
0
|
|
88.39%
|
11.61%
|
0%
|
0%
|
|
Possible
|
1
|
3
|
2
|
2
|
|
12.5%
|
37.5%
|
25%
|
25%
|
|
Probable
|
0
|
0
|
2
|
4
|
|
0%
|
0%
|
33.33%
|
66.67%
|
ct_t3 <- crosstab(ds_out$psh_tot_t3_cat, ds_out$nr_crisis_t3,
dnn = c("PSH", "Nr.Crisis"),
drop.levels = F, expected = F, prop.c = F, prop.r = TRUE, fisher = T, plot = FALSE)
ds_xtab_t3 <- as.data.frame(ct_t3$tab) %>%
spread(Nr.Crisis,Freq) %>%
mutate_if(is.numeric, as.character) %>%
bind_rows(as.data.frame(ct_t3$prop.row) %>%
spread(Nr.Crisis,Freq) %>%
mutate_if(is.numeric, .funs = list(~paste(round(.*100,2),"%", sep="")))) %>%
arrange(PSH)
kable(ds_xtab_t3, caption = "Table 2. Numbers of ISP Crisis by PSH Score at Admission") %>%
kable_styling(bootstrap_options = c("striped", "hover", "condensed")) %>%
collapse_rows(columns = 1, valign = "top") %>%
add_header_above(c(" ", "Nr. Crisis" = 4))
Table 2. Numbers of ISP Crisis by PSH Score at Admission
|
|
Nr. Crisis
|
|
PSH
|
none
|
<3
|
3-5
|
>5
|
|
Unlikely
|
115
|
12
|
3
|
1
|
|
87.79%
|
9.16%
|
2.29%
|
0.76%
|
|
Possible
|
0
|
3
|
1
|
0
|
|
0%
|
75%
|
25%
|
0%
|
|
Probable
|
0
|
0
|
1
|
2
|
|
0%
|
0%
|
33.33%
|
66.67%
|
g1 <- ggplot(data = ds_out %>%
select(nr_crisis_t0, gos_cat) %>%
filter(!is.na(nr_crisis_t0)) %>%
group_by(nr_crisis_t0, gos_cat) %>%
summarise(n = n()) %>%
mutate(freq = n/sum(n)) %>%
complete(nr_crisis_t0, nesting(gos_cat), fill = list(n = 0, freq = 0)),
aes(x = nr_crisis_t0, y = n, fill = gos_cat)) + geom_bar(stat="identity", position=position_dodge(), colour="black") +
scale_fill_manual(name = "GOS", values = cbPalette) +
labs(title = "A", x = "ISP Crisis during Week 1", y = "nr. of Patients") +
theme_bw(base_family="OpenSans") +
theme(text = element_text(size=16))
g2 <- ggplot(data = ds_out %>%
select(psh_tot_t0_cat, gos_cat) %>%
filter(!is.na(psh_tot_t0_cat)) %>%
group_by(psh_tot_t0_cat, gos_cat) %>%
summarise(n = n()) %>%
mutate(freq = n/sum(n)) %>%
complete(psh_tot_t0_cat, nesting(gos_cat), fill = list(n = 0, freq = 0)),
aes(x = psh_tot_t0_cat, y = n, fill = gos_cat)) + geom_bar(stat="identity", position=position_dodge(), colour="black") +
scale_fill_manual(name = "GOS", values = cbPalette) +
labs(title = "B", x = "PSH results during Week 1", y = "nr. of Patients") +
theme_bw(base_family="OpenSans") +
theme(text = element_text(size=16),
legend.text = element_text(size=11),
legend.title = element_text(size=12))
grid_arrange_shared_legend(g1,g2, ncol = 1, nrow = 2, position = "bottom")
g1 <- ggplot(data = ds_out %>%
select(isp_crisis, gos_cat) %>%
group_by(isp_crisis, gos_cat) %>%
summarise(n = n()) %>%
mutate(freq = n/sum(n)) %>%
complete(isp_crisis, nesting(gos_cat), fill = list(n = 0, freq = 0)),
aes(x = isp_crisis, y = n, fill = gos_cat)) + geom_bar(stat="identity", position=position_dodge(), colour="black") +
scale_fill_manual(name = "GOS", values = cbPalette) +
labs(title = "A", x = "ISP Crisis during follow-up", y = "nr. of Patients") +
theme_bw(base_family="OpenSans") +
theme(text = element_text(size=16))
g2 <- ggplot(data = ds_out %>%
select(psh_positive, gos_cat) %>%
group_by(psh_positive, gos_cat) %>%
summarise(n = n()) %>%
mutate(freq = n/sum(n)) %>%
complete(psh_positive, nesting(gos_cat), fill = list(n = 0, freq = 0)),
aes(x = psh_positive, y = n, fill = gos_cat)) + geom_bar(stat="identity", position=position_dodge(), colour="black") +
scale_fill_manual(name = "GOS", values = cbPalette) +
labs(title = "B", x = "PSH positivity during study", y = "nr. of Patients") +
theme_bw(base_family="OpenSans") +
theme(text = element_text(size=16),
legend.text = element_text(size=11),
legend.title = element_text(size=12))
grid_arrange_shared_legend(g1,g2, ncol = 1, nrow = 2, position = "bottom")
