slide 13 Readmission in 7 days

gistfile1.txt

####################################
# Readmission_7days ################
####################################

library(tidyverse)
library(lubridate)

data <- read.csv("/df.csv")

data$X <- NULL

#######################################################

# Code that converts all the columns names to lowercase 

df_tolower <-  function(x) { 
  colnames(x) <- tolower(colnames(x)) 
  x
  
}

#####################################################

df <- df_tolower(data )


################################################################################################################
# subseting data set 
################################################################################################################
dt_rmds<- df[c("pat_code","start_datetime", "end_datetime", "spell.type", "spell.number", "episode.order")]

colname_ct <- function(df, colname){
  df[,colname] = as.POSIXct(df[,colname])
  df
}


dt_rmds_ct <- dt_rmds %>%
  colname_ct("start_datetime") %>%
  colname_ct("end_datetime")

str(dt_rmds_ct)

dt_rmds_dates <-subset(dt_rmds_ct, (start_datetime >= "2013-05-01" & end_datetime <= "2014-10-01")) 

#################################################################################################################
# first, we create a table that contains the csn of the relevant cases
# and the days since last admission

#dt_rmds_4000 <- dt_rmds_dates[4000:5,]

#dt_rdms_40_filt_ep <- dt_rmds_40 %>%
  #dplyr::filter(spell.type == "Emergency")

dt_final <- dt_rmds_dates %>% # we take our data frame
  mutate(readm_date = as_date(end_datetime)) %>% # turn the dates into date format
  filter(spell.type == "Emergency") %>% # filter out cases within the time period wanted
  arrange(pat_code, readm_date) %>% # sort them first by mrn and then by admit_date
  group_by(pat_code) %>% # group them by mrn so we can for each patient...
  mutate(daysSinceLastAdmit = readm_date - lag(readm_date)) %>% # ...get the days since last admit
  mutate(daysSinceLastAdmit = as.integer(daysSinceLastAdmit)) %>% # turn this into an integer
  filter(daysSinceLastAdmit <= 7|is.na(daysSinceLastAdmit)) %>% # & daysSinceLastAdmit > 0???? keep the NA's as well yet get rid of zeros as this reveals there are no readmissions  
  mutate(readmissions_7days = if_else(is.na(daysSinceLastAdmit), "no", "yes")) %>% #create the wanted variable
  ungroup() %>% # ungroup it (must be done - don't ask)
  mutate(one_month = lubridate::round_date(start_datetime, "1 month", "month")) %>%
  select(pat_code, daysSinceLastAdmit, end_datetime, one_month, readmissions_7days, spell.type ) #readmissions_7days, start_datetime, one_month

#########################
#rounding the months ####
#########################

df_readm_7d <- dt_final %>%
  mutate(one_month = lubridate::round_date(end_datetime, "1 month", "month")) %>%
  filter(spell.type == "Emergency")

###################################################################
# taking out the NA and rows with zero and days bigger than 7 #####
###################################################################
df_na_omit <- dt_final  %>%
  filter(daysSinceLastAdmit > "0" & daysSinceLastAdmit <= "7")

# counting the Readmissions 
sum_readm <- df_na_omit %>%
  group_by(one_month) %>%
  dplyr::summarise(Readm = n()) 
 
# counting the discharges 

df_count_disc <- dt_final %>%
  group_by(one_month) %>%
  tally()

##############################################
# and joining with the readmission dataset ###
##############################################
df_read_and_count <- left_join(df_count_disc, sum_readm, by = "one_month")

######################################################
#This is in case we want all the dates on the plot ###
######################################################
# Set levels for date variable -- ensure they are unique
#TwoWeeks_levels <- format( unique (dt_final$TwoWeeks), "%d %m %Y")

#dt_final1 <- sum_readm %>%
  #convert date variable t a facot with set levels
  #mutate(one_month = factor(format(one_month, "%d %m %Y"), levels = one_month_levels))
######################################################  
#pct$data$x <- as.Date(pct$data$x, tz = 'Europe/London')
#pct1 <- ggplot(pct$data, aes(x,y, label =x))

#pct1


library(qicharts2)

######################################################
pct <- qic(Readm,
           x = one_month, 
           data = df_read_and_count, 
           chart = 'pp',
           #standardised = TRUE,
           #multiply= 100,
           title = "Readmissions within 7 days, counts n*",
           ylab = "Percent patients",
           xlab = "Readmissions within 7 days", 
           x.angle = 45) 

pct


#Error in qic(Count_readm_7days, n = Count_discharges, x = one_month, data = dt_join,  : 
#unused argument (standardised = TRUE)


####################################################
#Readmission in 7 days 
####################################################
plot3 <- qicharts2::qic(Readm,
                        n = n,
                        x = one_month, 
                        data = df_read_and_count, 
                        chart = 'pp',
                        ylab = "percent",
                        #freeze = 1,
                        #part = 11,
                        multiply = 100,
                        #x.format = "%d-%m-%Y", 
                        x.angle = 45,
                        xlab = "Readmissions within 7 days, Percentages %", 
                        title = "Readmissions within 7 days, 1st May, 2013 - 10th of October, 2014") 


plot3

p3 <- plot3$data


#######################################################################
# This isn't necessary but really hope you can help me finding  #######
# an answer to how I can plot all the dates                     #######
#######################################################################
# trying to plot the qic chart with ggplot                       ######
# so that I will be able to see all dates with their variation   ######
#######################################################################
library(ggplot2)
library(plyr)

p3.1 <- rename(p3, c("x" = "one_month"))
p3.1$one_month <- as.character(p3.1$x, format="%Y-%m-%d")

plot4 <- ggplot(p3.1, aes(x = one_month,y = y )) +
  geom_ribbon(ymin = p3.1$lcl, ymax = p3.1$ucl, alpha = 0 ) +   # fill = "" #
  geom_line(colour = "red", size = .75) + 
  geom_line(aes(one_month, cl)) +
  geom_point(colour = "blue" , fill = "blue", size = 1.5) +
  #x.axis(1, p3$x, format(p3$x, "%Y-%m-%d"), cex.axis = 0.7)+
  ggtitle(label = "Readmissions within 7 days") +
  labs(x = NULL,
       y = NULL)+
  theme_minimal()+
  theme(axis.text.x = element_text(angle = 45 , hjust = 1)) + #hjust = 1
  theme_bw()
 

plot4