R/plot_recruitment.R

In suvi-subra/viewTRIAL: Centralized monitoring RShiny template

#' Plot recruitment curve
#'
#' This function plots a recruitment curve by centers
#'
#' @param data dataframe
#' @param date.var variable containing dates
#' @param centre.var variable containing centers
#'
#' @return list of dataframes from secuTrial export
#' @import ggplot2
#' @import plotly
#' @import magrittr
#' @export
plot_recruit <- function(data, date.var, centre.var){

  enr <- data

  ## Order by signed consent date and select only 2 columns
  df <- enr[order(enr[[date.var]]), c(date.var, centre.var)]
  ## Include row numbers for df
  df <- data.frame(df, 1:nrow(enr))
  ## Allocate to another df name
  df.center <- df
  ## Group by center and allocate row id
  for(i in 1:nrow(df.center)) {
    df.center[i, "n"] <- sum(df.center[1:i, centre.var] == df.center[i, centre.var])
  }
  ## Allocate factor levels for centers
  #df.center$center <- factor(df.center$centre, levels = c("All", sort(levels(df.center$centre))))
  df.center.all <- df.center
  ## Get levels of centers without "All"
  ctr.levels <- levels(factor(df.center[[centre.var]]))

  df.center.all$centre.short <- "All"
  df.center.all$n <- df.center.all[,3]
  df.center <- rbind(df.center, df.center.all)
  names(df.center) <- c("Baseline", "Center", "total", "n")
  df.center$Center <- factor(df.center$Center, levels = c("All", ctr.levels))


  p <- ggplot() + geom_line(mapping = aes(x = Baseline, y = n, color = Center), data = df.center) +
    labs(x = "Date of randomization", y = "Patients included") + theme_bw()
  ggplotly(p)
}

get_recruitment_plot_df <- function(df){

  plot.df <- df %>%
    select(centre.short, rando_date.date) %>% arrange(rando_date.date) %>%
    ## Determine the  no of patients in EACH center
    group_by(centre.short) %>%
    mutate(n = rank(rando_date.date, ties.method = "first")) %>%
    ## Create new centre var with (n=)
    mutate(centre.n = paste0(centre.short, " (n=", max(n),")")) %>%
    ## Save output to bind later
    {. ->> output} %>%
    ungroup() %>%
    select(centre.n, rando_date.date) %>%
    ## Relabel all centers with "All" and determin total across all centers
    mutate(centre.n = paste0("All (n=", nrow(df),")"), n = 1:nrow(.)) %>%
    ## Bind BY center df to ALL center df
    bind_rows(output) %>%
    ## Factor levels according to frequency
    mutate(centre.n = forcats::fct_infreq(centre.n)) %>%
    ## Rename cols
    select("Center" = centre.n, "Baseline" = rando_date.date, n)

  return(plot.df)
}