R/00-data.R

In SubgrPlots: Graphical Displays for Subgroup Analysis in Clinical Trials

#' Prostate Cancer Dataset
#'
#' a prostate carcinoma dataset from a clinical trial David P. Byar and
#' Sylvan B. Green. The choice of treatment for cancer patients based on
#' covariate information: application to prostate cancer.
#' Bulletin du Cancer, 67:477–490, 1980.
#' The data can be found in the web:
#' \url{http://portal.uni-freiburg.de/imbi/Royston-Sauerbrei-book/index.html}.
#' We downloaded it form the supplementary material of Rosenkranz (2016)
#' https://onlinelibrary.wiley.com/doi/full/10.1002/bimj.201500147
#'  modified the data to keep relevant variables, and created categorical
#'  ones from age and weight.
#'
#' @usage data(prca)
#' @format A data frame with 475 rows and 15 variables:
#' \describe{
#'   \item{survtime}{survival time. Response variable}
#'   \item{cens}{The status indicator, 0=alive, 1=dead}
#'   \item{rx}{treatment received, 0=control, 1=treatment}
#'   \item{bm}{existence of bone metastasis}
#'   \item{hx}{history of cardiovascular events}
#'   \item{stage}{disease stage (3 or 4)}
#'   \item{pf}{performance}
#'   \item{age}{age}
#'   \item{weight}{weight in kg.}
#'   \item{age1}{dummy variable for 65 <= age < 75}
#'   \item{age2}{dummy variable for age >= 75}
#'   \item{wt1}{dummy variable for 90 <= weight < 110}
#'   \item{wt2}{dummy variable for weight >= 110}
#'   \item{age_group}{age categorized in 3 groups}
#'   \item{weight_group}{weight categorized in 3 groups}
#' }
#' @source \url{http://portal.uni-freiburg.de/imbi/Royston-Sauerbrei-book/index.html}
#' @source \url{https://onlinelibrary.wiley.com/doi/full/10.1002/bimj.201500147}
#'
#' @examples
#' \dontrun{
#' # Code used to download the dataset and create variables
#' library(haven)
#' l1 <- "https://onlinelibrary.wiley.com/action/"
#' l2 <- "downloadSupplement?doi=10.1002%2Fbimj.201500147&attachmentId=2173117089"
#' data_url <- paste0(l1,l2)
#' temp <- tempfile()
#' download.file(data_url,temp)
#' prca0 <- read_sas(unz(temp, "adv_prostate_ca.sas7bdat"))
#' # Select the variables that we use for the analysis
#' prca <- prca0[,c("SURVTIME","CENS","RX","BM","HX","STAGE","PF", "AGE", "WT")]
#'
#' # Change names of variables to lower case
#' names(prca)<- c("survtime","cens","rx","bm",
#'                 "hx","stage","pf","age", "wt")
#'
#' # Create subgroups for Age and Weight and Stage
#' prca$age1 <- 1 * (prca$age > 65 & prca$age <= 75)
#' prca$age2 <- 1 * (prca$age > 75)
#' prca$wt1  <- 1 * (prca$wt > 90 & prca$wt <= 110)
#' prca$wt2  <- 1 * (prca$wt > 110)
#'
#' # Create subgroups for Age and Weight and Stage with (-1,1) coding
#' prca$agegroup <- 1 + (1 * (prca$age > 65) + 1 * (prca$age > 75))
#' prca$wtgroup  <- 1 + (1 * (prca$wt > 90) + 1 * (prca$wt > 110))
#' dat = prca
#' dat$agegroup = factor(dat$agegroup)
#' dat$wtgroup = factor(dat$wtgroup)
#' range(dat$age)
#' range(dat$wt)
#' levels(dat$agegroup) = c("[48,65]","(65,75]","(75,89]")
#' levels(dat$wtgroup)  = c("[69,90]","(90,110]","(110,152]")
#' ## We need variables as factors
#' dat$bm    = factor(dat$bm)
#' dat$hx    = factor(dat$hx)
#' dat$stage = factor(dat$stage)
#' dat$pf    = factor(dat$pf)
#' dat$rx    = factor(dat$rx) # Treatment
#'
#' # Put labels to the variables so that they appear in the plot
#' names(dat)<- c("survtime",
#'                "cens",
#'                "rx",
#'                "bm",
#'                "hx",
#'                "stage",
#'                "pf",
#'                "age",
#'                "weight",
#'                "age1",
#'                "age2",
#'                "wt1",
#'                "wt2",
#'                "age_group",
#'                "weight_group")
#' prca <- dat
#' ## devtools::use_data(prca, overwrite = T) ## Use it as dataset for the package
#' }

"prca"