From 00a97a20d9e58d5de000feeab3866793c5c985e3 Mon Sep 17 00:00:00 2001
From: charlottebargain <charlotte.bargain@gmail.com>
Date: Wed, 5 Nov 2025 15:55:35 +0100
Subject: [PATCH 1/5] v1

---
 R/data.R | 144 ++++++++++++++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 143 insertions(+), 1 deletion(-)

diff --git a/R/data.R b/R/data.R
index a912803..b556b63 100644
--- a/R/data.R
+++ b/R/data.R
@@ -22,7 +22,9 @@ grstat_example = function(N=200, seed=42, ...){
 
   recist = example_rc(enrolres, seed, ...)
 
-  rtn = lst(enrolres, ae, recist) %>%
+  fu = example_fu(enrolres, recist, seed, ...)
+
+  rtn = lst(enrolres, ae, recist, fu) %>%
     imap(~.x %>% mutate(crfname=.y %>% set_label("Form name")))
   rtn$date_extraction = "2024/01/01"
   rtn$datetime_extraction = structure(1704067200, class = c("POSIXct", "POSIXt"),
@@ -159,6 +161,7 @@ example_ae = function(enrolres, seed, p_na=0,
 }
 
 
+
 #' Simulate a RECIST dataset
 #'
 #' Internal function that simulates a synthetic RECIST dataset following the
@@ -283,6 +286,145 @@ example_rc = function(enrolres, seed,
 }
 
 
+
+#' Generate Simulated Survival Data
+#'
+#' Internal function that simulates survival data, which depend on treatment arm
+#' and date of progression if applicable.
+#'
+#' @param enrolres the enrolment result table, from [example_enrol()].
+#' @param recist the recist table with progression date, from [example_rc()].
+#' @param seed Integer. Random seed for reproducibility (can be `NULL`).
+#' @param censor_min_days minimum time in days between last RECIST evaluation and administrative censoring
+#' @param censor_max_days maximum time in days between last RECIST evaluation and administrative censoring
+#' @param ... Additional arguments (currently unused).
+#'
+#' @return A tibble with `N` rows and the following columns:
+#'
+#' - `subjid`: Subject ID
+#' - `fu_status`: Status of the patient (0 - Alive/Censored , 1 - Dead)
+#' - `fu_date`: Date of latest news (death date if dead, otherwise censoring date)
+#'
+#' @keywords internal
+#' @importFrom dplyr arrange select mutate left_join if_else n
+#' @importFrom lubridate days
+#' @importFrom stats rexp
+#' @importFrom forcats as_factor fct_recode
+example_fu = function(enrolres, recist, seed, censor_min_days = 30, censor_max_days = 180, ...){
+
+  stopifnot(is.data.frame(enrolres), is.data.frame(recist))
+  if (!all(c("subjid", "date_inclusion") %in% names(enrolres))) {
+    stop("`enrolres` must contain columns `subjid` and `date_inclusion`.")
+  }
+  if (!all(c("subjid", "rcdt") %in% names(recist))) {
+    stop("`recist` must contain columns `subjid` and `rcdt` (date of RECIST evaluation).")
+  }
+  if(!("arm" %in% names(enrolres))) {
+    stop("`enrolres` must contain `arm`.")
+  }
+  if(!is.null(seed)) { set.seed(seed) }
+
+
+  enrolres = enrolres %>% arrange(subjid)
+  recist = recist %>% arrange(subjid)
+
+  last_recist = .date_last_visit(recist) #TODO considerer last_progressioon_date
+
+  # Survival rate assigned for each subject, according to arm of treatment
+  lambda = .surv_coef_trt(arm = enrolres$arm, lambda_control = 5) #TODO ne pas prendre des HR mais des beta
+
+  time_to_death = ceiling(rexp(n = nrow(enrolres),
+                               rate = lambda))
+
+
+  fu_data = enrolres %>%
+
+    select(subjid, date_inclusion) %>%
+    mutate(death_date = date_inclusion + days(time_to_death)) %>%
+
+    left_join(last_recist, by = "subjid") %>%
+    mutate(death_date = if_else(condition = !is.na(last_rcdt) & death_date <= last_rcdt,
+                                true = last_rcdt + days(ceiling(runif(n = n(), min = 7, max = 90))), #n() # ???
+                                false = death_date),
+
+           # Administrative censoring
+           censoring_days = round(runif(n(),
+                                        min = censor_min_days,
+                                        max = censor_max_days)),
+           censoring_date = last_rcdt + days(censoring_days),
+
+           # Status and date of latest news
+           fu_status = fct_recode(as_factor(as.integer(death_date <= censoring_date)),
+                                  "Alive/censored" = "0",
+                                  "Dead" = "1"),
+           fu_date = if_else(condition = death_date <= censoring_date,
+                             true = death_date,
+                             false = censoring_date)) %>%
+    select(subjid, fu_status, fu_date)
+
+  return(fu_data)
+}
+
+
+
+
+# Internals FU -----------------------------------------------------------------
+
+#' Used in `example_fu()`
+#' Extract the last date of RECIST evaluation for each patient and return a tibble
+#' @param recist a tibble with RECIST evaluation dates, with variables `subjid` and `rcdt` (Date)
+#' @return tibble with columns `subjid`, `last_rcdt` (last RECIST evaluation for each `subjid`)
+#' @noRd
+#' @keywords internal
+#' @importFrom dplyr select filter group_by arrange slice ungroup
+.date_last_visit = function(recist) {
+
+  stopifnot(is.data.frame(recist))
+  if (!all(c("subjid", "rcdt") %in% names(recist))) {
+    stop("`recist` must contain columns `subjid` and `rcdt` (RECIST evaluation dates).")
+  }
+
+  last_recist = recist %>%
+    select(subjid, rcdt) %>%
+    filter(!is.na(rcdt)) %>%
+    group_by(subjid) %>%
+    arrange(desc(rcdt), .by_group = TRUE) %>%
+    slice(1) %>%
+    ungroup() %>%
+    rename(last_rcdt = rcdt)
+}
+
+
+#' Used in `example_fu()`
+#' Assign a survival rate parameter (exponential lambda) for each arm of treatment
+#' @param arm factor vector of treatment arms (`Control` versus `Treatment`)
+#' @param lambda_control scale parameter of an exponential baseline hazard function
+#' @return numeric vector of rates (lambda) per subject, aligned with `arm`
+#' @noRd
+#' @keywords internal
+.surv_coef_trt = function(arm, lambda_control = 0.005) {
+
+  if(is.null(arm)) {stop("`arm` is NULL.")}
+  if(!is.factor(arm)) arm = as_factor(arm)
+
+  arm_levels = levels(arm)
+  n_arm = nlevels(arm)
+
+  if (!"Control" %in% arm_levels) { stop("One level of `arm` must be 'Control'. Found: ", paste(arm_levels, collapse = ", ")) }
+
+
+  if(n_arm == 2) { # "Control" versus "Treatment"
+
+    if (!"Treatment" %in% arm_levels) { stop("One level of `arm` must be 'Treatment'. Found: ", paste(arm_levels, collapse = ", ")) }
+    HR = c("Control" = 1, "Treatment" = 0.7)
+  } else { stop("Wrong number of arms of treatment: ", n_arm, ". Only two arms of treatment are supported.") }
+
+  lambda = as.numeric(lambda_control*HR[arm])
+
+  lambda
+}
+
+
 # Internals RC ------------------------------------------------------------
 
 #' Used in `example_rc()`

From 96d414585cb9d92a5c5e99d56e666c89522efb17 Mon Sep 17 00:00:00 2001
From: charlottebargain <charlotte.bargain@gmail.com>
Date: Thu, 13 Nov 2025 19:30:54 +0100
Subject: [PATCH 2/5] Ajout table `fu` dans `grstat_example()`
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

La table fu possède trois colonnes : `subjid` (l'identifiant patient), `fu_date` (date de dernier statut connu) et `fu_status` (dernier statut connu). Le statut est déterminé en fonction du bras de traitement (`arm` dans la table `enrolres` et du statut de progression de la tumeur connu dans la table `rc`.

Exemple :

db = grstat_example()

data_km = db$enrolres %>%
  left_join(db$fu, by = "subjid") %>%
  mutate(time_OS = as.numeric(fu_date - date_inclusion)/365.25,
         event_OS = 1*(fu_status == "Dead"))

km = survival::survfit(survival::Surv(time_OS, event_OS) ~ arm, data = data_km)
p = ggsurvfit::ggsurvfit(km) +
  ggsurvfit::add_censor_mark() +
  ggsurvfit::add_confidence_interval() +
  ggsurvfit::add_risktable()

print(p)
---
 NAMESPACE                             |   5 +
 R/data.R                              | 128 ++++++++++++++++----------
 grstat.Rproj                          |   1 -
 man/example_fu.Rd                     |  43 +++++++++
 tests/testthat/helper-zzz_chaRlotte.R |  14 +++
 5 files changed, 143 insertions(+), 48 deletions(-)
 create mode 100644 man/example_fu.Rd
 create mode 100644 tests/testthat/helper-zzz_chaRlotte.R

diff --git a/NAMESPACE b/NAMESPACE
index bbc1826..fba45ad 100644
--- a/NAMESPACE
+++ b/NAMESPACE
@@ -99,6 +99,7 @@ importFrom(flextable,width)
 importFrom(forcats,as_factor)
 importFrom(forcats,fct_drop)
 importFrom(forcats,fct_infreq)
+importFrom(forcats,fct_recode)
 importFrom(forcats,fct_relevel)
 importFrom(forcats,fct_reorder)
 importFrom(forcats,fct_rev)
@@ -146,6 +147,9 @@ importFrom(ggplot2,unit)
 importFrom(ggplot2,vars)
 importFrom(glue,glue)
 importFrom(lifecycle,deprecated)
+importFrom(lubridate,as_date)
+importFrom(lubridate,days)
+importFrom(lubridate,ymd)
 importFrom(purrr,discard)
 importFrom(purrr,imap)
 importFrom(purrr,iwalk)
@@ -181,6 +185,7 @@ importFrom(scales,breaks_width)
 importFrom(scales,label_percent)
 importFrom(stats,na.omit)
 importFrom(stats,rbinom)
+importFrom(stats,rexp)
 importFrom(stats,rnorm)
 importFrom(stats,runif)
 importFrom(stringr,fixed)
diff --git a/R/data.R b/R/data.R
index b556b63..a585980 100644
--- a/R/data.R
+++ b/R/data.R
@@ -29,6 +29,8 @@ grstat_example = function(N=200, seed=42, ...){
   rtn$date_extraction = "2024/01/01"
   rtn$datetime_extraction = structure(1704067200, class = c("POSIXct", "POSIXt"),
                                       tzone = "Europe/Paris")
+
+  rtn = .remove_post_event(rtn)
   rtn
 }
 
@@ -287,6 +289,11 @@ example_rc = function(enrolres, seed,
 
 
 
+
+
+# Calculer le temps de décès selon bras et statut progression. Puis en dernier, tronquer les visites recist et le suivi ae a la date de deces/censure
+
+
 #' Generate Simulated Survival Data
 #'
 #' Internal function that simulates survival data, which depend on treatment arm
@@ -307,17 +314,17 @@ example_rc = function(enrolres, seed,
 #'
 #' @keywords internal
 #' @importFrom dplyr arrange select mutate left_join if_else n
-#' @importFrom lubridate days
+#' @importFrom lubridate ymd as_date days
 #' @importFrom stats rexp
 #' @importFrom forcats as_factor fct_recode
-example_fu = function(enrolres, recist, seed, censor_min_days = 30, censor_max_days = 180, ...){
+example_fu = function(enrolres, recist, seed, lambda_censor = 0.3, lambda_control = 0.2, beta_arm = -0.6, beta_prog_status = 0.5, ...){
 
   stopifnot(is.data.frame(enrolres), is.data.frame(recist))
   if (!all(c("subjid", "date_inclusion") %in% names(enrolres))) {
     stop("`enrolres` must contain columns `subjid` and `date_inclusion`.")
   }
-  if (!all(c("subjid", "rcdt") %in% names(recist))) {
-    stop("`recist` must contain columns `subjid` and `rcdt` (date of RECIST evaluation).")
+  if (!all(c("subjid", "rcdt", "rcresp") %in% names(recist))) {
+    stop("`recist` must contain columns `subjid`, `rcdt` (date of RECIST evaluation) and `rcresp` (RECIST global response).")
   }
   if(!("arm" %in% names(enrolres))) {
     stop("`enrolres` must contain `arm`.")
@@ -328,30 +335,24 @@ example_fu = function(enrolres, recist, seed, censor_min_days = 30, censor_max_d
   enrolres = enrolres %>% arrange(subjid)
   recist = recist %>% arrange(subjid)
 
-  last_recist = .date_last_visit(recist) #TODO considerer last_progressioon_date
-
-  # Survival rate assigned for each subject, according to arm of treatment
-  lambda = .surv_coef_trt(arm = enrolres$arm, lambda_control = 5) #TODO ne pas prendre des HR mais des beta
+  prog_status = .progression_status(recist)
 
-  time_to_death = ceiling(rexp(n = nrow(enrolres),
-                               rate = lambda))
+  # Survival rate assigned for each subject, according to arm of treatment and progression status
+  lambda = .surv_coef(arm = enrolres$arm, prog_status = prog_status$status, lambda_control = lambda_control,
+                      beta_arm = beta_arm, beta_prog_status = beta_prog_status)
 
+  time_to_death = rexp(n = nrow(enrolres), rate = lambda)
+  time_to_censor = rexp(n = nrow(enrolres), rate = lambda_censor)
 
   fu_data = enrolres %>%
 
     select(subjid, date_inclusion) %>%
-    mutate(death_date = date_inclusion + days(time_to_death)) %>%
+    left_join(prog_status, by = join_by(subjid)) %>%
 
-    left_join(last_recist, by = "subjid") %>%
-    mutate(death_date = if_else(condition = !is.na(last_rcdt) & death_date <= last_rcdt,
-                                true = last_rcdt + days(ceiling(runif(n = n(), min = 7, max = 90))), #n() # ???
-                                false = death_date),
+    mutate(date_inclusion = as.Date(date_inclusion),
 
-           # Administrative censoring
-           censoring_days = round(runif(n(),
-                                        min = censor_min_days,
-                                        max = censor_max_days)),
-           censoring_date = last_rcdt + days(censoring_days),
+           death_date = date_inclusion + days(ceiling(time_to_death*365.25)),
+           censoring_date = date_inclusion + days(ceiling(time_to_censor*365.25)),
 
            # Status and date of latest news
            fu_status = fct_recode(as_factor(as.integer(death_date <= censoring_date)),
@@ -360,71 +361,104 @@ example_fu = function(enrolres, recist, seed, censor_min_days = 30, censor_max_d
            fu_date = if_else(condition = death_date <= censoring_date,
                              true = death_date,
                              false = censoring_date)) %>%
-    select(subjid, fu_status, fu_date)
+    select(subjid, fu_status, fu_date) %>%
+    apply_labels(subjid = "Subject ID",
+                 fu_status = "Status",
+                 fu_date = "Date of last known status")
 
-  return(fu_data)
+    return(fu_data)
 }
 
 
 
-
 # Internals FU -----------------------------------------------------------------
 
+
 #' Used in `example_fu()`
-#' Extract the last date of RECIST evaluation for each patient and return a tibble
-#' @param recist a tibble with RECIST evaluation dates, with variables `subjid` and `rcdt` (Date)
-#' @return tibble with columns `subjid`, `last_rcdt` (last RECIST evaluation for each `subjid`)
+#' Extract the progression status of RECIST evaluation for each patient and return a tibble
+#' @param recist a tibble with RECIST progression status for each patients
+#' @return tibble with columns `subjid`, `prog_status` (binary variable : progressive disease or not)
 #' @noRd
 #' @keywords internal
-#' @importFrom dplyr select filter group_by arrange slice ungroup
-.date_last_visit = function(recist) {
+#' @importFrom dplyr select summarise
+.progression_status = function(recist) {
 
   stopifnot(is.data.frame(recist))
-  if (!all(c("subjid", "rcdt") %in% names(recist))) {
-    stop("`recist` must contain columns `subjid` and `rcdt` (RECIST evaluation dates).")
+  if (!all(c("subjid", "rcresp") %in% names(recist))) {
+    stop("`recist` must contain columns `subjid`, and `rcresp` (RECIST global response).")
   }
 
-  last_recist = recist %>%
-    select(subjid, rcdt) %>%
-    filter(!is.na(rcdt)) %>%
-    group_by(subjid) %>%
-    arrange(desc(rcdt), .by_group = TRUE) %>%
-    slice(1) %>%
-    ungroup() %>%
-    rename(last_rcdt = rcdt)
+  rcresp_levels = levels(recist$rcresp)
+  if (!"Progressive disease" %in% rcresp_levels) { stop("One level of `rcresp_levels` must be 'Progressive disease'. Found: ", paste(rcresp_levels, collapse = ", ")) }
+
+  progression_status = recist %>%
+    select(subjid, rcresp) %>%
+    summarise(status = any(rcresp == "Progressive disease"), .by = 'subjid') %>%
+    mutate(status = as_factor(if_else(!is.na(status), true = "Progressive disease", false = "No progression")))
 }
 
 
 #' Used in `example_fu()`
-#' Assign a survival rate parameter (exponential lambda) for each arm of treatment
+#' Assign a survival rate parameter (exponential lambda) for each patient, depending on treatment and progression status
 #' @param arm factor vector of treatment arms (`Control` versus `Treatment`)
+#' @param prog_status factor vector of progression status (`No progressive disease` versus `progressive disease`)
 #' @param lambda_control scale parameter of an exponential baseline hazard function
-#' @return numeric vector of rates (lambda) per subject, aligned with `arm`
+#' @return numeric vector of rates (lambda) per subject, aligned with `arm` and `prog_status`
 #' @noRd
 #' @keywords internal
-.surv_coef_trt = function(arm, lambda_control = 0.005) {
+.surv_coef = function(arm, prog_status, lambda_control, beta_arm, beta_prog_status) {
 
   if(is.null(arm)) {stop("`arm` is NULL.")}
+  if(is.null(prog_status)) {stop("`prog_status` is NULL.")}
+
   if(!is.factor(arm)) arm = as_factor(arm)
+  if(!is.factor(prog_status)) prog_status = mutate(prog_status, status = as_factor(status))
 
   arm_levels = levels(arm)
   n_arm = nlevels(arm)
-
   if (!"Control" %in% arm_levels) { stop("One level of `arm` must be 'Control'. Found: ", paste(arm_levels, collapse = ", ")) }
 
+  prog_status_levels = levels(prog_status)
+  n_prog_status = nlevels(prog_status)
+  if (!"Progressive disease" %in% prog_status_levels) { stop("One level of `prog_status` must be 'Progressive disease'. Found: ", paste(prog_status_levels, collapse = ", ")) }
+
 
-  if(n_arm == 2) { # "Control" versus "Treatment"
+  if(n_arm == 2 & n_prog_status == 2) { # "Control" versus "Treatment" and "Progressive disease" versus "No Progressive disease (stable or response)"
 
     if (!"Treatment" %in% arm_levels) { stop("One level of `arm` must be 'Treatment'. Found: ", paste(arm_levels, collapse = ", ")) }
-    HR = c("Control" = 1, "Treatment" = 0.7)
-  } else { stop("Wrong number of arms of treatment: ", n_arm, ". Only two arms of treatment are supported.") }
 
-  lambda = as.numeric(lambda_control*HR[arm])
+    lambda = lambda_control * exp(beta_arm*(arm == "Treatment") + beta_prog_status*(prog_status == "Progressive disease"))
+
+  } else if(n_arm != 2) { stop("Wrong number of arms of treatment: ", n_arm, ". Only two arms of treatment are supported.")
+  } else if(n_prog_status != 2) { stop("Wrong number of progression status: ", n_prog_status, ". Progression status must be a binary variable.")}
 
   lambda
 }
 
 
+#' Used in `grstat_example()`
+#' Remove recist evaluation rows posterior to the end of the follow-up
+#' @param rtn list of data frames with enrolment, adverse event, recist evaluation and follow-up information
+#' @noRd
+#' @return list of data frames with enrolment, adverse event, recist evaluation and follow-up information with recist evaluation truncated with last follow-up date
+#' @keywords internal
+#' @importFrom dplyr left_join select filter
+.remove_post_event = function(rtn){
+
+  if(!is.list(rtn)) {stop("`rtn` must be a list of data frames.")}
+
+  if (! ("recist" %in% names(rtn)) ) { stop("One data frame in `rtn` list must be 'recist'. Found: ", names(rtn, collapse = ", ")) }
+  if (! ("fu" %in% names(rtn)) ) { stop("One data frame in `rtn` list must be 'fu'. Found: ", names(rtn, collapse = ", ")) }
+
+  rtn$recist <- rtn$recist %>%
+    left_join(rtn$fu %>% select(- crfname), by = join_by(subjid)) %>%
+    filter(rcdt <= fu_date) %>%
+    select(- c(fu_date, fu_status))
+
+  rtn
+}
+
+
 # Internals RC ------------------------------------------------------------
 
 #' Used in `example_rc()`
@@ -442,7 +476,7 @@ example_fu = function(enrolres, recist, seed, censor_min_days = 30, censor_max_d
                    percent_change_per_month > 0 ~ 1 / rc_coef_treatement,
                    .default = rc_coef_treatement)
   percent_change_per_month = percent_change_per_month * coef
-  subj_delai = 42 * seq(rc_num_timepoints) + runif(rc_num_timepoints, -7, 7)
+  subj_delai = 15 * seq(rc_num_timepoints) + runif(rc_num_timepoints, -7, 7)
   percent_change = percent_change_per_month * subj_delai / 30.5
   percent_change = percent_change + rnorm(rc_num_timepoints, 0, rc_sd_tlsum_noise)
   percent_change = c(0, percent_change)
diff --git a/grstat.Rproj b/grstat.Rproj
index 028250a..6515568 100644
--- a/grstat.Rproj
+++ b/grstat.Rproj
@@ -1,5 +1,4 @@
 Version: 1.0
-ProjectId: d350bac8-4795-42a0-a32d-ccb93a467950
 
 RestoreWorkspace: No
 SaveWorkspace: No
diff --git a/man/example_fu.Rd b/man/example_fu.Rd
new file mode 100644
index 0000000..5d225ef
--- /dev/null
+++ b/man/example_fu.Rd
@@ -0,0 +1,43 @@
+% Generated by roxygen2: do not edit by hand
+% Please edit documentation in R/data.R
+\name{example_fu}
+\alias{example_fu}
+\title{Generate Simulated Survival Data}
+\usage{
+example_fu(
+  enrolres,
+  recist,
+  seed,
+  lambda_censor = 0.3,
+  lambda_control = 0.2,
+  beta_arm = -0.6,
+  beta_prog_status = 0.5,
+  ...
+)
+}
+\arguments{
+\item{enrolres}{the enrolment result table, from \code{\link[=example_enrol]{example_enrol()}}.}
+
+\item{recist}{the recist table with progression date, from \code{\link[=example_rc]{example_rc()}}.}
+
+\item{seed}{Integer. Random seed for reproducibility (can be \code{NULL}).}
+
+\item{...}{Additional arguments (currently unused).}
+
+\item{censor_min_days}{minimum time in days between last RECIST evaluation and administrative censoring}
+
+\item{censor_max_days}{maximum time in days between last RECIST evaluation and administrative censoring}
+}
+\value{
+A tibble with \code{N} rows and the following columns:
+\itemize{
+\item \code{subjid}: Subject ID
+\item \code{fu_status}: Status of the patient (0 - Alive/Censored , 1 - Dead)
+\item \code{fu_date}: Date of latest news (death date if dead, otherwise censoring date)
+}
+}
+\description{
+Internal function that simulates survival data, which depend on treatment arm
+and date of progression if applicable.
+}
+\keyword{internal}
diff --git a/tests/testthat/helper-zzz_chaRlotte.R b/tests/testthat/helper-zzz_chaRlotte.R
new file mode 100644
index 0000000..40debbb
--- /dev/null
+++ b/tests/testthat/helper-zzz_chaRlotte.R
@@ -0,0 +1,14 @@
+db = grstat_example()
+
+data_km = db$enrolres %>%
+  left_join(db$fu, by = "subjid") %>%
+  mutate(time_OS = as.numeric(fu_date - date_inclusion)/365.25,
+         event_OS = 1*(fu_status == "Dead"))
+
+km = survival::survfit(survival::Surv(time_OS, event_OS) ~ arm, data = data_km)
+p = ggsurvfit::ggsurvfit(km) +
+  ggsurvfit::add_censor_mark() +
+  ggsurvfit::add_confidence_interval() +
+  ggsurvfit::add_risktable()
+
+print(p)

From c7c91d234091e539e68e6bf5c062404b16b1e770 Mon Sep 17 00:00:00 2001
From: Dan Chaltiel <15105152+DanChaltiel@users.noreply.github.com>
Date: Mon, 24 Nov 2025 10:32:04 +0100
Subject: [PATCH 3/5] c'est les vacances, faisons du shiny!
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Comme promis, l'appli pour tweaker tes paramètres en temps réel.
- Le prompt IA, si ça t'intéresse de voir comment j'ai fait : https://chatgpt.com/share/69242598-61fc-8003-b30f-72e50d93da0a
- J'ai mis des données bidon pour la PFS, je te laisse cuisiner get_plot()
- rc_delay n'est pas fonctionnel tant qu'il n'existe pas dans `example_rc()` (cf commentaire dans ma PR), mais on en aura sans doute besoin pour avoir une PFS jolie. Tu peux évidemment le renommer, je n'y ai pas beaucoup réfléchi.
- j'ai un tout petit peu modifié `grstat_example()` pour que l'appel soit plus robuste car j'avais une erreur incompréhensible.

[skip_ci]
---
 R/data.R                               |   8 +-
 tests/testthat/helper-zzz_shiny_data.R | 232 +++++++++++++++++++++++++
 2 files changed, 236 insertions(+), 4 deletions(-)
 create mode 100644 tests/testthat/helper-zzz_shiny_data.R

diff --git a/R/data.R b/R/data.R
index a585980..f2596c7 100644
--- a/R/data.R
+++ b/R/data.R
@@ -16,13 +16,13 @@
 #' @importFrom tibble lst
 grstat_example = function(N=200, seed=42, ...){
 
-  enrolres = example_enrol(N, seed, ...)
+  enrolres = example_enrol(N=N, seed=seed, ...)
 
-  ae = example_ae(enrolres, seed, ...)
+  ae = example_ae(enrolres=enrolres, seed=seed, ...)
 
-  recist = example_rc(enrolres, seed, ...)
+  recist = example_rc(enrolres=enrolres, seed=seed, ...)
 
-  fu = example_fu(enrolres, recist, seed, ...)
+  fu = example_fu(enrolres=enrolres, recist=recist, seed=seed, ...)
 
   rtn = lst(enrolres, ae, recist, fu) %>%
     imap(~.x %>% mutate(crfname=.y %>% set_label("Form name")))
diff --git a/tests/testthat/helper-zzz_shiny_data.R b/tests/testthat/helper-zzz_shiny_data.R
new file mode 100644
index 0000000..7eb4aa7
--- /dev/null
+++ b/tests/testthat/helper-zzz_shiny_data.R
@@ -0,0 +1,232 @@
+library(shiny)
+
+grstat_example_ui = function() {
+  fluidPage(
+    titlePanel("Paramètres de grstat_example"),
+
+    fluidRow(
+      column(
+        width = 4,
+        h4("Paramètres d'entrée"),
+        actionButton("run_sim", "Lancer la simulation"),
+        br(), br(),
+        tabsetPanel(
+          id = "input_tabs",
+
+          tabPanel(
+            "Plot",
+            fluidRow(
+              column(
+                12,
+                numericInput("N", "Nombre de patients (N)", value = 200, min = 1, step = 1),
+                numericInput("seed", "Graine aléatoire (seed)", value = 42, step = 1)
+              )
+            )
+          ),
+
+          tabPanel(
+            "Enrol",
+            fluidRow(
+              column(
+                12,
+                sliderInput("r", "Proportion contrôle pour arm (r)", min = 0, max = 1, value = 0.5, step = 0.01),
+                sliderInput("r2", "Proportion contrôle pour arm3 (r2)", min = 0, max = 1, value = 1/3, step = 0.01)
+              )
+            )
+          ),
+
+          tabPanel(
+            "AE",
+            fluidRow(
+              column(
+                6,
+                sliderInput("p_na", "Proportion de NA (p_na)", min = 0, max = 1, value = 0, step = 0.01),
+                sliderInput("p_sae", "Proportion SAE contrôle (p_sae)", min = 0, max = 1, value = 0.1, step = 0.01),
+                sliderInput("p_sae_trt", "Proportion SAE traitement (p_sae_trt)", min = 0, max = 1, value = 0.1, step = 0.01),
+                numericInput("n_max", "AE max par patient contrôle (n_max)", value = 15, min = 1, step = 1),
+                numericInput("n_max_trt", "AE max par patient traité (n_max_trt)", value = 15, min = 1, step = 1),
+              ),
+              column(
+                6,
+                numericInput("w_soc", "Poids SOC contrôle (w_soc)", value = 1, step = 0.1),
+                numericInput("w_soc_trt", "Poids SOC traitement (w_soc_trt)", value = 1, step = 0.1),
+                numericInput("beta0", "Intercept grade (beta0)", value = -1, step = 0.1),
+                numericInput("beta_trt", "Effet traitement (beta_trt)", value = 0.4, step = 0.1),
+                numericInput("beta_sae", "Effet SAE (beta_sae)", value = 1, step = 0.1)
+              )
+            )
+          ),
+
+          tabPanel(
+            "RECIST",
+            fluidRow(
+              column(
+                6,
+                numericInput("rc_delay", "Délai entre 2 évaluations (TODO rc_delay)", value = 15, step = 1),
+                numericInput("rc_num_timepoints", "Nombre de visites (rc_num_timepoints)", value = 5, min = 2, step = 1),
+                sliderInput("rc_p_new_lesions", "Proba nouvelles lésions (rc_p_new_lesions)", min = 0, max = 1, value = 0.09, step = 0.01),
+                sliderInput("rc_p_na", "Proba NA (rc_p_na)", min = 0, max = 1, value = 0.005, step = 0.001),
+                sliderInput("rc_p_nt_lesions_yn", "Présence NT (rc_p_nt_lesions_yn)", min = 0, max = 1, value = 0.5, step = 0.01),
+                numericInput("rc_sd_tlsum_noise", "SD bruit TLSUM (rc_sd_tlsum_noise)", value = 0.5, step = 0.1),
+                numericInput("rc_coef_treatement", "Coefficient traitement (rc_coef_treatement)", value = 3, step = 0.1),
+              ),
+              column(
+                6,
+                h4("Probabilités réponse NT (rc_p_nt_lesions_resp)"),
+                sliderInput("rc_p_nt_CR", "CR (rc_p_nt_CR)", min = 0, max = 1, value = 0.73, step = 0.01),
+                sliderInput("rc_p_nt_SD", "SD (rc_p_nt_SD)", min = 0, max = 1, value = 0.25, step = 0.01),
+                sliderInput("rc_p_nt_PD", "PD (rc_p_nt_PD)", min = 0, max = 1, value = 0.01, step = 0.01),
+                sliderInput("rc_p_nt_NE", "NE (rc_p_nt_NE)", min = 0, max = 1, value = 0.01, step = 0.01)
+              )
+            )
+          ),
+
+          tabPanel(
+            "Follow-up",
+            fluidRow(
+              column(
+                12,
+                numericInput("lambda_censor", "Lambda censure (lambda_censor)", value = 0.3, step = 0.01),
+                numericInput("lambda_control", "Lambda contrôle (lambda_control)", value = 0.2, step = 0.01),
+                numericInput("beta_arm", "Effet bras (beta_arm)", value = -0.6, step = 0.1),
+                numericInput("beta_prog_status", "Effet statut progression (beta_prog_status)", value = 0.5, step = 0.1)
+              )
+            )
+          )
+        )
+      ),
+
+      column(
+        width = 8,
+        tabsetPanel(
+          id = "output_tabs",
+          tabPanel("Plot", plotOutput("sim_plot")),
+          tabPanel("Enrol", tableOutput("enrol_table")),
+          tabPanel("AE", tableOutput("ae_table")),
+          tabPanel("RECIST", tableOutput("recist_table")),
+          tabPanel("Follow-up", tableOutput("fu_table"))
+        )
+      )
+    )
+  )
+}
+
+
+
+
+grstat_example_server = function(input, output, session) {
+
+  sim = eventReactive(input$run_sim, {
+    grstat_example(
+      N = input$N,
+      seed = input$seed,
+
+      r = input$r,
+      r2 = input$r2,
+
+      p_na = input$p_na,
+      p_sae = input$p_sae,
+      p_sae_trt = input$p_sae_trt,
+      n_max = input$n_max,
+      n_max_trt = input$n_max_trt,
+      w_soc = input$w_soc,
+      w_soc_trt = input$w_soc_trt,
+      beta0 = input$beta0,
+      beta_trt = input$beta_trt,
+      beta_sae = input$beta_sae,
+
+      rc_delay = input$rc_delay,
+      rc_num_timepoints = input$rc_num_timepoints,
+      rc_p_new_lesions = input$rc_p_new_lesions,
+      rc_p_na = input$rc_p_na,
+      rc_p_nt_lesions_yn = input$rc_p_nt_lesions_yn,
+      rc_p_nt_lesions_resp = list(
+        CR = input$rc_p_nt_CR,
+        SD = input$rc_p_nt_SD,
+        PD = input$rc_p_nt_PD,
+        NE = input$rc_p_nt_NE
+      ),
+      rc_sd_tlsum_noise = input$rc_sd_tlsum_noise,
+      rc_coef_treatement = input$rc_coef_treatement,
+
+      lambda_censor = input$lambda_censor,
+      lambda_control = input$lambda_control,
+      beta_arm = input$beta_arm,
+      beta_prog_status = input$beta_prog_status
+    )
+  })
+
+  output$enrol_table = renderTable({
+    x = sim()
+    if (is.null(x)) return(NULL)
+    head(x$enrolres, 20)
+  })
+
+  output$ae_table = renderTable({
+    x = sim()
+    if (is.null(x)) return(NULL)
+    head(x$ae, 20)
+  })
+
+  output$recist_table = renderTable({
+    x = sim()
+    if (is.null(x)) return(NULL)
+    head(x$recist, 20)
+  })
+
+  output$fu_table = renderTable({
+    x = sim()
+    if (is.null(x)) return(NULL)
+    head(x$fu, 20)
+  })
+
+  output$sim_plot = renderPlot({
+    x = sim()
+    if (is.null(x)) return(NULL)
+    get_plot(x)
+  })
+}
+
+
+# Make the plot -------------------------------------------------------------------------------
+
+
+get_plot = function(db){
+  set.seed(42)
+  data_km = db$enrolres %>%
+    left_join(db$fu, by = "subjid") %>%
+    mutate(
+      time_OS = as.numeric(fu_date - date_inclusion)/365.25,
+      event_OS = fu_status == "Dead",
+      time_PFS = time_OS - rnorm(n(), 1, 0.5),
+      event_PFS = fu_status == "Dead" | rbinom(n(), 1, 0.1)
+    )
+
+  km_os = survival::survfit(survival::Surv(time_OS, event_OS) ~ arm, data = data_km)
+  p_os = ggsurvfit::ggsurvfit(km_os) +
+    ggsurvfit::add_censor_mark() +
+    ggsurvfit::add_confidence_interval() +
+    ggsurvfit::add_risktable() +
+    labs(title="OS")
+
+  #TODO : faire un KM avec la PFS RECIST
+  km_pfs = survival::survfit(survival::Surv(time_PFS, event_PFS) ~ arm, data = data_km)
+  p_pfs = ggsurvfit::ggsurvfit(km_pfs) +
+    ggsurvfit::add_censor_mark() +
+    ggsurvfit::add_confidence_interval() +
+    ggsurvfit::add_risktable() +
+    labs(title="PFS (données totalement aléatoires)")
+
+  patchwork::wrap_plots(p_os, p_pfs, ncol=1, guides="collect")
+}
+
+
+# Run the app ---------------------------------------------------------------------------------
+
+app = shinyApp(
+  ui = grstat_example_ui,
+  server = grstat_example_server
+)
+#FIXME décommente cette ligne pour que l'app soit lancée à chaque Ctrl+Shift+L
+#Commande pour lancer l'appli
+# runApp(app)

From e13fb121c8d501c0c01f0e092a695789f9fcc75d Mon Sep 17 00:00:00 2001
From: charlottebargain <charlotte.bargain@gmail.com>
Date: Thu, 11 Jun 2026 12:27:42 +0200
Subject: [PATCH 4/5] Add data set `fu` with OS and PFS status and delay to
 `grstat_example`

The new data set `fu` contains for each subject/patient simulated with `grstat_example` :
- Overall survival status and delay from inclusion
- Progression-free survival status and delay from inclusion

Survival outcomes depends on :
- Treatment arm
- RECIST progression dates

Censoring is explicitly simulated using an exponential distribution
controlled by the `lambda_censor` parameter.
---
 DESCRIPTION                |   2 +-
 R/data.R                   | 114 ++++++++++++++++++++++++++-----------
 man/ae_plot_grade.Rd       |   2 +-
 man/example_fu.Rd          |  20 ++++---
 man/example_rc.Rd          |   1 +
 man/grstat-package.Rd      |   7 ++-
 man/grstat_example.Rd      |   2 +-
 man/reexports.Rd           |   6 +-
 tests/testthat/test-data.R |  13 +++++
 9 files changed, 121 insertions(+), 46 deletions(-)

diff --git a/DESCRIPTION b/DESCRIPTION
index 2bebbca..21956c5 100644
--- a/DESCRIPTION
+++ b/DESCRIPTION
@@ -64,7 +64,7 @@ Suggests:
     vdiffr
 Encoding: UTF-8
 Roxygen: list(markdown = TRUE)
-RoxygenNote: 7.3.2
 Config/testthat/edition: 3
 Config/testthat/parallel: true
 VignetteBuilder: knitr
+Config/roxygen2/version: 8.0.0
diff --git a/R/data.R b/R/data.R
index f2596c7..15217cb 100644
--- a/R/data.R
+++ b/R/data.R
@@ -6,7 +6,7 @@
 #'
 #' @param N the number of patients
 #' @param seed the random seed (can be `NULL`)
-#' @param ... passed on to internal functions. See [example_enrol()],  [example_ae()], and  [example_rc()] for the argument names.
+#' @param ... passed on to internal functions. See [example_enrol()],  [example_ae()], [example_rc()], and [example_fu()] for the argument names.
 #'
 #' @returns A list of datasets, like in EDCimport.
 #'
@@ -30,7 +30,7 @@ grstat_example = function(N=200, seed=42, ...){
   rtn$datetime_extraction = structure(1704067200, class = c("POSIXct", "POSIXt"),
                                       tzone = "Europe/Paris")
 
-  rtn = .remove_post_event(rtn)
+  # rtn = .remove_post_event(rtn)
   rtn
 }
 
@@ -202,12 +202,14 @@ example_rc = function(enrolres, seed,
                       rc_p_nt_lesions_resp = list("CR"=0.73, "SD"=0.25, "PD"=0.01, "NE"=0.01),
                       rc_sd_tlsum_noise = 0.5,
                       rc_coef_treatement = 3,
+                      rc_subj_delai = 42,
                       ...) {
   set.seed(seed)
   recist_data = enrolres %>%
     mutate(
       data = .simulate_one_patient(arm, rc_num_timepoints, rc_p_na,
-                                   rc_sd_tlsum_noise, rc_coef_treatement),
+                                   rc_sd_tlsum_noise, rc_coef_treatement,
+                                   rc_subj_delai),
       .by = subjid
     ) %>%
     unnest(data) %>%
@@ -291,7 +293,7 @@ example_rc = function(enrolres, seed,
 
 
 
-# Calculer le temps de décès selon bras et statut progression. Puis en dernier, tronquer les visites recist et le suivi ae a la date de deces/censure
+# Calculer le temps de décès selon bras et statut progression.
 
 
 #' Generate Simulated Survival Data
@@ -302,22 +304,28 @@ example_rc = function(enrolres, seed,
 #' @param enrolres the enrolment result table, from [example_enrol()].
 #' @param recist the recist table with progression date, from [example_rc()].
 #' @param seed Integer. Random seed for reproducibility (can be `NULL`).
-#' @param censor_min_days minimum time in days between last RECIST evaluation and administrative censoring
-#' @param censor_max_days maximum time in days between last RECIST evaluation and administrative censoring
+#' @param lambda_censor numeric. Rate parameter of an exponential distribution used to simulate censoring times.
+#' @param lambda_control numeric. Scale parameter of an exponential baseline hazard function.
+#' @param beta_arm numeric. Parameter associated to the treatment effect.
+#' @param beta_prog_status numeric. Parameter associated to a known progression on a RECIST assessment.
 #' @param ... Additional arguments (currently unused).
 #'
 #' @return A tibble with `N` rows and the following columns:
 #'
 #' - `subjid`: Subject ID
-#' - `fu_status`: Status of the patient (0 - Alive/Censored , 1 - Dead)
-#' - `fu_date`: Date of latest news (death date if dead, otherwise censoring date)
+#' - `pfs_status`: Progression-free survival status of the patient
+#' - `pfs_date`: Date of latest news before progression/death
+#' - `os_status`: Overall survival status of the patient (0 - Alive/Censored , 1 - Dead)
+#' - `os_date`: Date of latest news (death date if dead, otherwise censoring date)
 #'
 #' @keywords internal
 #' @importFrom dplyr arrange select mutate left_join if_else n
 #' @importFrom lubridate ymd as_date days
 #' @importFrom stats rexp
 #' @importFrom forcats as_factor fct_recode
-example_fu = function(enrolres, recist, seed, lambda_censor = 0.3, lambda_control = 0.2, beta_arm = -0.6, beta_prog_status = 0.5, ...){
+example_fu = function(enrolres, recist, seed,
+                      lambda_censor = 0.2, lambda_control = 0.2,
+                      beta_arm = -0.6, beta_prog_status = 2, ...){
 
   stopifnot(is.data.frame(enrolres), is.data.frame(recist))
   if (!all(c("subjid", "date_inclusion") %in% names(enrolres))) {
@@ -331,15 +339,16 @@ example_fu = function(enrolres, recist, seed, lambda_censor = 0.3, lambda_contro
   }
   if(!is.null(seed)) { set.seed(seed) }
 
-
   enrolres = enrolres %>% arrange(subjid)
   recist = recist %>% arrange(subjid)
 
   prog_status = .progression_status(recist)
+  progression_date = .progression_date(recist)
 
   # Survival rate assigned for each subject, according to arm of treatment and progression status
-  lambda = .surv_coef(arm = enrolres$arm, prog_status = prog_status$status, lambda_control = lambda_control,
-                      beta_arm = beta_arm, beta_prog_status = beta_prog_status)
+  lambda = .surv_coef(arm = enrolres$arm, prog_status = prog_status$status,
+                      lambda_control = lambda_control, beta_arm = beta_arm,
+                      beta_prog_status = beta_prog_status)
 
   time_to_death = rexp(n = nrow(enrolres), rate = lambda)
   time_to_censor = rexp(n = nrow(enrolres), rate = lambda_censor)
@@ -348,25 +357,40 @@ example_fu = function(enrolres, recist, seed, lambda_censor = 0.3, lambda_contro
 
     select(subjid, date_inclusion) %>%
     left_join(prog_status, by = join_by(subjid)) %>%
+    left_join(progression_date, by = join_by(subjid)) %>%
+
+    mutate(
+      date_inclusion = as.Date(date_inclusion),
 
-    mutate(date_inclusion = as.Date(date_inclusion),
+      origin_date = if_else(!is.na(prog_date), prog_date, date_inclusion),
 
-           death_date = date_inclusion + days(ceiling(time_to_death*365.25)),
-           censoring_date = date_inclusion + days(ceiling(time_to_censor*365.25)),
+      death_date = origin_date + days(ceiling(time_to_death*365.25)),
+      censoring_date = origin_date + days(ceiling(time_to_censor*365.25)),
 
-           # Status and date of latest news
-           fu_status = fct_recode(as_factor(as.integer(death_date <= censoring_date)),
-                                  "Alive/censored" = "0",
-                                  "Dead" = "1"),
-           fu_date = if_else(condition = death_date <= censoring_date,
-                             true = death_date,
-                             false = censoring_date)) %>%
-    select(subjid, fu_status, fu_date) %>%
+      # Status and date of progression
+      pfs_status = case_when(
+        !is.na(prog_date) & prog_date <= death_date & prog_date <= censoring_date ~ 1,
+        death_date <= censoring_date ~ 1,
+        .default = 0
+      ),
+      pfs_date = pmin(prog_date, death_date, censoring_date, na.rm = TRUE),
+
+      # Status and date of latest news
+      os_status = fct_recode(as_factor(as.integer(death_date <= censoring_date)),
+                             "Alive/censored" = "0",
+                             "Dead" = "1"),
+      os_date = if_else(condition = death_date <= censoring_date,
+                        true = death_date,
+                        false = censoring_date)
+    ) %>%
+    select(subjid, pfs_status, pfs_date, os_status, os_date) %>%
     apply_labels(subjid = "Subject ID",
-                 fu_status = "Status",
-                 fu_date = "Date of last known status")
+                 pfs_status = "Status PFS",
+                 pfs_date = "Date of progression",
+                 os_status = "Status OS",
+                 os_date = "Date of last known OS status")
 
-    return(fu_data)
+  return(fu_data)
 }
 
 
@@ -394,7 +418,29 @@ example_fu = function(enrolres, recist, seed, lambda_censor = 0.3, lambda_contro
   progression_status = recist %>%
     select(subjid, rcresp) %>%
     summarise(status = any(rcresp == "Progressive disease"), .by = 'subjid') %>%
-    mutate(status = as_factor(if_else(!is.na(status), true = "Progressive disease", false = "No progression")))
+    mutate(status = as_factor(if_else(!is.na(status),
+                                      true = "Progressive disease",
+                                      false = "No progression")))
+}
+
+
+#' Used in `example_fu()`
+#' Extract first progression date per patient
+#' @param recist a tibble with RECIST progression status for each patients
+#' @return tibble with columns `subjid`, `prog_date`
+#' @noRd
+#' @keywords internal
+#' @importFrom dplyr filter summarise
+.progression_date = function(recist) {
+
+  stopifnot(is.data.frame(recist))
+
+  progression_date = recist %>%
+    summarise(
+      prog_date = if(any(rcresp == "Progressive disease", na.rm = TRUE)) {
+        min(rcdt[rcresp == "Progressive disease"], na.rm = TRUE)
+      } else { NA },
+      .by = subjid)
 }
 
 
@@ -403,6 +449,8 @@ example_fu = function(enrolres, recist, seed, lambda_censor = 0.3, lambda_contro
 #' @param arm factor vector of treatment arms (`Control` versus `Treatment`)
 #' @param prog_status factor vector of progression status (`No progressive disease` versus `progressive disease`)
 #' @param lambda_control scale parameter of an exponential baseline hazard function
+#' @param beta_arm parameter associated to the treatment effect
+#' @param beta_prog_status parameter associated to a known progression on a RECIST assessment
 #' @return numeric vector of rates (lambda) per subject, aligned with `arm` and `prog_status`
 #' @noRd
 #' @keywords internal
@@ -427,7 +475,8 @@ example_fu = function(enrolres, recist, seed, lambda_censor = 0.3, lambda_contro
 
     if (!"Treatment" %in% arm_levels) { stop("One level of `arm` must be 'Treatment'. Found: ", paste(arm_levels, collapse = ", ")) }
 
-    lambda = lambda_control * exp(beta_arm*(arm == "Treatment") + beta_prog_status*(prog_status == "Progressive disease"))
+    beta = beta_arm*(arm == "Treatment") + beta_prog_status*(prog_status == "Progressive disease")
+    lambda = lambda_control * exp(beta)
 
   } else if(n_arm != 2) { stop("Wrong number of arms of treatment: ", n_arm, ". Only two arms of treatment are supported.")
   } else if(n_prog_status != 2) { stop("Wrong number of progression status: ", n_prog_status, ". Progression status must be a binary variable.")}
@@ -452,8 +501,8 @@ example_fu = function(enrolres, recist, seed, lambda_censor = 0.3, lambda_contro
 
   rtn$recist <- rtn$recist %>%
     left_join(rtn$fu %>% select(- crfname), by = join_by(subjid)) %>%
-    filter(rcdt <= fu_date) %>%
-    select(- c(fu_date, fu_status))
+    filter(rcdt <= os_date) %>%
+    select(- c(os_date, os_status))
 
   rtn
 }
@@ -467,7 +516,8 @@ example_fu = function(enrolres, recist, seed, lambda_censor = 0.3, lambda_contro
 #' @keywords internal
 #' @importFrom tibble tibble
 .simulate_one_patient = function(arm, rc_num_timepoints, rc_p_na,
-                                 rc_sd_tlsum_noise, rc_coef_treatement) {
+                                 rc_sd_tlsum_noise, rc_coef_treatement,
+                                 rc_subj_delai) {
   rc_num_timepoints = rc_num_timepoints-1 #add baseline later
   rctlsum_b = rnorm(1, 50, 30)
   rctlsum_b = ifelse(rctlsum_b <10, runif(1, 10, 150), rctlsum_b)
@@ -476,7 +526,7 @@ example_fu = function(enrolres, recist, seed, lambda_censor = 0.3, lambda_contro
                    percent_change_per_month > 0 ~ 1 / rc_coef_treatement,
                    .default = rc_coef_treatement)
   percent_change_per_month = percent_change_per_month * coef
-  subj_delai = 15 * seq(rc_num_timepoints) + runif(rc_num_timepoints, -7, 7)
+  subj_delai = rc_subj_delai * seq(rc_num_timepoints) + runif(rc_num_timepoints, -7, 7)
   percent_change = percent_change_per_month * subj_delai / 30.5
   percent_change = percent_change + rnorm(rc_num_timepoints, 0, rc_sd_tlsum_noise)
   percent_change = c(0, percent_change)
diff --git a/man/ae_plot_grade.Rd b/man/ae_plot_grade.Rd
index cbd764f..5e17547 100644
--- a/man/ae_plot_grade.Rd
+++ b/man/ae_plot_grade.Rd
@@ -26,7 +26,7 @@ ae_plot_grade(
 
 \item{variant}{one or several of \code{c("max", "sup", "eq")}. \code{max} computes the maximum AE grade per patient, \code{sup} computes the number of patients having experienced at least one AE of grade higher or equal to X, and \code{eq} computes the number of patients having experienced at least one AE of grade equal to X.}
 
-\item{position}{Position adjustment (cf. \code{\link[ggplot2:geom_bar]{ggplot2::geom_col()}})}
+\item{position}{Position adjustment (cf. \code{\link[ggplot2:geom_col]{ggplot2::geom_col()}})}
 
 \item{type}{whether to present patients as proportions (\code{relative}) or as counts (\code{absolute})}
 
diff --git a/man/example_fu.Rd b/man/example_fu.Rd
index 5d225ef..d41dd2e 100644
--- a/man/example_fu.Rd
+++ b/man/example_fu.Rd
@@ -8,10 +8,10 @@ example_fu(
   enrolres,
   recist,
   seed,
-  lambda_censor = 0.3,
+  lambda_censor = 0.2,
   lambda_control = 0.2,
   beta_arm = -0.6,
-  beta_prog_status = 0.5,
+  beta_prog_status = 2,
   ...
 )
 }
@@ -22,18 +22,24 @@ example_fu(
 
 \item{seed}{Integer. Random seed for reproducibility (can be \code{NULL}).}
 
-\item{...}{Additional arguments (currently unused).}
+\item{lambda_censor}{numeric. Rate parameter of an exponential distribution used to simulate censoring times.}
+
+\item{lambda_control}{numeric. Scale parameter of an exponential baseline hazard function.}
 
-\item{censor_min_days}{minimum time in days between last RECIST evaluation and administrative censoring}
+\item{beta_arm}{numeric. Parameter associated to the treatment effect.}
 
-\item{censor_max_days}{maximum time in days between last RECIST evaluation and administrative censoring}
+\item{beta_prog_status}{numeric. Parameter associated to a known progression on a RECIST assessment.}
+
+\item{...}{Additional arguments (currently unused).}
 }
 \value{
 A tibble with \code{N} rows and the following columns:
 \itemize{
 \item \code{subjid}: Subject ID
-\item \code{fu_status}: Status of the patient (0 - Alive/Censored , 1 - Dead)
-\item \code{fu_date}: Date of latest news (death date if dead, otherwise censoring date)
+\item \code{pfs_status}: Progression-free survival status of the patient
+\item \code{pfs_date}: Date of latest news before progression/death
+\item \code{os_status}: Overall survival status of the patient (0 - Alive/Censored , 1 - Dead)
+\item \code{os_date}: Date of latest news (death date if dead, otherwise censoring date)
 }
 }
 \description{
diff --git a/man/example_rc.Rd b/man/example_rc.Rd
index 9403d52..02e0d6b 100644
--- a/man/example_rc.Rd
+++ b/man/example_rc.Rd
@@ -14,6 +14,7 @@ example_rc(
   rc_p_nt_lesions_resp = list(CR = 0.73, SD = 0.25, PD = 0.01, NE = 0.01),
   rc_sd_tlsum_noise = 0.5,
   rc_coef_treatement = 3,
+  rc_subj_delai = 42,
   ...
 )
 }
diff --git a/man/grstat-package.Rd b/man/grstat-package.Rd
index 36f4467..01c05de 100644
--- a/man/grstat-package.Rd
+++ b/man/grstat-package.Rd
@@ -20,9 +20,14 @@ Useful links:
 \author{
 \strong{Maintainer}: Dan Chaltiel \email{dan.chaltiel@gustaveroussy.fr} (\href{https://orcid.org/0000-0003-3488-779X}{ORCID})
 
+Authors:
+\itemize{
+  \item Dan Chaltiel \email{dan.chaltiel@gustaveroussy.fr} (\href{https://orcid.org/0000-0003-3488-779X}{ORCID})
+}
+
 Other contributors:
 \itemize{
-  \item Team Oncostat (01ed4t417) [copyright holder, funder]
+  \item Team Oncostat (\href{https://ror.org/01ed4t417}{ROR}) [copyright holder, funder]
   \item Baptiste Archambaud \email{baptiste.archambaud@gustaveroussy.fr} [contributor]
   \item Charlotte Bargain \email{charlotte.bargain@gustaveroussy.fr} [contributor]
   \item Aldéric Fraslin \email{alderic.fraslin@gustaveroussy.fr} (\href{https://orcid.org/0000-0003-2235-0012}{ORCID}) [contributor]
diff --git a/man/grstat_example.Rd b/man/grstat_example.Rd
index 6d21e1a..0249a14 100644
--- a/man/grstat_example.Rd
+++ b/man/grstat_example.Rd
@@ -11,7 +11,7 @@ grstat_example(N = 200, seed = 42, ...)
 
 \item{seed}{the random seed (can be \code{NULL})}
 
-\item{...}{passed on to internal functions. See \code{\link[=example_enrol]{example_enrol()}},  \code{\link[=example_ae]{example_ae()}}, and  \code{\link[=example_rc]{example_rc()}} for the argument names.}
+\item{...}{passed on to internal functions. See \code{\link[=example_enrol]{example_enrol()}},  \code{\link[=example_ae]{example_ae()}}, \code{\link[=example_rc]{example_rc()}}, and \code{\link[=example_fu]{example_fu()}} for the argument names.}
 }
 \value{
 A list of datasets, like in EDCimport.
diff --git a/man/reexports.Rd b/man/reexports.Rd
index f0281b4..ddffaaa 100644
--- a/man/reexports.Rd
+++ b/man/reexports.Rd
@@ -13,10 +13,10 @@ These objects are imported from other packages. Follow the links
 below to see their documentation.
 
 \describe{
-  \item{dplyr}{\code{\link[dplyr:reexports]{\%>\%}}}
+  \item{dplyr}{\code{\link[dplyr:\%>\%]{\%>\%}}}
 
-  \item{flextable}{\code{\link[flextable]{as_flextable}}}
+  \item{flextable}{\code{\link[flextable:as_flextable]{as_flextable()}}}
 
-  \item{tibble}{\code{\link[tibble]{tibble}}}
+  \item{tibble}{\code{\link[tibble:tibble]{tibble()}}}
 }}
 
diff --git a/tests/testthat/test-data.R b/tests/testthat/test-data.R
index 349a706..906b612 100644
--- a/tests/testthat/test-data.R
+++ b/tests/testthat/test-data.R
@@ -98,6 +98,19 @@ test_that("RECIST plots", {
   rc = db$recist %>%
     left_join(x$enrolres, by="subjid", suffix=c("_bak", ""))
 
+  data_km = rc %>%
+    summarise(
+      date_pd = min_narm(rcdt[rcresp=="Progressive disease"]),
+      last_date = max_narm(rcdt),
+      .by=subjid
+    ) %>%
+    left_join(db$enrolres, by="subjid", suffix=c("_rc", "")) %>%
+    mutate(
+      date_pfs = pmin(date_pd, last_date, na.rm=TRUE),
+      time_pfs = as.numeric(date_pfs - date_inclusion)/30.4,
+      event_pfs = !is.na(date_pd)
+    )
+
   # KM plot
   km = survival::survfit(survival::Surv(time_pfs, event_pfs) ~ arm, data=data_km)
   ggsurvfit::ggsurvfit(km) +

From 6b6db22ed8d3fb736114ae0dd774e517be7c593e Mon Sep 17 00:00:00 2001
From: charlottebargain <charlotte.bargain@gmail.com>
Date: Thu, 11 Jun 2026 15:06:52 +0200
Subject: [PATCH 5/5] Simplified defensive checks and parameters names

Simplified defensive checks in internal functions :
- Removed redundant column validation
- Inputs are guaranteed by `grstat_example`

Prefix the parameters of `example_fu` function with `fu_` to avoid confusion about their role.
---
 R/data.R          | 99 +++++++++++++++++++----------------------------
 man/example_fu.Rd | 16 ++++----
 2 files changed, 48 insertions(+), 67 deletions(-)

diff --git a/R/data.R b/R/data.R
index 15217cb..8c7a2de 100644
--- a/R/data.R
+++ b/R/data.R
@@ -304,10 +304,10 @@ example_rc = function(enrolres, seed,
 #' @param enrolres the enrolment result table, from [example_enrol()].
 #' @param recist the recist table with progression date, from [example_rc()].
 #' @param seed Integer. Random seed for reproducibility (can be `NULL`).
-#' @param lambda_censor numeric. Rate parameter of an exponential distribution used to simulate censoring times.
-#' @param lambda_control numeric. Scale parameter of an exponential baseline hazard function.
-#' @param beta_arm numeric. Parameter associated to the treatment effect.
-#' @param beta_prog_status numeric. Parameter associated to a known progression on a RECIST assessment.
+#' @param fu_lambda_censor numeric. Rate parameter of an exponential distribution used to simulate censoring times.
+#' @param fu_lambda_control numeric. Scale parameter of an exponential baseline hazard function.
+#' @param fu_beta_arm numeric. Parameter associated to the treatment effect.
+#' @param fu_beta_prog_status numeric. Parameter associated to a known progression on a RECIST assessment.
 #' @param ... Additional arguments (currently unused).
 #'
 #' @return A tibble with `N` rows and the following columns:
@@ -324,40 +324,32 @@ example_rc = function(enrolres, seed,
 #' @importFrom stats rexp
 #' @importFrom forcats as_factor fct_recode
 example_fu = function(enrolres, recist, seed,
-                      lambda_censor = 0.2, lambda_control = 0.2,
-                      beta_arm = -0.6, beta_prog_status = 2, ...){
+                      fu_lambda_censor = 0.2, fu_lambda_control = 0.2,
+                      fu_beta_arm = -0.6, fu_beta_prog_status = 2, ...){
 
   stopifnot(is.data.frame(enrolres), is.data.frame(recist))
-  if (!all(c("subjid", "date_inclusion") %in% names(enrolres))) {
-    stop("`enrolres` must contain columns `subjid` and `date_inclusion`.")
-  }
-  if (!all(c("subjid", "rcdt", "rcresp") %in% names(recist))) {
-    stop("`recist` must contain columns `subjid`, `rcdt` (date of RECIST evaluation) and `rcresp` (RECIST global response).")
-  }
-  if(!("arm" %in% names(enrolres))) {
-    stop("`enrolres` must contain `arm`.")
-  }
-  if(!is.null(seed)) { set.seed(seed) }
+  assert_names_exists(enrolres, c("subjid", "date_inclusion", "arm"))
+  assert_names_exists(recist, c("subjid", "rcdt", "rcresp"))
 
-  enrolres = enrolres %>% arrange(subjid)
-  recist = recist %>% arrange(subjid)
+  assert_not_null(seed)
+  set.seed(seed)
 
-  prog_status = .progression_status(recist)
-  progression_date = .progression_date(recist)
+  data_progression_status = .progression_status(recist)
+  data_progression_date = .progression_date(recist)
 
   # Survival rate assigned for each subject, according to arm of treatment and progression status
-  lambda = .surv_coef(arm = enrolres$arm, prog_status = prog_status$status,
-                      lambda_control = lambda_control, beta_arm = beta_arm,
-                      beta_prog_status = beta_prog_status)
+  lambda = .surv_coef(arm = enrolres$arm, prog_status = data_progression_status$status,
+                      fu_lambda_control = fu_lambda_control, fu_beta_arm = fu_beta_arm,
+                      fu_beta_prog_status = fu_beta_prog_status)
 
   time_to_death = rexp(n = nrow(enrolres), rate = lambda)
-  time_to_censor = rexp(n = nrow(enrolres), rate = lambda_censor)
+  time_to_censor = rexp(n = nrow(enrolres), rate = fu_lambda_censor)
 
   fu_data = enrolres %>%
 
     select(subjid, date_inclusion) %>%
-    left_join(prog_status, by = join_by(subjid)) %>%
-    left_join(progression_date, by = join_by(subjid)) %>%
+    left_join(data_progression_status, by = join_by(subjid)) %>%
+    left_join(data_progression_date, by = join_by(subjid)) %>%
 
     mutate(
       date_inclusion = as.Date(date_inclusion),
@@ -383,6 +375,7 @@ example_fu = function(enrolres, recist, seed,
                         true = death_date,
                         false = censoring_date)
     ) %>%
+    arrange(subjid) %>%
     select(subjid, pfs_status, pfs_date, os_status, os_date) %>%
     apply_labels(subjid = "Subject ID",
                  pfs_status = "Status PFS",
@@ -408,16 +401,15 @@ example_fu = function(enrolres, recist, seed,
 .progression_status = function(recist) {
 
   stopifnot(is.data.frame(recist))
-  if (!all(c("subjid", "rcresp") %in% names(recist))) {
-    stop("`recist` must contain columns `subjid`, and `rcresp` (RECIST global response).")
-  }
+  assert_names_exists(recist, c("subjid", "rcresp"))
 
-  rcresp_levels = levels(recist$rcresp)
-  if (!"Progressive disease" %in% rcresp_levels) { stop("One level of `rcresp_levels` must be 'Progressive disease'. Found: ", paste(rcresp_levels, collapse = ", ")) }
+  if(!any(recist$rcresp == "Progressive disease")) {
+    cli::cli_abort("No patient has 'Progressive disease' in `recist$rcresp`.")
+  }
 
   progression_status = recist %>%
     select(subjid, rcresp) %>%
-    summarise(status = any(rcresp == "Progressive disease"), .by = 'subjid') %>%
+    summarise(status = any(rcresp == "Progressive disease", na.rm = TRUE), .by = 'subjid') %>%
     mutate(status = as_factor(if_else(!is.na(status),
                                       true = "Progressive disease",
                                       false = "No progression")))
@@ -434,6 +426,11 @@ example_fu = function(enrolres, recist, seed,
 .progression_date = function(recist) {
 
   stopifnot(is.data.frame(recist))
+  assert_names_exists(recist, c("subjid", "rcresp", "rcdt"))
+
+  if(!any(recist$rcresp == "Progressive disease")) {
+    cli::cli_abort("No patient has 'Progressive disease' in `recist$rcresp`.")
+  }
 
   progression_date = recist %>%
     summarise(
@@ -448,39 +445,23 @@ example_fu = function(enrolres, recist, seed,
 #' Assign a survival rate parameter (exponential lambda) for each patient, depending on treatment and progression status
 #' @param arm factor vector of treatment arms (`Control` versus `Treatment`)
 #' @param prog_status factor vector of progression status (`No progressive disease` versus `progressive disease`)
-#' @param lambda_control scale parameter of an exponential baseline hazard function
-#' @param beta_arm parameter associated to the treatment effect
-#' @param beta_prog_status parameter associated to a known progression on a RECIST assessment
+#' @param fu_lambda_control scale parameter of an exponential baseline hazard function
+#' @param fu_beta_arm parameter associated to the treatment effect
+#' @param fu_beta_prog_status parameter associated to a known progression on a RECIST assessment
 #' @return numeric vector of rates (lambda) per subject, aligned with `arm` and `prog_status`
 #' @noRd
 #' @keywords internal
-.surv_coef = function(arm, prog_status, lambda_control, beta_arm, beta_prog_status) {
+.surv_coef = function(arm, prog_status, fu_lambda_control, fu_beta_arm, fu_beta_prog_status) {
 
-  if(is.null(arm)) {stop("`arm` is NULL.")}
-  if(is.null(prog_status)) {stop("`prog_status` is NULL.")}
+  assert_not_null(arm)
+  assert_not_null(prog_status)
 
-  if(!is.factor(arm)) arm = as_factor(arm)
-  if(!is.factor(prog_status)) prog_status = mutate(prog_status, status = as_factor(status))
-
-  arm_levels = levels(arm)
-  n_arm = nlevels(arm)
-  if (!"Control" %in% arm_levels) { stop("One level of `arm` must be 'Control'. Found: ", paste(arm_levels, collapse = ", ")) }
-
-  prog_status_levels = levels(prog_status)
-  n_prog_status = nlevels(prog_status)
-  if (!"Progressive disease" %in% prog_status_levels) { stop("One level of `prog_status` must be 'Progressive disease'. Found: ", paste(prog_status_levels, collapse = ", ")) }
-
-
-  if(n_arm == 2 & n_prog_status == 2) { # "Control" versus "Treatment" and "Progressive disease" versus "No Progressive disease (stable or response)"
-
-    if (!"Treatment" %in% arm_levels) { stop("One level of `arm` must be 'Treatment'. Found: ", paste(arm_levels, collapse = ", ")) }
-
-    beta = beta_arm*(arm == "Treatment") + beta_prog_status*(prog_status == "Progressive disease")
-    lambda = lambda_control * exp(beta)
-
-  } else if(n_arm != 2) { stop("Wrong number of arms of treatment: ", n_arm, ". Only two arms of treatment are supported.")
-  } else if(n_prog_status != 2) { stop("Wrong number of progression status: ", n_prog_status, ". Progression status must be a binary variable.")}
+  if(nlevels(arm) != 2) {
+    cli::cli_abort("Number of arm of treatment must be 2 (`Control` versus `Treatment`")
+  }
 
+  beta = fu_beta_arm*(arm == "Treatment") + fu_beta_prog_status*(prog_status == "Progressive disease")
+  lambda = fu_lambda_control * exp(beta)
   lambda
 }
 
diff --git a/man/example_fu.Rd b/man/example_fu.Rd
index d41dd2e..c76fbab 100644
--- a/man/example_fu.Rd
+++ b/man/example_fu.Rd
@@ -8,10 +8,10 @@ example_fu(
   enrolres,
   recist,
   seed,
-  lambda_censor = 0.2,
-  lambda_control = 0.2,
-  beta_arm = -0.6,
-  beta_prog_status = 2,
+  fu_lambda_censor = 0.2,
+  fu_lambda_control = 0.2,
+  fu_beta_arm = -0.6,
+  fu_beta_prog_status = 2,
   ...
 )
 }
@@ -22,13 +22,13 @@ example_fu(
 
 \item{seed}{Integer. Random seed for reproducibility (can be \code{NULL}).}
 
-\item{lambda_censor}{numeric. Rate parameter of an exponential distribution used to simulate censoring times.}
+\item{fu_lambda_censor}{numeric. Rate parameter of an exponential distribution used to simulate censoring times.}
 
-\item{lambda_control}{numeric. Scale parameter of an exponential baseline hazard function.}
+\item{fu_lambda_control}{numeric. Scale parameter of an exponential baseline hazard function.}
 
-\item{beta_arm}{numeric. Parameter associated to the treatment effect.}
+\item{fu_beta_arm}{numeric. Parameter associated to the treatment effect.}
 
-\item{beta_prog_status}{numeric. Parameter associated to a known progression on a RECIST assessment.}
+\item{fu_beta_prog_status}{numeric. Parameter associated to a known progression on a RECIST assessment.}
 
 \item{...}{Additional arguments (currently unused).}
 }