Merge branch 'main' into 2141_supersede_derive_param_extreme_record

.github/workflows/templates.yml

@@ -6,7 +6,7 @@ on:
   pull_request:
     branches:
       - main
-      - devel
+      - patch
 
 jobs:
   templates:

NEWS.md

@@ -58,6 +58,13 @@ were enhanced such that more than one summary variable can be derived, e.g.,
 
 ## Various
 
+# admiral 0.12.3
+
+- Fixed a bug in `derive_var_dthcaus()` where if a subject has observations in
+more than one of the sources, the one from the last source was selected
+regardless of the date. Now the function works as described in its
+documentation. (#2154)
+
 # admiral 0.12.2
 
 - A unit test for `derive_param_computed()` was modified in anticipation of major user-facing changes to R version 4.4 (#2147)

R/data.R

@@ -88,7 +88,7 @@
 #' Note: Variables `SOC`, `TERM`, `Grade 1`, `Grade 2`,`Grade 3`,`Grade 4`,`Grade 5`, `Definition`
 #' are from the source document on NCI-CTC website defining the grading criteria.
 #' [**Common Terminology Criteria for Adverse Events (CTCAE)v4.0**](https://ctep.cancer.gov/protocoldevelopment/electronic_applications/ctc.htm#ctc_40)
-#' From these variables only 'TERM' is used in the {admiral} code, the rest are for information and
+#' From these variables only 'TERM' is used in the `{admiral}` code, the rest are for information and
 #' traceability only.
 #'
 #'
@@ -124,7 +124,7 @@
 #' Note: Variables `SOC`, `TERM`, `Grade 1`, `Grade 2`,`Grade 3`,`Grade 4`,`Grade 5`, `Definition`
 #' are from the source document on NCI-CTC website defining the grading criteria.
 #' [**Common Terminology Criteria for Adverse Events (CTCAE)v5.0**](https://ctep.cancer.gov/protocoldevelopment/electronic_applications/ctc.htm#ctc_50)
-#' From these variables only 'TERM' is used in the {admiral} code, the rest are for information and
+#' From these variables only 'TERM' is used in the `{admiral}` code, the rest are for information and
 #' traceability only.
 #'
 #'
@@ -162,7 +162,7 @@
 #' Note: Variables `SOC`, `TERM`, `SUBGROUP`, `Grade 1`, `Grade 2`,`Grade 3`,`Grade 4`,`Grade 5`, `Definition`
 #' are from the source document on DAIDS website defining the grading criteria.
 #' [Division of AIDS (DAIDS) Table for Grading the Severity of Adult and Pediatric Adverse Events
-#' From these variables only 'TERM' is used in the {admiral} code, the rest are for information and
+#' From these variables only 'TERM' is used in the `{admiral}` code, the rest are for information and
 #' traceability only.
 #'
 #'

R/derive_date_vars.R

@@ -1248,7 +1248,7 @@ compute_tmf <- function(dtc,
 #' The date can be imputed (see `date_imputation` argument)
 #' and the date imputation flag ('`--DTF'`) can be added.
 #'
-#' In {admiral} we don't allow users to pick any single part of the date/time to
+#' In `{admiral}` we don't allow users to pick any single part of the date/time to
 #' impute, we only enable to impute up to a highest level, i.e. you couldn't
 #' choose to say impute months, but not days.
 #'
@@ -1456,7 +1456,7 @@ derive_vars_dt <- function(dataset, # nolint: cyclocomp_linter
 #' The date and time can be imputed (see `date_imputation`/`time_imputation` arguments)
 #' and the date/time imputation flag (`'--DTF'`, `'--TMF'`) can be added.
 #'
-#' In {admiral} we don't allow users to pick any single part of the date/time to
+#' In `{admiral}` we don't allow users to pick any single part of the date/time to
 #' impute, we only enable to impute up to a highest level, i.e. you couldn't
 #' choose to say impute months, but not days.
 #'

R/derive_var_dthcaus.R

@@ -160,6 +160,8 @@ derive_var_dthcaus <- function(dataset,
 
   # process each source
   add_data <- vector("list", length(sources))
+  tmp_source_nr <- get_new_tmp_var(dataset)
+  tmp_date <- get_new_tmp_var(dataset)
   for (ii in seq_along(sources)) {
     source_dataset_name <- sources[[ii]]$dataset_name
     source_dataset <- source_datasets[[source_dataset_name]]
@@ -183,8 +185,6 @@ derive_var_dthcaus <- function(dataset,
     )
 
     # if several death records, use the first/last according to 'mode'
-    tmp_source_nr <- get_new_tmp_var(dataset)
-    tmp_date <- get_new_tmp_var(dataset)
     add_data[[ii]] <- add_data[[ii]] %>%
       filter_extreme(
         order = exprs(!!date_var, !!!sources[[ii]]$order),

R/derive_vars_dtm_to_tm.R

@@ -12,7 +12,7 @@
 #' @details
 #' The names of the newly added variables are automatically set by replacing the
 #' `--DTM` suffix of the `source_vars` with `--TM`. The `--TM` variables are created
-#' using the {hms} package.
+#' using the `{hms}` package.
 #'
 #' @return
 #' A data frame containing the input dataset with the corresponding time

R/duplicates.R

@@ -4,7 +4,7 @@
 #'
 #'
 #' @details
-#' Many {admiral} function check that the input dataset contains only one record
+#' Many `{admiral}` function check that the input dataset contains only one record
 #' per `by_vars` group and throw an error otherwise. The `get_duplicates_dataset()`
 #' function allows one to retrieve the duplicate records that lead to an error.
 #'

tests/testthat/test-derive_var_dthcaus.R

@@ -414,8 +414,8 @@ test_that("derive_var_dthcaus Test 8: `dataset` is sorted using the `order` para
   expect_dfs_equal(expected_output, actual_output, keys = "USUBJID")
 })
 
-## Test 9: returns a error when traceability_vars is used ----
-test_that("derive_var_dthcaus Test 9: returns a error when traceability_vars is used", {
+## Test 9: returns a warning when traceability_vars is used ----
+test_that("derive_var_dthcaus Test 9: returns a warning when traceability_vars is used", {
   ae <- tibble::tribble(
     ~STUDYID, ~USUBJID, ~AESEQ, ~AEDECOD, ~AEOUT, ~AEDTHDTC,
     "TEST01", "PAT01", 12, "SUDDEN DEATH", "FATAL", "2021-04-04"
@@ -437,3 +437,115 @@ test_that("derive_var_dthcaus Test 9: returns a error when traceability_vars is
     class = "lifecycle_error_deprecated"
   )
 })
+
+## Test 10: multiple observations from different sources ----
+test_that("derive_var_dthcaus Test 10: multiple observations from different sources", {
+  expected <- tibble::tribble(
+    ~STUDYID, ~USUBJID, ~DTHCAUS,
+    "TEST01", "PAT01",  "SUDDEN DEATH",
+    "TEST01", "PAT02",  NA_character_,
+    "TEST01", "PAT03",  "DEATH DUE TO progression of disease"
+  )
+
+  adsl <- select(expected, -DTHCAUS)
+
+  ae <- tibble::tribble(
+    ~STUDYID, ~USUBJID, ~AESEQ, ~AEDECOD, ~AEOUT, ~AEDTHDTC,
+    "TEST01", "PAT01", 12, "SUDDEN DEATH", "FATAL", "2021-04-04"
+  ) %>%
+    mutate(
+      AEDTHDT = ymd(AEDTHDTC)
+    )
+
+  ds <- tibble::tribble(
+    ~STUDYID, ~USUBJID, ~DSSEQ, ~DSDECOD, ~DSTERM, ~DSSTDTC,
+    "TEST01", "PAT01", 4, "DEATH", "DEATH DUE TO progression of disease", "2021-04-05",
+    "TEST01", "PAT02", 1, "INFORMED CONSENT OBTAINED", "INFORMED CONSENT OBTAINED", "2021-04-02",
+    "TEST01", "PAT02", 2, "RANDOMIZATION", "RANDOMIZATION", "2021-04-11",
+    "TEST01", "PAT02", 3, "COMPLETED", "PROTOCOL COMPLETED", "2021-12-01",
+    "TEST01", "PAT03", 1, "DEATH", "DEATH DUE TO progression of disease", "2021-04-07",
+    "TEST01", "PAT03", 2, "RANDOMIZATION", "RANDOMIZATION", "2021-04-11",
+    "TEST01", "PAT03", 3, "COMPLETED", "PROTOCOL COMPLETED", "2021-12-01"
+  )
+
+  # Derive `DTHCAUS` only - for on-study deaths only
+  src_ae <- dthcaus_source(
+    dataset_name = "ae",
+    filter = AEOUT == "FATAL",
+    date = convert_dtc_to_dt(AEDTHDTC),
+    mode = "first",
+    dthcaus = AEDECOD
+  )
+
+  src_ds <- dthcaus_source(
+    dataset_name = "ds",
+    filter = DSDECOD == "DEATH" & grepl("DEATH DUE TO", DSTERM),
+    date = convert_dtc_to_dt(DSSTDTC),
+    mode = "first",
+    dthcaus = DSTERM
+  )
+  actual <- adsl %>%
+    derive_var_dthcaus(src_ae, src_ds, source_datasets = list(ae = ae, ds = ds))
+
+  expect_dfs_equal(
+    base = expected,
+    compare = actual,
+    keys = c("USUBJID")
+  )
+})
+
+## Test 11: multiple observations from different sources with same date ----
+test_that("derive_var_dthcaus Test 11: multiple observations with same date", {
+  expected <- tibble::tribble(
+    ~STUDYID, ~USUBJID, ~DTHCAUS,
+    "TEST01", "PAT01",  "SUDDEN DEATH",
+    "TEST01", "PAT02",  NA_character_,
+    "TEST01", "PAT03",  "DEATH DUE TO progression of disease"
+  )
+
+  adsl <- select(expected, -DTHCAUS)
+
+  ae <- tibble::tribble(
+    ~STUDYID, ~USUBJID, ~AESEQ, ~AEDECOD, ~AEOUT, ~AEDTHDTC,
+    "TEST01", "PAT01", 12, "SUDDEN DEATH", "FATAL", "2021-04-05"
+  ) %>%
+    mutate(
+      AEDTHDT = ymd(AEDTHDTC)
+    )
+
+  ds <- tibble::tribble(
+    ~STUDYID, ~USUBJID, ~DSSEQ, ~DSDECOD, ~DSTERM, ~DSSTDTC,
+    "TEST01", "PAT01", 4, "DEATH", "DEATH DUE TO progression of disease", "2021-04-05",
+    "TEST01", "PAT02", 1, "INFORMED CONSENT OBTAINED", "INFORMED CONSENT OBTAINED", "2021-04-02",
+    "TEST01", "PAT02", 2, "RANDOMIZATION", "RANDOMIZATION", "2021-04-11",
+    "TEST01", "PAT02", 3, "COMPLETED", "PROTOCOL COMPLETED", "2021-12-01",
+    "TEST01", "PAT03", 1, "DEATH", "DEATH DUE TO progression of disease", "2021-04-07",
+    "TEST01", "PAT03", 2, "RANDOMIZATION", "RANDOMIZATION", "2021-04-11",
+    "TEST01", "PAT03", 3, "COMPLETED", "PROTOCOL COMPLETED", "2021-12-01"
+  )
+
+  # Derive `DTHCAUS` only - for on-study deaths only
+  src_ae <- dthcaus_source(
+    dataset_name = "ae",
+    filter = AEOUT == "FATAL",
+    date = convert_dtc_to_dt(AEDTHDTC),
+    mode = "first",
+    dthcaus = AEDECOD
+  )
+
+  src_ds <- dthcaus_source(
+    dataset_name = "ds",
+    filter = DSDECOD == "DEATH" & grepl("DEATH DUE TO", DSTERM),
+    date = convert_dtc_to_dt(DSSTDTC),
+    mode = "first",
+    dthcaus = DSTERM
+  )
+  actual <- adsl %>%
+    derive_var_dthcaus(src_ae, src_ds, source_datasets = list(ae = ae, ds = ds))
+
+  expect_dfs_equal(
+    base = expected,
+    compare = actual,
+    keys = c("USUBJID")
+  )
+})