Skip to content

Commit

Permalink
Merge branch 'main' into 32_advs_template_growth_parameters
Browse files Browse the repository at this point in the history
  • Loading branch information
rossfarrugia authored May 24, 2024
2 parents f16050f + 8241b8b commit a93a2cc
Show file tree
Hide file tree
Showing 8 changed files with 257 additions and 230 deletions.
2 changes: 1 addition & 1 deletion DESCRIPTION
Original file line number Diff line number Diff line change
@@ -1,7 +1,7 @@
Type: Package
Package: admiralpeds
Title: Pediatrics Extension Package for ADaM in 'R' Asset Library
Version: 0.1.0.9009
Version: 0.1.0.9010
Authors@R:
person("Ross", "Farrugia", , "[email protected]", role = c("aut", "cre"))
Description: A toolbox for programming Clinical Data Standards Interchange
Expand Down
2 changes: 1 addition & 1 deletion NEWS.md
Original file line number Diff line number Diff line change
@@ -1,3 +1,3 @@
# admiralpeds 0.1.0.9009
# admiralpeds 0.1.0.9010

- Initial package release focused mainly on child growth/development charts.
98 changes: 33 additions & 65 deletions data-raw/dm_peds.R
Original file line number Diff line number Diff line change
@@ -1,76 +1,44 @@
# Dataset: dm_ped
# Dataset: dm_peds
# Description: Create DM test SDTM dataset for pediatric studies

# Load libraries -----
library(pharmaversesdtm) # TODO remove when this script is moved to pharmaversesdtm
library(dplyr)
library(purrr)
library(admiral)
library(pharmaversesdtm)

# Create DM for pediatric ----

## Read in original data ----
# Read input test data from pharmaversesdtm ----
data("dm")

## Make dm_ped dataset
dm_peds <- tibble::tribble(
~USUBJID, ~BRTHDTC, ~AGE, ~AGEU, ~RFSTDTC, ~RFENDTC, ~SEX,
"PEDS-1001", "2022-09-10", 0, "YEARS", "2022-09-30", "2023-12-30", "M",
"PEDS-1002", "2022-06-10", 0, "YEARS", "2022-09-12", "2024-02-15", "F",
"PEDS-1003", "2022-07-10", 0, "YEARS", "2023-01-08", "2023-10-25", "F",
"PEDS-1005", "2019-07-10", 2, "YEARS", "2021-07-09", "2024-01-13", "F",
"PEDS-1006", "2021-08-10", 2, "YEARS", "2023-08-11", "2024-02-15", "F",
"PEDS-1010", "2019", 2, "YEARS", "2021-07-09", "2024-02-15", "M",
"PEDS-1012", "2016-10-10", 6, "YEARS", "2023-06-23", "2024-02-15", "M",
"PEDS-1013", "2012-01-10", 12, "YEARS", "2024-01-10", "2024-02-15", "F",
"PEDS-1009", "2005-10-25", 20, "YEARS", "2024-01-12", "2024-02-15", "M",
) %>%
mutate(
STUDYID = "PEDS SAMPLE STUDY",
DOMAIN = "DM",
SUBJID = substr(USUBJID, 5, 9),
SITEID = substr(USUBJID, 5, 7),
RFXSTDTC = format(admiral::convert_dtc_to_dt(RFSTDTC) + 21, "%Y-%m-%d"),
RFXENDTC = format(admiral::convert_dtc_to_dt(RFENDTC) + 62, "%Y-%m-%d"),
RFICDTC = RFSTDTC,
RFPENDTC = RFENDTC,
DTHDTC = NA_character_,
DTHFL = NA_character_,
RACE = case_when(
grepl("100", USUBJID) ~ "WHITE",
grepl("101", USUBJID) ~ "BLACK OR AFRICAN AMERICAN",
TRUE ~ NA_character_
),
ETHNIC = case_when(
grepl("100", USUBJID) ~ "NOT HISPANIC OR LATINO",
grepl("101", USUBJID) ~ "NOT HISPANIC OR LATINO",
TRUE ~ NA_character_
),
COUNTRY = case_when(
grepl("100", USUBJID) ~ "USA",
grepl("101", USUBJID) ~ "ITA",
TRUE ~ NA_character_
),
ARMCD = "A",
ARM = "Arm A",
ACTARM = ARM,
ACTARMCD = ARMCD,
DMDTC = NA_character_,
DMDY = NA_integer_
) %>%
select(
STUDYID, DOMAIN, USUBJID, SUBJID, RFSTDTC, RFENDTC, RFXSTDTC, RFXENDTC,
RFICDTC, RFPENDTC, DTHDTC, DTHFL, SITEID, BRTHDTC, AGE, AGEU,
SEX, RACE, ETHNIC, ARMCD, ARM, ACTARMCD, ACTARM, COUNTRY, DMDTC, DMDY
)

# get common column names
common_cols <- seq_along(intersect(names(dm_peds), names(dm)))
# Apply label
lapply(common_cols, function(x) {
attr(dm_peds[[common_cols[x]]], "label") <- attr(dm[[common_cols[x]]], "label")
})

# Convert blank to NA ----
dm <- convert_blanks_to_na(dm)

# Subset to first 5 patients only (which is enough for our examples) ----
dm_subset <- dm %>%
filter(USUBJID %in% c(
"01-701-1015", "01-701-1023", "01-701-1028",
"01-701-1033", "01-701-1034"
))

# Add birth dates/age realistic for pediatrics in line with treatment dates ----
dm_peds <- dm_subset %>%
mutate(BRTHDTC = case_when(
USUBJID == "01-701-1015" ~ "2013-01-02",
USUBJID == "01-701-1023" ~ "2010-08-05",
USUBJID == "01-701-1028" ~ "2010-07-19",
USUBJID == "01-701-1033" ~ "2014-01-01",
USUBJID == "01-701-1034" ~ "2014-06-01"
)) %>%
mutate(AGE = case_when(
USUBJID == "01-701-1015" ~ 1,
USUBJID == "01-701-1023" ~ 2,
USUBJID == "01-701-1028" ~ 3,
USUBJID == "01-701-1033" ~ 0,
USUBJID == "01-701-1034" ~ 0
))

# Variable labels ----
attr(dm_peds$BRTHDTC, "label") <- "Date/Time of Birth"
attr(dm_peds$AGE, "label") <- "Age"

# Label dataset ----
attr(dm_peds, "label") <- "Demographics"
Expand Down
Loading

0 comments on commit a93a2cc

Please sign in to comment.