Creating ADCE"

In admiralvaccine: Vaccine Extension Package for ADaM in 'R' Asset Library

# Other variables
admiral_homepage <- "https://pharmaverse.github.io/admiral"

Introduction

This article describes creating an ADCE ADaM for the analysis of Vaccine Reactogenicity Data collected in SDTM CE domain. The current presented example is tested using CE SDTM domains and ADSL ADaM domain. However, other domains could be used if needed (eg temperature data collected in VS).

Note: All examples assume CDISC SDTM and/or ADaM format as input unless otherwise specified.

Programming Flow

• Read in Data
• Pre-processing of Input Dataset
• Create Reference Dataset for Periods
• Derivation of Analysis Dates
• Join with the Periods Reference Dataset and Derive Relative Day in Period
• Creation of Analysis Version for GRADING Variable (Either TOXGR or SEV)
• Creation of Analysis Sequence Number
• Final Step : Get All the Remaining Variables from ADSL

Read in Data {#readdata}

Assumption: The CE domain has already been merged with the SUPPCE dataset. If this is not yet the case, join SUPPCE onto parent CE domain using metatools::combine_supp(CE, SUPPCE).

library(admiraldev)
library(admiral)
library(dplyr)
library(lubridate)
library(admiralvaccine)
library(pharmaversesdtm)

data("ce_vaccine")
data("admiralvaccine_adsl")

adsl <- admiralvaccine_adsl
ce <- ce_vaccine

ce <- convert_blanks_to_na(ce)
adsl <- convert_blanks_to_na(adsl)

dataset_vignette(
  adsl,
  display_vars = exprs(USUBJID, TRTSDT, TRTEDT, TRT01A, AP01SDT, AP01EDT, AP02SDT, AP02EDT)
)

Pre-processing of Input Dataset {#cefilter}

This step involves company-specific pre-processing of required input dataset for further analysis. In this step, we will filter records that has only reactogenicity events.

adce <- ce %>%
  filter(CECAT == "REACTOGENICITY")

Create Reference Dataset for Periods {#adperiods}

Create period dataset - for joining period information onto CE records. Need to remove datetime variables as otherwise causes duplicate issues.

adsl2 <- adsl %>%
  select(-c(starts_with("AP") & ends_with("DTM")))

adperiods <- create_period_dataset(
  adsl2,
  new_vars = exprs(APERSDT = APxxSDT, APEREDT = APxxEDT)
)

dataset_vignette(
  adperiods,
  display_vars = exprs(USUBJID, APERIOD, APERSDT, APEREDT)
)

Derivation of Analysis Dates {#adates}

At this step, it may be useful to join ADSL to your CE domain. Only the ADSL variables used for derivations are selected at this step. The rest of the relevant ADSL variables would be added later.

adsl_vars <- exprs(TRTSDT, TRTEDT)

adce <- adce %>%
  # join ADSL to CE
  derive_vars_merged(
    dataset_add = adsl,
    new_vars = adsl_vars,
    by = get_admiral_option("subject_keys")
  ) %>%
  derive_vars_dt(
    dtc = CESTDTC,
    new_vars_prefix = "AST",
    highest_imputation = "n"
  ) %>%
  derive_vars_dt(
    dtc = CEENDTC,
    new_vars_prefix = "AEN",
    highest_imputation = "n"
  ) %>%
  derive_vars_dy(
    reference_date = TRTSDT,
    source_vars = exprs(ASTDT, AENDT)
  )

dataset_vignette(
  adce,
  display_vars = exprs(USUBJID, TRTSDT, CESTDTC, CEENDTC, ASTDT, AENDT, ASTDY, AENDY)
)

Join with the Periods Reference Dataset and Derive Relative Day in Period {#mergeperiods}

Also add analysis version of CEREL(AREL).

adce <-
  derive_vars_joined(
    adce,
    dataset_add = adperiods,
    by_vars = get_admiral_option("subject_keys"),
    filter_join = ASTDT >= APERSDT & ASTDT <= APEREDT,
    join_type = "all"
  ) %>%
  mutate(
    APERSTDY = as.integer(ASTDT - APERSDT) + 1,
    AREL = CEREL
  )

dataset_vignette(
  adce,
  display_vars = exprs(USUBJID, TRTSDT, ASTDT, AENDT, ASTDY, AENDY, APERIOD, APERSDT, APERSTDY)
)

Creation of Analysis Version for GRADING Variable (Either TOXGR or SEV) {#agrade}

Depending on which variable is collected for the Grading (TOXGR or SEV) in CE domain, derive the associated analysis version. In current example, SEV is collected, so the code is using this as an example. In addition, derivation of Extreme Flags: in current example: flag the first occurrence of the most severe grade within a Period (AOCC01FL).

adce <- adce %>%
  mutate(
    ASEV = CESEV,
    ASEVN = as.integer(factor(ASEV,
      levels = c("MILD", "MODERATE", "SEVERE", "DEATH THREATENING")
    ))
  ) %>%
  restrict_derivation(
    derivation = derive_var_extreme_flag,
    args = params(
      by_vars = exprs(USUBJID, APERIOD),
      order = exprs(desc(ASEVN), ASTDY, CEDECOD),
      new_var = AOCC01FL,
      mode = "first"
    ),
    filter = !is.na(APERIOD) & !is.na(ASEV)
  )

dataset_vignette(
  adce,
  display_vars = exprs(
    USUBJID, TRTSDT, ASTDT, APERIOD, APERSDT,
    APERSTDY, CEDECOD, ASEVN, AOCC01FL, CESEQ
  )
)

Creation of Analysis Sequence Number {#aseq}

adce <- adce %>%
  derive_var_obs_number(
    new_var = ASEQ,
    by_vars = get_admiral_option("subject_keys"),
    order = exprs(CEDECOD, CELAT, CETPTREF, APERIOD),
    check_type = "error"
  ) %>%
  derive_vars_duration(
    new_var = ADURN,
    new_var_unit = ADURU,
    start_date = ASTDT,
    end_date = AENDT,
    in_unit = "days",
    out_unit = "days",
    add_one = TRUE,
    trunc_out = FALSE
  )

dataset_vignette(
  adce,
  display_vars = exprs(
    USUBJID, TRTSDT, ASTDT, APERIOD, APERSDT,
    APERSTDY, CEDECOD, ASEVN, AOCC01FL, CESEQ, ASEQ
  )
)

Final Step : Get All the Remaining Variables from ADSL {#jadsl_list}

Get list of ADSL vars as per trial specific which needs to be adjusted when using the template

adsl_list <- adsl %>%
  select(STUDYID, USUBJID, TRT01A, TRT01P, AGE, AGEU, SEX, RACE, COUNTRY, ETHNIC, SITEID, SUBJID)

adce <- adce %>%
  derive_vars_merged(
    dataset_add = adsl_list,
    by_vars = get_admiral_option("subject_keys")
  )

dataset_vignette(
  adce,
  display_vars = exprs(
    USUBJID, TRTSDT, ASTDT, APERIOD, APERSDT, APERSTDY,
    CEDECOD, ASEVN, AOCC01FL, CESEQ, ASEQ, AGE, SEX
  )
)

Example Script

ADaM | Sample Code ---- | -------------- ADCE | ad_adce.R{target="_blank"}

admiralvaccine documentation built on Sept. 11, 2024, 6:35 p.m.