Skip to contents

Simulate a Portfolio of Studies

SimulatePortfolio.Rd

Creates a portfolio of multiple synthetic studies by resampling from source data. Each study is generated using ResampleStudy with configurable parameters.

Usage

SimulatePortfolio(
  lRaw,
  dfConfig = NULL,
  nStudies = 5,
  seed = NULL,
  vSubjectIDs = c("subjid", "subjectenrollmentnumber", "subject_nsv", "subjectname",
    "subjectid"),
  vSiteIDs = c("invid", "siteid", "site_num")
)

Arguments

lRaw

list. Named list of raw data domains (must include Raw_SUBJ).

dfConfig

data.frame. Optional configuration for each study. Must contain columns 'studyid' and 'nSubjects'. See Details.

nStudies

numeric. Number of studies to generate if dfConfig is NULL. Default: 5.

seed

numeric. Random seed for reproducibility. Default: NULL.

vSubjectIDs

Character vector defining hierarchical order for subject ID column lookup. Default: c("subjid", "subjectenrollmentnumber", "subject_nsv")

vSiteIDs

Character vector defining hierarchical order for site ID column lookup. Default: c("invid", "siteid", "site_num")

Value

list. Combined portfolio with same structure as lRaw, containing data from all simulated studies row-bound together.

Details

If dfConfig is provided, it should contain:

  • studyid: Unique study identifier

  • nSubjects: Number of subjects per study

  • TargetSiteCount: (Optional) Target number of sites

  • strOversampleDomain: (Optional) Domain for stratified sampling

  • vOversamplQuantileRange_min/max: (Optional) Quantile range for stratification

  • replacement: (Optional) Sample with replacement (default TRUE)

If dfConfig is NULL, generates default configuration with random parameters.

Domains are automatically processed based on columns present:

  • Subject-level domains (have vSubjectIDs columns) are filtered to sampled subjects

  • Site-level domains (have vSiteIDs columns but no subject columns) are filtered to used sites

  • Other domains have studyid updated only

Examples

# Simple portfolio with default parameters
lRaw <- list(
  Raw_SUBJ = clindata::rawplus_dm,
  Raw_AE = clindata::rawplus_ae,
  Raw_SITE = clindata::ctms_site
)

# Add required columns to Raw_SUBJ
lRaw$Raw_SUBJ$subjectname <- lRaw$Raw_SUBJ$subject_nsv
lRaw$Raw_SUBJ$subjectenrollmentnumber <- lRaw$Raw_SUBJ$subjid

lPortfolio <- SimulatePortfolio(lRaw, nStudies = 3, seed = 123)

# Custom configuration
dfConfig <- data.frame(
  studyid = c("TRIAL001", "TRIAL002"),
  nSubjects = c(50, 75),
  TargetSiteCount = c(15, 20)
)
lPortfolio_custom <- SimulatePortfolio(lRaw, dfConfig = dfConfig, seed = 456)