rbmiUtils (version 0.3.0)

validate_data: Validate Data Before Imputation

Description

Pre-flight validation of data, variable specification, and intercurrent event data before calling rbmi::draws(). Collects all issues and reports them together in a single error message.

Usage

validate_data(data, vars, data_ice = NULL)

Value

Invisibly returns TRUE if all checks pass. Throws an error with collected messages if any issues are found.

Arguments

data

A data.frame containing the analysis dataset.

vars

A vars object as created by rbmi::set_vars().

data_ice

An optional data.frame of intercurrent events. If provided, must contain columns corresponding to vars$subjid, vars$visit, and vars$strategy. Can be created using prepare_data_ice().

Details

The following checks are performed:

• data is a data.frame

• All columns named in vars exist in data

• subjid, visit, and group columns are factors

• outcome column is numeric

• Covariate columns have no missing values

• Data has one row per subject-visit combination

• If data_ice is provided: correct columns, valid subjects, valid visits, recognised strategies, and at most one row per subject

Recommended Workflow:

1. Call validate_data() to check your data

2. Use prepare_data_ice() to create ICE data if needed

3. Review missingness with summarise_missingness()

4. Proceed with rbmi::draws() for imputation

See Also

• rbmi::draws() which requires validated input data

• prepare_data_ice() to create intercurrent event data from flags

• summarise_missingness() to understand missing data patterns

Examples

library(rbmi)

dat <- data.frame(
  USUBJID = factor(rep(c("S1", "S2", "S3"), each = 3)),
  AVISIT = factor(rep(c("Week 4", "Week 8", "Week 12"), 3),
                  levels = c("Week 4", "Week 8", "Week 12")),
  TRT = factor(rep(c("Placebo", "Drug A", "Drug A"), each = 3)),
  CHG = c(1.1, 2.2, 3.3, 0.5, NA, NA, 1.0, 2.0, NA),
  BASE = rep(c(10, 12, 11), each = 3),
  STRATA = factor(rep(c("A", "B", "A"), each = 3))
)

vars <- set_vars(
  subjid = "USUBJID",
  visit = "AVISIT",
  group = "TRT",
  outcome = "CHG",
  covariates = c("BASE", "STRATA")
)

validate_data(dat, vars)