Toy Longitudinal Survival Data

A toy person-period dataset for demonstrating time-varying quantile g-computation with a five-component exposure mixture.

Usage

data(toy_data)

Format

A data.frame with 5,584 rows and 28 variables:

UniqID

Simulated participant identifier.

BC, NIT, SO4, NH4, OM

Continuous time-varying mixture components in arbitrary simulated concentration units.

Time, TimeInn, TimeOut

Follow-up interval variables covering times 0 through 12.

status

Interval event indicator.

sex, visible_minority, IndigenousIdentity, marsth, immigration_status, landed_migration, Occupation, employment, Education

Simulated demographic and socioeconomic covariates.

age, income_inadequacy

Covariates that vary over follow-up in this toy dataset.

CanadianRegion, Urban_form, CSize, dependency, deprivation, ethnicconcentration, instability

Time-varying categorical covariates.

Details

The dataset contains 500 simulated individuals followed for 1 to 13 years. Exactly 100 participants die during follow-up, corresponding to 20 percent cumulative mortality. It is provided solely for examples and testing and contains no real participant records. The exposure variables remain continuous in the public dataset and are converted to quartiles internally by tvcQGComp_survival(). Precomputed lag columns are intentionally omitted; set auto_history = TRUE in the configuration to generate required lag histories. An RDS copy is installed at system.file("extdata", "toy_data_500.RDS", package = "tvcQGComp").

Source

Simulated for the tvcQGComp methods project.