carts (version 0.1.0)

carts-package: carts: Simulation-Based Assessment of Covariate Adjustment in Randomized Trials

Description

Monte Carlo simulation framework for different randomized clinical trial designs with a special emphasis on estimators based on covariate adjustment. The package implements regression-based covariate adjustment (Rosenblum & van der Laan (2010) tools:::Rd_expr_doi("10.2202/1557-4679.1138")) and a one-step estimator (Van Lancker et al (2024) tools:::Rd_expr_doi("10.48550/arXiv.2404.11150")) for trials with continuous, binary and count outcomes. The estimation of the minimum sample-size required to reach a specified statistical power for a given estimator uses bisection to find an initial rough estimate, followed by stochastic approximation (Robbins-Monro (1951) tools:::Rd_expr_doi("10.1214/aoms/1177729586")) to improve the estimate, and finally, a grid search to refine the estimate in the neighborhood of the current best solution.

Author

Maintainer: Benedikt Sommer

Authors:

• Klaus K. Holst

• Foroogh Shamsi

Other contributors:

• Novo Nordisk A/S [copyright holder]

See Also

Useful links:

• https://novonordisk-opensource.github.io/carts/

• https://github.com/NovoNordisk-OpenSource/carts

• Report bugs at https://github.com/NovoNordisk-OpenSource/carts/issues

Examples

# \donttest{
trial <- Trial$new(
  covariates = \(n) data.frame(a = rbinom(n, 1, 0.5), x = rnorm(n)),
  outcome = setargs(outcome_count, par = c(1, 0.5, 1), overdispersion = 0.7)
)

trial$estimators(
  unadjusted = est_glm(family = "poisson"),
  adjusted = est_glm(family = "poisson", covariates = "x")
)

trial$run(n = 200, R = 100)
trial$summary()
# }