Repository files navigation

psborrow2 is an R package that for conducting Bayesian dynamic borrowinganalyses and simulation studies (Lewis et al 2019, Viele et al 2014)psborrow2 helps the user:

1. Apply Bayesian dynamic borrowing methods.psborrow2 has a user-friendly interface forconducting Bayesian dynamic borrowing analyses using the hierarchical commensurate prior approachthat handles the computationally-difficult MCMC samplingon behalf of the user.

2. Conduct simulation studies of Bayesian dynamic borrowing methods.psborrow2 includes aframework to compare different trial and borrowing characteristics in a unified wayin simulation studies to inform trial design.

3. Generate data for simulation studies.psborrow2 includes a set of functions to generatedata for simulation studies.

You can install the latest version ofpsborrow2 on CRAN with:

install.packages('psborrow2')

or you can install the development version with:

remotes::install_github("Genentech/psborrow2")

Please note thatcmdstanr is highly recommended, but will not be installed by default when installingpsborrow2.To installcmdstanr, follow the instructions outlined by thecmdstanr documentation or use:

install.packages("cmdstanr",repos= c("https://stan-dev.r-universe.dev", getOption("repos")))

To learn how to use thepsborrow2 R package, refer to thepackage website (https://genentech.github.io/psborrow2/).

psborrow2 is the successor topsborrow.psborrowis still freely available onCRAN with thesame validated functionality; however, the package is not actively developed.Major updates inpsborrow2 include:

• New, more flexible user interface
• New MCMC software (STAN)
• Expanded functionality (e.g., more outcomes, more flexibility in priors, more flexibility in data generation, etc.)

The namepsborrow combines propensity scoring (ps) and Bayesian dynamicborrowing. As one might expect, bothpsborrow andpsborrow2 can be used to combine dynamicborrowing and propensity-score adjustment/weighting methods.

Lewis CJ, Sarkar S, Zhu J, Carlin BP. Borrowing from historical control datain cancer drug development: a cautionary tale and practical guidelines.Statistics in biopharmaceutical research. 2019 Jan 2;11(1):67-78.

Viele K, Berry S, Neuenschwander B, Amzal B, Chen F, Enas N, Hobbs B,Ibrahim JG, Kinnersley N, Lindborg S, Micallef S. Use of historical controldata for assessing treatment effects in clinical trials. Pharmaceuticalstatistics. 2014 Jan;13(1):41-54.