The hardware and bandwidth for this mirror is donated by METANET, the Webhosting and Full Service-Cloud Provider.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]metanet.ch.
In this vignette, you’ll learn about the scope of
psborrow2
and where to find additional information on how
to implement analyses in psborrow2
.
While the randomized controlled trial (RCT) comparing experimental and control arms remains the gold standard for evaluating the efficacy of a novel therapy, one may want to leverage relevant existing external control data to inform the study outcome. External control data can help increase study power and thereby shorten trial duration or reduce the number of subjects needed. However, analysis of external control data can also introduce bias. One method for incorporating external control data to mitigate bias is Bayesian dynamic borrowing (BDB), in which external control data is borrowed to the extent that the external and RCT control arms have similar outcomes. See @viele2014use for a summary.
Implementing BDB is computationally involved and requires Markov
chain Monte Carlo (MCMC) sampling methods, which in turn may require
knowledge of MCMC sampling software. To overcome these technical
barriers and we developed psborrow2
, an R package which
facilitates the use of the MCMC sampling program Stan (via CMD
Stan).
psborrow2
helps the user:
Apply Bayesian dynamic borrowing methods.
psborrow2
has a user-friendly interface for conducting
Bayesian dynamic borrowing analyses using the hierarchical commensurate
prior approach that handles the computationally-difficult MCMC sampling
on behalf of the user.
Conduct simulation studies of Bayesian dynamic borrowing
methods. psborrow2
includes a framework to compare
different trial and borrowing characteristics in a unified way in
simulation studies to inform trial design.
Generate data for simulation studies.
psborrow2
includes a set of functions to generate data for
simulation studies.
psborrow2
supports time-to-event, binary, and continuous
endpoints.
psborrow2
can implement BDB in a scenario wherein a
two-arm RCT is supplemented with external data on the control arm. Three
arms are required to implement BDB in psborrow2
. They
are:
Such scenarios are common in drug development because the comparator arm for a novel therapy is often the standard of care, for which data exists from electronic health care records or from previous phase III registrational trials.
Refer to the “dataset” article for more information on how to implement BDB analyses on your own data: (https://genentech.github.io/psborrow2/articles/dataset.html)[https://genentech.github.io/psborrow2/articles/dataset.html]
Refer to the “simulation study” article for more information on how to create a simulation study involving BDB and other innovative trial designs: https://genentech.github.io/psborrow2/articles/simulation_study.html
Refer to the “data generation” article for more information on how to generate data for simulation studies: https://genentech.github.io/psborrow2/articles/data_simulation.html
Please refer to https://genentech.github.io/psborrow2/articles/index.html
for additional articles on psborrow2
functionality.
cmdstanr
cmdstanr
is highly recommended for use with
psborrow2
. To install cmdstanr
, follow the
instructions outlined by the cmdstanr
documentation or use:
install.packages("cmdstanr", repos = c("https://mc-stan.org/r-packages/", getOption("repos")))
These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.