Skip to contents

Bayesian models to estimate causal effects of biological treatments on time-to-event endpoints in clinical trials with principal strata defined by the occurrence of antidrug antibodies.

Scope

  • The package contains functions to fit Bayesian principal stratification models and to perform clinical trial simulations to determine operating characteristics for given scenarios.
  • Two-arm clinical trials of biological therapies are considered
    • with an intercurrent event (determining the principal stratum of interest) that can only occur in the treated arm (such as the development of antidrug antibodies), and
    • with a time-to-event endpoint that is assumed to follow an exponential distribution.
  • Effect estimators are hazard ratios and restricted mean survival times.
  • Potential predictors of the intercurrent event can be taken into account.
  • The models are fitted by Monte Carlo Markov Chain (MCMC) sampling, they are coded in Stan and precompiled.
  • More flexible time-to-event distributions (piecewise-exponential and Weibull) will be considered in future versions of the package.

Principal stratification methodology

  • Principal stratification is an approach to estimate causal effects in partitions of subjects determined by post-treatment events. It was introduced in the biostatistical literature by Frangakis and Rubin (2002).
  • The ICH E9 (R1) addendum on estimands and sensitivity analysis in clinical trials proposed principal stratification as one approach to deal with intercurrent events in clinical trials (International Council for Harmonisation (ICH) (2020)).
  • Principal stratum membership is typically not known with certainty. A Bayesian approach may be particularly suited to deal with this type of uncertainty. Following a proposal by Imbens and Rubin (1997), principal stratum membership can be treated as a latent mixture variable.
  • Motivated by scientific questions arising in clinical trials of biological therapies, in this package the approach by Imbens and Rubin (1997) is adapted to a specific clinical trial setting with a time-to-event endpoint and the intercurrent event only occurring in the treated group.
  • For recent reviews of applications to clinical trials see Lipkovich et al. (2022) and Bornkamp et al. (2021).

References:

Bornkamp, B., Rufibach, K., Lin, J., Liu, Y., Mehrotra, D. V., Roychoudhury, S., Schmidli, H., Shentu, Y., and Wolbers, M. (2021), “Principal stratum strategy: Potential role in drug development,” Pharm Stat, 20, 737–751. https://doi.org/10.1002/pst.2104.
Frangakis, C. E., and Rubin, D. B. (2002), “Principal stratification in causal inference,” Biometrics, 58, 21–29. https://doi.org/10.1111/j.0006-341x.2002.00021.x.
Imbens, G. W., and Rubin, D. B. (1997), “Bayesian Inference for Causal Effects in Randomized Experiments with Noncompliance,” Ann Stat, 25, 305–327. https://doi.org/10.1214/aos/1034276631.
International Council for Harmonisation (ICH) (2020), “ICH E9 (R1) addendum on estimands and sensitivity analysis in clinical trials to the guideline on statistical principles for clinical trials.” https://database.ich.org/sites/default/files/E9-R1_Step4_Guideline_2019_1203.pdf.
Lipkovich, I., Ratitch, B., Qu, Y., Zhang, X., Shan, M., and Mallinckrodt, C. (2022), “Using principal stratification in analysis of clinical trials,” Stat Med, 41, 3837–387. https://doi.org/10.1002/sim.9439.

Installation

The current stable version of the package can be installed from CRAN with:

install.packages("BPrinStratTTE")

The development version of the package can be installed from GitHub with:

if (!require("remotes")) {install.packages("remotes")}
remotes::install_github("Boehringer-Ingelheim/BPrinStratTTE")