Tipping Point Analysis for Bayesian Dynamic Borrowing

Tipping point analysis for clinical trials that employ Bayesian dynamic borrowing via robust meta-analytic predictive (MAP) priors. Further functions facilitate expert elicitation of a primary weight of the informative component of the robust MAP prior and computation of operating characteristics. Intended use is the planning, analysis and interpretation of extrapolation studies in pediatric drug development, but applicability is generally wider.

tipmap

The tipmap-package facilitates the planning and analysis of partial extrapolation studies in pediatric drug development. It provides an implementation of a Bayesian tipping point approach that can be used in analyses based on robust meta-analytic predictive (MAP) priors. Further functions facilitate expert elicitation of a primary (pre-specified) weight of the informative component of the MAP prior and the computation of operating characteristics.

Installation

CRAN

You can install the current stable version from CRAN with:

install.packages("tipmap")

GitHub

You can install the current development version from GitHub with:

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

Getting started

Load the package:

library(tipmap)

The prior data (collected in the source population):

prior_data <- create_prior_data(
  n_total = c(160, 240, 320),
  est = c(1.23, 1.40, 1.51),
  se = c(0.4, 0.36, 0.31)
)

The data from the new trial (collected in the target population):

ped_trial <- create_new_trial_data(
  n_total = 30, 
  est = 1.27, 
  se = 0.95
)

Derivation of the meta-analytic predictive (MAP) prior:

uisd <- sqrt(ped_trial["n_total"]) * ped_trial["se"]
g_map <-
  RBesT::gMAP(
    formula = cbind(est, se) ~ 1 | study_label,
    data = prior_data,
    family = gaussian,
    weights = n_total,
    tau.dist = "HalfNormal",
    tau.prior = cbind(0, uisd / 16),
    beta.prior = cbind(0, uisd)
  )

map_prior <- RBesT::automixfit(
  sample = g_map,
  Nc = seq(1, 4),
  k = 6,
  thresh = -Inf
)

Computing the posterior distribution for weights of the informative component of the MAP prior ranging from 0 to 1:

posterior <- create_posterior_data(
  map_prior = map_prior,
  new_trial_data = ped_trial,
  sigma = uisd)

Creating data for a tipping point analysis (tipping point plot):

tipmap_data <- create_tipmap_data(
  new_trial_data = ped_trial,
  posterior = posterior,
  map_prior = map_prior)

Create tipping point plot:

tipmap_plot(tipmap_data = tipmap_data)

Get tipping points:

get_tipping_points(
  tipmap_data, 
  quantile = c(0.025, 0.05, 0.1, 0.2), 
  null_effect = 0.1)

Citing tipmap

To cite tipmap in publications please use: Christian Stock and Morten Dreher (2023). tipmap: Tipping Point Analysis for Bayesian Dynamic Borrowing. R package version 0.5.2. URL: https://CRAN.R-project.org/package=tipmap