Sample Size Estimation for Bio-Equivalence Trials Through Simulation

Sample size estimation for bio-equivalence trials is supported through a simulation-based approach that extends the Two One-Sided Tests (TOST) procedure. The methodology provides flexibility in hypothesis testing, accommodates multiple treatment comparisons, and accounts for correlated endpoints. Users can model complex trial scenarios, including parallel and crossover designs, intra-subject variability, and different equivalence margins. Monte Carlo simulations enable accurate estimation of power and type I error rates, ensuring well-calibrated study designs. The statistical framework builds on established methods for equivalence testing and multiple hypothesis testing in bio-equivalence studies, as described in Schuirmann (1987) , Mielke et al. (2018) , Shieh (2022) , and Sozu et al. (2015) . Comprehensive documentation and vignettes guide users through implementation and interpretation of results.

SimTOST

SimTOST is an R package specifically designed for bioequivalence studies, providing simulation-based sample size estimation for the Two One-Sided Tests (TOST) procedure. It offers flexible options to handle complex study designs, including trials with multiple correlated primary endpoints, multiple hypotheses, and treatment arms. Multiplicity adjustments include Bonferroni (adjust = "bon"), Sidak (adjust = "sid"), the weak k-adjustment (adjust = "k"), Mielke’s strong k-out-of-m adjustment (adjust = "t"), and sequential hierarchical testing (adjust = "seq"). By incorporating correlations between endpoints, SimTOST ensures accurate and robust planning of bioequivalence trials, making it a powerful tool for studies with intricate requirements.

Installation

SimTOST is available on CRAN and can be installed by running the following code.

install.packages("SimTOST")

You can also install the development version of SimTOST from GitHub with:

# install.packages("devtools")
devtools::install_github("smartdata-analysis-and-statistics/SimTOST")

Vignettes

The main features of this package is sampleSize function which can be used to calculate sample size for individual and multiple endpoints. Various worked examples are available as vignettes, with an introduction provided at Introduction. The Introduction vignette describes the available multiplicity adjustments, including Mielke’s strong k-out-of-m adjustment (adjust = "t").