Sanaya Shroff1, Boris Grinshpun2, Philip Delff1
1Vertex Pharmaceuticals Incorporated, 2Certara, Inc
Introduction/Objectives The new R package, NMsim [1,2], provides the capability to perform simulations of NONMEM models, including with parameter uncertainty, directly from R without the need for model reimplementation. Using only paths to control stream(s) and a simulation data set provided by the user, NMsim can incorporate model parameter uncertainty through several different methods, from a covariance step or bootstrap results, allowing calculations of confidence intervals or covariate effect evaluation, e.g. forest plots. Clinical Trial Simulation (CTS) is an application of pharmacokinetic (PK) or pharmacokinetic/pharmacodynamic (PK/PD) modeling to assist in the design of clinical trials. NMsim can aid in effective CTS and trial design by simple, R-based, simultaneous multi-model simulation in NONMEM. We aim to show these features through examples that are highly relevant in pharmacometrics. Methods: Using simple adjustments to function calls NMsim can: * Incorporate model parameter uncertainty from bootstrap results for an example typical subject * Incorporate model parameter uncertainty from a covariance step for an example typical subject * Conduct a full CTS using external methods like simpar. * Summarize and compare simulation results from different methods * Use simulation with parameter uncertainty to build a forest plot and evaluate covariate effects (e.g. using the coveffectsplot R package) The seamless workflow of NMsim allows code to be readily applicable to most NONMEM models, the user interface entirely within R. Results: NMsim provides a well-designed, model-agnostic interface to rapidly conduct simulations with parameter uncertainty, including full clinical trial simulations using a variety of different approaches. The results illustrate how easily a model-based approach to questions commonly posed in drug discovery and development, such as covariate effect evaluation, can be performed through simulation. Conclusions: The features of NMsim provide a powerful tool to easily simulate with parameter uncertainty, e.g. simulations for forest plots, or to conduct clinical trial simulation. Using NMsim, one can simply and effectively conduct simulation with parameter uncertainty based on both non-parametric bootstrap, and parametric methods such as NONMEM’s native NWPRI. Externally-sampled parameter values can also easily be fed into NMsim and NMsim even includes automation of parameter sampling with simpar by supplying only the control stream path and number of desired samples. We hope these examples will aid pharmacometricians in making simulation with uncertainty and clinical trial simulations more easily accessible by automating analyses, and can help support the design of clinical studies during the drug development process.
[1] Delff, Philip. 2024. NMsim: An R package that can simulate Nonmem models. https://philipdelff.github.io/NMsim [2] Delff, Philip. 2024. NMsim: Seamless Nonmem Simulation Platform. https://cran.r-project.org/web/packages/NMsim
Reference: PAGE 33 (2025) Abstr 11456 [www.page-meeting.org/?abstract=11456]
Poster: Methodology - New Modelling Approaches