The goal of
nlmixr, or more accurately
nlmixr2, is to support easy and robust nonlinear mixed effects models (NLMEMs) in R.
NLMEMs are used to help identify and explain the relationships between drug exposure, safety, and efficacy and the differences among population subgroups. Most often, they are built using longitudinal PK and pharmacodynamic (PD) data collected during clinical studies. These models characterize the relationships between dose, exposure and biomarker and/or clinical endpoint response over time, variability between individuals and groups, residual variability, and uncertainty.
NLMEM development in the pharmaceutical space is dominated by a small number of proprietary, commercial software tools. Although this kind of approach to software has some advantages, adopting an open-source, open-science paradigm also has benefits - third-party auditing or adjustments are possible, and the precise model-fitting methodology employed can be determined by anyone with the time and energy to review the source code. We see
nlmixr2 being especially useful in being able to integrate into the rich R ecosystem, and it is well suited for use in scripted, literate-programming workflows of the kind flourishing in the R ecosystem by means of packages such as
nlmixr2 2.1.0 was released and I promised to talk about the new features. One of the things that can impact many peoples work-flow is new estimation methods for population-only data. Many people use population-only estimation methods before changing the model to a mixed effect model, so I believe these can be useful for many people trying to find the best model to the data at hand. I will talk about the new ones (and why you may want to use them).Read more