In this episode I discuss R Shiny and how it can be used for building clinical pharmacology tools. I provide an overview of the technology, suggest a few example use cases, and then walk through a specific practical example of predicting AUC and Cmax for future doses from observed data. I end with a discussion of the benefits and challenges of using R Shiny for clinical pharmacology tools.
Links discussed in the show:
Copyright Teuscher Solutions LLC
All Rights Reserved
In this episode I discuss noncompartmental analysis. I categorize PK parameters calculated by NCA methods as either Important, Useful, or Questionable. I also share my thoughts on how to report PK parameters calculated using NCA methods in nonclinical and clinical reports. I want to hear your thoughts about this episode. Do you agree or disagree with my categorization of PK parameters? Why? Use the links below to let me know.
Links discussed in the show:
Copyright Teuscher Solutions LLC
All Rights Reserved
In this episode I review the process I follow for fitting a base pharmacokinetic (PK) model. I talk from the perspective of an individual PK model, but include some differences associated with population PK models. I go over exploratory data analysis, getting initial estimates, and how to choose between different base models.
Links discussed in the show:
Copyright Teuscher Solutions LLC
All Rights Reserved
In this episode I discuss different algorithms used for pharmacometrics modeling. I describe difference between maximum likelihood and expectation maximization methods. I review the FO and FOCEI maximum likelihood algorithms. I then review SAEM, IMP, and QRPEM expectation maximization algorithms that are available. I conclude with an brief explanation of the difference between parameter estimation and parameter uncertainty.
Links discussed in the show:
Copyright Teuscher Solutions LLC
All Rights Reserved
In this episode I discuss my experience as an independent consultant in the clinical pharmacology and pharmacometrics space. I talk about how I got from high school to become a consultant. I touched on the many different activities I do as an independent consultant. I shared a bit about how I handle compensation with clients and how I pay myself. And I ended with some brief comments about getting customers.
Links discussed in the show:
Copyright Teuscher Solutions LLC
All Rights Reserved
In this episode I discuss using nlmixr for population PK modeling in R. I give an overview of the software package, describe the installation process, and walk through a simple example. Overall, nlmixr is a nice package that can be used for most modeling activities you encounter. It eliminates the need for a license since it is open-source. There are some challenges associated with using R for modeling with large datasets and long-running models. But once those are overcome, this could be a great package for all-around use.
Links discussed in the show:
Copyright Teuscher Solutions LLC
All Rights Reserved
In this episode I discuss power calculations using the R package PowerTOST. I gave an introduction to power calculations and the statistical premise. I reviewed bioequivalence study designs that are commonly used for generic drug development, food effect evaluation, and drug-drug interaction studies.
Links discussed in the show:
Copyright Teuscher Solutions LLC
All Rights Reserved
In this episode I share a couple short clips from my conversation with Chad Briscoe and Greg Austin from BioTalk Unzipped (www.biotalkunzipped.com). We chat about a variety of topics including artificial intelligence and challenges with cell therapy development. You can hear the entire conversation on the BioTalk Unizipped podcast or YouTube channel in January. For now, enjoy these clips from the upcoming show!
Links discussed in the show:
In this episode I discuss toxicokinetic analysis. This is the analysis of exposure data to support animal toxicology studies. I review the purpose of toxicology studies and how exposure metrics are used. Then I discuss different blood sampling schemes, and how to calculate exposure parameters for both full profiles and mean profiles.
Links discussed in the show:
Copyright Teuscher Solutions LLC
All Rights Reserved
This episode is a discussion of steady state pharmacokinetics. Steady state occurs when the amount of drug input is equivalent to the amount of drug elimination in a dosing interval. I discuss some key PK parameters at steady state and how to use them. I also discuss different methods of confirming that you have reached steady state.
Links discussed in the show:
Copyright Teuscher Solutions LLC
All Rights Reserved
In this episode I talk about dose proportionality, which is a surrogate for determining if we have linear clearance across a range of dose levels. I discuss 3 different methods that people use to determine dose proportionality along with a recommendation to use the power model method.
Links discussed in the show:
This episode is a discussion of covariate search methods. I give definitions of covariates and predictors, and then I describe 3 different covariate search methods. For each method I describe the pros and cons associated with each one, including bias and time limitations.
Links discussed in the show:
Copyright Teuscher Solutions LLC
All Rights Reserved
In this episode I describe stimulatory indirect response PK-PD models. I describe how they work, how to set up the dataset for NONMEM, and how to code the model in NONMEM. This is the last of 4 episodes on different PK-PD models.
Links discussed in the show:
Copyright Teuscher Solutions LLC
All Rights Reserved
In this episode I describe inhibitory indirect response PK-PD models. I describe how they work, how to set up the dataset for NONMEM, and how to code the model in NONMEM. This is the third of 4 episodes on different PK-PD models.
Links discussed in the show:
Copyright Teuscher Solutions LLC
All Rights Reserved
In this episode I describe effect compartment PK-PD models. I describe how they work, a method of creating exploratory plots, how to set up the dataset for NONMEM, and how to code the model in NONMEM. This is the 2nd of 4 episodes on different PK-PD models.
Links discussed in the show:
Copyright Teuscher Solutions LLC
All Rights Reserved
In this episode I describe direct effect PK-PD models. I describe how they work, a method of creating exploratory plots to identify direct effect models, how to set up the dataset for NONMEM, and how to code the model in NONMEM. This is the first of 4 episodes on different PK-PD models.
Links discussed in the show:
Copyright Teuscher Solutions LLC
All Rights Reserved
In this episode I discuss different methods of presenting pharmacokinetic and pharmacometrics results. I spoke about using PowerPoint and R Markdown derived HTML files. While I prefer HTML files in nearly all cases, I appreciate that sometimes PowerPoint format is required by some companies. I included multiple suggestions for both PowerPoint and HTML files that can help make the information presentation more effective.
Links discussed in the show:
Copyright Teuscher Solutions LLC
All Rights Reserved
This episode covers the terminal elimination rate constant calculated using non-compartmental analysis techniques. I reviewed how the rate constant is used to estimate half-life and extrapolate AUC to infinity, reviewed the methods for calculating the terminal rate constant value, and some important points about sample selection. Finally, I gave my reasons that we should ultimately stop talking about this and fretting over the method of calculation. Instead, let’s design better PK experiments with more optimal sampling to capture decline of the drugs we are studying.
Links discussed in the show:
Copyright Teuscher Solutions LLC
All Rights Reserved
This episode covered parallelization in modeling analyses. I spoke about across model parallelization and within model parallelization. I recommended some approaches for choosing how many cores to use in parallelization and also discussed different modeling computer systems you can use.
Links discussed in the show:
Copyright Teuscher Solutions LLC
All Rights Reserved
Today I covered 2 different topics. First, I spoke about residual error models when modeling. I discussed additive, proportional, and power models. I also reviewed combined models and when they can be helpful. Then I spoke about modeling report submission packages. I explained what I put in those packages, how they are used, and some potential changes coming in the future.
Links discussed in the show:
Copyright Teuscher Solutions LLC
All Rights Reserved
