Population PK/PD Modeling and Simulation

An applied course for building, evaluating, and using population PK/PD models in R

Learn population PK/PD modeling with nlmixr2, rxode2, diagnostics, covariate modeling, simulation, and reproducible workflows.

Welcome

Population PK/PD Modeling and Simulation is the applied flagship course in the AEAcademy PMx curriculum.

This course teaches you how to move from pharmacometric data to population models, diagnostics, PK/PD interpretation, simulation, and model-informed decisions using modern open-source R workflows.

The focus is not only on running nlmixr2.

The focus is on understanding the full modeling workflow:

Data

↓

Model

↓

Evaluation

↓

Simulation

↓

Decision

You will learn how to build models, evaluate assumptions, interpret outputs, and use simulations responsibly.

Get Access

Enroll — Population PK/PD Modeling and Simulation Try Free Previews First

Free introductory modules are available with a free AEAcademy Member account — no credit card required.

Who This Course Is For

This course is designed for:

pharmacometricians and clinical pharmacologists
PK/PD and translational scientists
PharmD, MS, and PhD trainees
quantitative scientists entering population modeling
analysts moving from NCA or exploratory PK analysis into model-based workflows
industry scientists who want reproducible open-source R workflows

Learning Objectives

By the end of this course, you will be able to:

prepare analysis-ready datasets for population modeling
visualize PK and PK/PD data before modeling
build population PK models using nlmixr2
interpret fixed effects, random effects, and residual error
understand FOCEi and SAEM at a practical level
evaluate models using diagnostics, residuals, VPCs, and qualification thinking
explore and interpret covariate relationships
introduce direct and delayed PK/PD response models
fit a joint PK/PD model with multiple endpoints
simulate typical subjects, virtual populations, and dose scenarios
simulate from fitted models using estimated variability
communicate model results and assumptions clearly

Recommended Preparation

You do not need to be an expert modeler before starting.

You should be comfortable with:

basic PK concepts such as clearance, volume, half-life, absorption, and exposure
basic R syntax
dplyr and ggplot2
reading concentration-time profiles
basic NCA concepts

Recommended AEAcademy preparation:

Foundations of Pharmacometrics
Computational Foundations for Pharmacometrics
Noncompartmental Analysis, if you want stronger exposure-analysis context

Software Used

The main modeling tools are:

Code

library(nlmixr2)
library(rxode2)

Other packages used throughout the course include:

Code

library(tidyverse)
library(nlmixr2data)
library(ggPMX)
library(nlmixr2plot)

You do not need to know all of these packages before starting.

They are introduced as they become useful in the workflow.

Dataset Strategy

To keep the course coherent, we use a small number of recurring datasets.

Theophylline

Theophylline is the main population PK teaching dataset.

It is used for:

exploratory PK visualization
data preparation
structural PK modeling
first nlmixr2 model fits
variability modeling
residual error modeling
diagnostics
covariate modeling
simulation

Warfarin

Warfarin is used as the introductory PK/PD example.

It introduces:

multiple endpoints
PK and PD observations
delayed response
indirect response modeling
ODE implementation in nlmixr2
joint PK/PD diagnostics

Course Structure

The course follows the applied modeling workflow from orientation to simulation.

Module 1: Orientation to Population PK/PD Modeling

Learn the big picture.

Topics include:

what population modeling is
individual vs population thinking
fixed effects and random effects
structural and statistical model components
why mixed-effects modeling matters
overview of the nlmixr2 workflow

Module 2: Data Preparation and Exploratory PK Visualization

Prepare and inspect modeling-ready data.

Topics include:

observation and dosing records
ID, TIME, DV, AMT, EVID, and MDV
analysis-ready datasets
individual concentration-time profiles
dose-stratified visualization
early QC before modeling

Module 3: Structural PK Models

Build structural PK intuition.

Topics include:

one-compartment oral models
absorption, clearance, and volume
differential equation thinking
CL/V parameterization
simulating structural profiles
naive pooled model intuition

Module 4: Estimation and Building Population PK Models

Fit the first population PK models.

Topics include:

writing nlmixr2 model functions
initial estimates
FOCEi estimation
maximum likelihood intuition
random effects
residual error models
model output interpretation

Module 5: Variability and Random Effects

Understand variability in population models.

Topics include:

interindividual variability
ETA interpretation
residual variability
additive, proportional, and exponential random effects
random-effect covariance
variability interpretation

Module 6: Covariate Modeling

Use patient characteristics to explain variability.

Topics include:

continuous and categorical covariates
covariate centering
weight effects and allometric thinking
covariate model structures
clinical vs statistical interpretation
avoiding overfitting

Module 7: Model Diagnostics

Evaluate model adequacy.

Topics include:

why diagnostics matter
goodness-of-fit plots
individual prediction plots
residual diagnostics
visual predictive checks
parameter uncertainty
model qualification

Module 8: PK/PD Modeling

Extend population PK models into response modeling.

Topics include:

PK vs PD
multiple endpoints
direct effect models
Emax and sigmoid Emax models
delayed response
ODE intuition
indirect response models
joint warfarin PK/PD modeling

Module 9: Simulation and Model-Based Decisions

Use models to explore scenarios.

Topics include:

why simulation matters
individual vs population simulation
virtual subjects
dose scenario exploration
simulation from fitted models
population predictions using fitted variability
interpreting simulation results
communicating assumptions and limitations

Course Philosophy

This course is built around three principles.

1. Modeling is a workflow

A model is not only a code block.

A model sits inside a workflow that includes data understanding, assumptions, estimation, diagnostics, simulation, and communication.

2. Diagnostics are evidence

Plots and diagnostics are not decorations.

They help determine whether a model is useful, limited, biased, or misleading.

3. Simulation connects models to decisions

Once a model is adequate for its purpose, simulation lets us ask practical questions about dose, variability, exposure, response, and uncertainty.

What Makes This Course Different

This course is designed to teach the way applied pharmacometricians think.

You will repeatedly practice:

checking data structure before modeling
starting with simple models
interpreting model components biologically
distinguishing fit, precision, accuracy, and qualification
diagnosing models before trusting them
simulating scenarios only after understanding assumptions
communicating conclusions clearly

How to Use This Course

Each lesson follows a consistent structure:

Objectives

↓

Key Ideas

↓

Worked Examples

↓

Interpretation

↓

Common Mistakes

↓

Practice

↓

Solutions

You are encouraged to:

run the code
inspect intermediate objects
modify examples
compare your outputs to the rendered lesson
focus on interpretation, not just execution

Certificate of Completion

A certificate of completion is available after completing the course.

The certificate reflects completion of the applied Population PK/PD Modeling and Simulation course.

Suggested Next Steps

After this course, natural next topics include:

advanced pharmacometrics
advanced PK/PD response models
TMDD
Bayesian population modeling
clinical trial simulation
PBPK and QSP
regulatory and MIDD communication

Get Started

Start with Module 1: Orientation to Population PK/PD Modeling.

Build the mental model first.

Then move step by step toward fitting, diagnosing, simulating, and interpreting population PK/PD models.

--- title: "Population PK/PD Modeling and Simulation" subtitle: "An applied course for building, evaluating, and using population PK/PD models in R" description: "Learn population PK/PD modeling with nlmixr2, rxode2, diagnostics, covariate modeling, simulation, and reproducible workflows." format: html: toc: true toc-depth: 2 number-sections: false code-fold: true code-tools: true execute: echo: true warning: false message: false --- ## Welcome **Population PK/PD Modeling and Simulation** is the applied flagship course in the AEAcademy PMx curriculum. This course teaches you how to move from pharmacometric data to population models, diagnostics, PK/PD interpretation, simulation, and model-informed decisions using modern open-source R workflows. The focus is not only on running `nlmixr2`. The focus is on understanding the full modeling workflow: ```text Data ↓ Model ↓ Evaluation ↓ Simulation ↓ Decision ``` You will learn how to build models, evaluate assumptions, interpret outputs, and use simulations responsibly. --- ## Get Access <div style="display: flex; gap: 0.75rem; flex-wrap: wrap; margin-top: 0.75rem;"> <a href="#" data-ms-modal="signup" data-ms-price:add="prc_population-pkpd-lifetime-access-ggn0fj1" class="btn-primary" > Enroll — Population PK/PD Modeling and Simulation </a> <a href="/free/" class="btn-outline">Try Free Previews First</a> </div> <p style="margin-top: 0.65rem; font-size: 0.9rem; color: #6b7280;">Free introductory modules are available with a free AEAcademy Member account — no credit card required.</p> --- ## Who This Course Is For This course is designed for: - pharmacometricians and clinical pharmacologists - PK/PD and translational scientists - PharmD, MS, and PhD trainees - quantitative scientists entering population modeling - analysts moving from NCA or exploratory PK analysis into model-based workflows - industry scientists who want reproducible open-source R workflows --- ## Learning Objectives By the end of this course, you will be able to: - prepare analysis-ready datasets for population modeling - visualize PK and PK/PD data before modeling - build population PK models using `nlmixr2` - interpret fixed effects, random effects, and residual error - understand FOCEi and SAEM at a practical level - evaluate models using diagnostics, residuals, VPCs, and qualification thinking - explore and interpret covariate relationships - introduce direct and delayed PK/PD response models - fit a joint PK/PD model with multiple endpoints - simulate typical subjects, virtual populations, and dose scenarios - simulate from fitted models using estimated variability - communicate model results and assumptions clearly --- ## Recommended Preparation You do not need to be an expert modeler before starting. You should be comfortable with: - basic PK concepts such as clearance, volume, half-life, absorption, and exposure - basic R syntax - `dplyr` and `ggplot2` - reading concentration-time profiles - basic NCA concepts Recommended AEAcademy preparation: - **Foundations of Pharmacometrics** - **Computational Foundations for Pharmacometrics** - **Noncompartmental Analysis**, if you want stronger exposure-analysis context --- ## Software Used The main modeling tools are: ```{r} library(nlmixr2) library(rxode2) ``` Other packages used throughout the course include: ```{r} library(tidyverse) library(nlmixr2data) library(ggPMX) library(nlmixr2plot) ``` You do not need to know all of these packages before starting. They are introduced as they become useful in the workflow. --- ## Dataset Strategy To keep the course coherent, we use a small number of recurring datasets. ### Theophylline Theophylline is the main population PK teaching dataset. It is used for: - exploratory PK visualization - data preparation - structural PK modeling - first `nlmixr2` model fits - variability modeling - residual error modeling - diagnostics - covariate modeling - simulation ### Warfarin Warfarin is used as the introductory PK/PD example. It introduces: - multiple endpoints - PK and PD observations - delayed response - indirect response modeling - ODE implementation in `nlmixr2` - joint PK/PD diagnostics --- ## Course Structure The course follows the applied modeling workflow from orientation to simulation. ### Module 1: Orientation to Population PK/PD Modeling Learn the big picture. Topics include: - what population modeling is - individual vs population thinking - fixed effects and random effects - structural and statistical model components - why mixed-effects modeling matters - overview of the `nlmixr2` workflow --- ### Module 2: Data Preparation and Exploratory PK Visualization Prepare and inspect modeling-ready data. Topics include: - observation and dosing records - ID, TIME, DV, AMT, EVID, and MDV - analysis-ready datasets - individual concentration-time profiles - dose-stratified visualization - early QC before modeling --- ### Module 3: Structural PK Models Build structural PK intuition. Topics include: - one-compartment oral models - absorption, clearance, and volume - differential equation thinking - CL/V parameterization - simulating structural profiles - naive pooled model intuition --- ### Module 4: Estimation and Building Population PK Models Fit the first population PK models. Topics include: - writing `nlmixr2` model functions - initial estimates - FOCEi estimation - maximum likelihood intuition - random effects - residual error models - model output interpretation --- ### Module 5: Variability and Random Effects Understand variability in population models. Topics include: - interindividual variability - ETA interpretation - residual variability - additive, proportional, and exponential random effects - random-effect covariance - variability interpretation --- ### Module 6: Covariate Modeling Use patient characteristics to explain variability. Topics include: - continuous and categorical covariates - covariate centering - weight effects and allometric thinking - covariate model structures - clinical vs statistical interpretation - avoiding overfitting --- ### Module 7: Model Diagnostics Evaluate model adequacy. Topics include: - why diagnostics matter - goodness-of-fit plots - individual prediction plots - residual diagnostics - visual predictive checks - parameter uncertainty - model qualification --- ### Module 8: PK/PD Modeling Extend population PK models into response modeling. Topics include: - PK vs PD - multiple endpoints - direct effect models - Emax and sigmoid Emax models - delayed response - ODE intuition - indirect response models - joint warfarin PK/PD modeling --- ### Module 9: Simulation and Model-Based Decisions Use models to explore scenarios. Topics include: - why simulation matters - individual vs population simulation - virtual subjects - dose scenario exploration - simulation from fitted models - population predictions using fitted variability - interpreting simulation results - communicating assumptions and limitations --- ## Course Philosophy This course is built around three principles. ### 1. Modeling is a workflow A model is not only a code block. A model sits inside a workflow that includes data understanding, assumptions, estimation, diagnostics, simulation, and communication. --- ### 2. Diagnostics are evidence Plots and diagnostics are not decorations. They help determine whether a model is useful, limited, biased, or misleading. --- ### 3. Simulation connects models to decisions Once a model is adequate for its purpose, simulation lets us ask practical questions about dose, variability, exposure, response, and uncertainty. --- ## What Makes This Course Different This course is designed to teach the way applied pharmacometricians think. You will repeatedly practice: - checking data structure before modeling - starting with simple models - interpreting model components biologically - distinguishing fit, precision, accuracy, and qualification - diagnosing models before trusting them - simulating scenarios only after understanding assumptions - communicating conclusions clearly --- ## How to Use This Course Each lesson follows a consistent structure: ```text Objectives ↓ Key Ideas ↓ Worked Examples ↓ Interpretation ↓ Common Mistakes ↓ Practice ↓ Solutions ``` You are encouraged to: - run the code - inspect intermediate objects - modify examples - compare your outputs to the rendered lesson - focus on interpretation, not just execution --- ## Certificate of Completion A certificate of completion is available after completing the course. The certificate reflects completion of the applied Population PK/PD Modeling and Simulation course. --- ## Suggested Next Steps After this course, natural next topics include: - advanced pharmacometrics - advanced PK/PD response models - TMDD - Bayesian population modeling - clinical trial simulation - PBPK and QSP - regulatory and MIDD communication --- ## Get Started Start with **Module 1: Orientation to Population PK/PD Modeling**. Build the mental model first. Then move step by step toward fitting, diagnosing, simulating, and interpreting population PK/PD models.