• Accueil

Agenda

Publié le 7 May 2009, mise à jour le 4 July 2011


DAY 1 (9h00-18h00):

 Definitions
  • What is a population? An individual?
  • Random variable, distribution of a random variable, parameters
  • Gaussian distribution: unidimensional vs. multidimensional Gaussian, mixture of Gaussians
  • What is a model? A statistical model?


 Estimation
  • Definition of an estimator
  • Maximum likelihood estimator
  • Confidence vs. prediction intervals
  • Hypotheses testing (definition of hypotheses, risk, decision rule)
DAY 2 (9h00-18h00):

 Linear statistical model
  • Definitions (linear vs. nonlinear models, fixed effects, random effects and mixed effects models)
  • Underlying assumptions (independence, normality, homoscedasticity)
  • Estimation (least squares) and hypothesis testing
  • Model diagnostics (residuals, normality graphs and tests)
  • Application to covariate models in pharmacokinetics


 Nonlinear models
  • Definitions
  • Underlying assumptions
  • Estimation methods (criteria and their optimisation: handling NONMEM warnings and error messages)
  • Diagnostics
  • Exercises: application to classical (individual) pharmacokinetic analysis using NONMEM
DAY3 (9h00-18h00):

 Nonlinear mixed effects models (hierarchical population models)
  • Definitions and sources of variability
  • Underlying assumptions
  • Likelihood & estimation methods: everything you always wanted to know about two-stage, FO, FOCE, FOCE-I, SAEM... methods but were afraid to ask
  • What are empirical Bayes estimates (EBEs)? How to handle eta- or epsilon-shrinkage?
  • Monte-Carlo simulations
  • Exercises in population pharmacokinetics using NONMEM
DAY4 (9h00-18h00):

 Nonlinear mixed effects models (ctd’)
  • Standard diagnostics (IPRED, PRED, WRES, IWRES)
  • Model building strategy, model selection criteria (AIC, BIC), Likelihood Ratio Test, Wald Test
  • Model validation (internal vs. external, cross-validation, visual predictive checks, NPDE)
  • Modeling covariate effects: a clear view on covariate selection
  • Exercises in population pharmacokinetics using NONMEM
DAY5 (9h00-17h00):

 Nonlinear mixed effects models (ctd’)
  • Modeling inter-occasion variability
  • Exercises in population pharmacokinetics using NONMEM
DAY6 (9h00-12h00):
  • FINAL EXAM: data analysis on a case study

calle calle