Transformations and Variance Functions in Nonlinear Mixed-Effects Models

• Activity Number: 635
• Type: Invited
• Date/Time: Thursday, August 2, 2012 : 10:30 AM to 12:20 PM
• Sponsor: WNAR
• Author(s): Jose Carlos Pinheiro*+
• Companies: Johnson & Johnson
• Keywords: Heteroscedasticity ; Random effects ; Pharmacokinetics ; Transform both sides

Abstract

Nonlinear mixed-effects (NLME) models typically assume additive normal within-subject errors with constant variance and normal random effects. However, when real data are analyzed, departures from these assumptions are frequently observed, such as non-normality of within-subject errors and/or random effects and heteroscedasticity. In classical regression analysis with independent data, non-normality has been typically dealt with by transforming the response (and sometimes also the model, as in the "transform both sides" approach) and heteroscedasticy has been addressed through either data transformation or the use of variance functions (leading to weighted regression methods). This talk discusses and contrasts the use of transformations and variance functions in the context of NLME models, for the purpose of correcting suspected departures from the model assumptions. Transform both sides strategies for NLME models are described and compared to the use of variance functions, in the context of heteroscedastic data. Real and simulated data from population PK models are used to illustrate the methods described.