SOME SIMPLE METHODS FOR ESTIMATING INTRAINDIVIDUAL VARIABILITY IN NONLINEAR MIXED EFFECTS MODELS

Methods are proposed to improve both population and individual inferen ce within the nonlinear random coefficient regression framework by inc orporating possible heterogeneity in the intraindividual variance stru cture. These methods extend existing variance function estimation tech niques by pooling information across individuals. The methods are appr opriate when it is reasonable to assume that there exists a common int raindividual variance structure. Two common situations in which this a ssumption may be reasonable are assay development and calibration and pharmacokinetic modeling. The proposed methods are illustrated using e xamples from both of these fields.