Table 4.

Covariate, inter-individual variability, and residual error relationships considered during model building

	Mathematical function	Compounds tested
Covariate relationshipa
Continuous covariate
No relationship		Citalopram, DMAG, Escitalopram, Olanzapine, Perphenazine, Risperidone, Ziprasidone
Additive		Citalopram, DMAG, Olanzapine, Perphenazine
Proportional		Citalopram, DMAG, Escitalopram, Olanzapine
Exponential		Citalopram, DMAG, Escitalopram, Perphenazine, Risperidone, Olanzapine, Ziprasidone
Power-law		Citalopram, Escitalopram, Olanzapine, Perphenazine
Michaelis–Menten		DMAG
Discrete
No relationship		Citalopram, DMAG, Escitalopram, Olanzapine, Perphenazine, Risperidone, Ziprasidone
Additive		Citalopram, DMAG, Escitalopram, Olanzapine, Perphenazine, Ziprasidone
Proportional		Perphenazine
Exponential		Risperidone, Perphenazine
Type of inter-individual variabilityb
No relationship		Citalopram, DMAG, Escitalopram, Olanzapine, Perphenazine, Risperidone, Ziprasidone
Exponential		Citalopram, DMAG,Escitalopram, Olanzapine, Perphenazine, Risperidone, Ziprasidone
Type of residual variabilityc
Additive		Citalopram, DMAG, Escitalopram, Olanzapine, Perphenazine, Risperidone, Ziprasidone
Proportional		Citalopram, DMAG, Escitalopram, Olanzapine, Perphenazine, Risperidone, Ziprasidone
Combined additive and proportional		Citalopram, DMAG, Escitalopram, Olanzapine, Perphenazine, Risperidone, Ziprasidone

^aModification of a typical value for the example of clearance of the ith individual by the jth covariate or where x _i,j is the value of the jth covariate for the ith patient, is the median value of the jth covariate, is the category (counting) number of the jth covariate for the ith patient, and is a parameter for functional form n that describe the relationship of the jth covariate

^bInter-individual function forms considered for the example of the clearance of the ith individual. The takes a standard normal distribution with standard deviation

^cResidual variability function forms considered. The variable Y _i,j is the jth observation for the ith patient, F _i,j is the corresponding model prediction, and and take standard normal distributions with variances and , respectively