. 2014 Jul;58(7):4094–4102. doi: 10.1128/AAC.02664-14

TABLE 2.

Evaluation of the predictive performance of piperacillin and tazobactam population models

Model	No. of variables	Log likelihood	χ² compared to linear model	No. of cycles to convergence	AIC	Linear regression of observed-predicted values for each patient
Model	No. of variables	Log likelihood	χ² compared to linear model	No. of cycles to convergence	AIC	R²	Intercept	Slope
Linear	4	−2,973.5	N/A	1,310	2,988	0.931	2.43	0.903
Michaelis-Menten	5	−2,892.5	2.26E−19	1,490	2,914	0.933	3.35	0.921
Parallel linear/MM	6	−2,894.5	7.00E−18	2,182	2,922	0.933	3.42	0.908
Linear with covariates^a	6	−2,899.0	6.65E−17	1,061	2,930	0.925	2.27	0.918

The model selected for the dose optimization software validation.