The aim of this study was to characterise the population pharmacokinetics of efavirenz in a representative patient population and to identify patient characteristics influencing the pharmacokinetics of efavirenz, with the ultimate goal of further developing techniques that can be applied to optimise therapeutic drug monitoring of antiretroviral agents. Ambulatory HIV-1-infected patients using an efavirenz-containing regimen were included. During regular visits, blood samples were collected for efavirenz plasma concentrations and clinical chemistry parameters. Concentrations of efavirenz were quantitatively assessed by a validated high-performance liquid chromatographic with ultraviolet detection method. Using nonlinear mixed-effect modelling (NONMEM), the pharmacokinetics of efavirenz were described. Disposition of efavirenz was described by a two-compartment model and absorption was modelled using a chain of three transition compartments. Apparent clearance (CL/F), volume of distribution after oral administration (V(d)/F), intercompartmental clearance, the peripheral volume of distribution and the intercompartmental transition rate constant (k(tr)) were estimated. Furthermore, interindividual, interoccasion and residual variability were estimated. The influence of patient characteristics on the pharmacokinetic parameters of efavirenz was explored. From 172 patients, 40 full pharmacokinetic curves and 315 efavirenz plasma concentrations at a single timepoint were available, resulting in a database of 1009 efavirenz plasma concentrations. CL/F, V(d)/F, and k(tr) were 11.7 L/h (4.3% relative standard error [RSE]), 189L (14.6% RSE) and 3.07 h(-1) (11.2% RSE), respectively. Residual variability in the model was composed of 0.14 mg/L additive error and 8.85% proportional error. Asian race and baseline total bilirubin (TBR) increased the relative bioavailability of efavirenz by 56% and 57%, respectively. No significant covariates were found for CL/F or V(d)/F. The pharmacokinetic parameters of efavirenz were adequately described with the developed population pharmacokinetic model. Asian race and baseline TBR were found to be significantly correlated with the bioavailability of efavirenz. The described model will be an essential tool in further optimisation of efavirenz-containing antiretroviral therapy, e.g. by the use of Bayesian estimation of individual pharmacokinetic parameters.