The objectives of this study were to assess the difference between actual a
nd nominal significance levels. as judged by the likelihood ratio lest, for
hypothesis tests regarding covariate effects using NONMEM, and to study wh
at factors influence these levels. Also. a strategy for obtaining closer ag
reement between nominal and actual significance levels was investigated. Ph
armacokinetic (PK) data without covariate relationships were simulated from
a one compartment iv bolus model for 50 individuals. Models with and witho
ut covariate relationships were then fitted to the data, and differences in
the objective function values were calculated. Alterations were made to th
e simulation settings; the structural and error models, the number of indiv
iduals, the number of samples per individual and the covariate distribution
. Different estimation methods in NONMEM were also tried. In addition, a st
rategy for estimating the actual significance levels for a specific data se
t, model and parameter was investigated using covariate randomization and a
real data set. Under most conditions when the first-order (FO) method was
used, the actual significance level for including a covariate relationship
in a model was higher than the nominal significance level. Among factors wi
th high impact were frequency of sampling and residual error magnitude. The
use of the first-order conditional estimation method with interaction (FOC
E-INTER) resulted in close agreement between actual and nominal significanc
e levels. The results from the covariate randomization procedure of the rea
l data set were in agreement with the results from the simulation study. Wi
th the FO method the actual significance levels were higher than the nomina
l, independent of the covariate type, but depending on the parameter influe
nced When using FOCE-INTER the actual and nominal levels were similar. The
most important factors influencing the actual significance levels for the F
O method are the approximation of the influence of the random effects in a
nonlinear model, a heteroscedastic error structure in which an existing int
eraction between interindividual and residual variability is nor accounted
for in the model, and a lognormal distribution of the residual error which
is approximated by a symmetric distribution. Estimation with FOCE-INTER and
the covariate randomization procedure provide means to achieve agreement b
etween nominal and actual significance levels.