We propose a generalized linear mixed model to estimate and test margi
nal effects on titers repeatedly measured by serial dilution assays. T
he link is log-log and the titer is assumed to follow a gamma distribu
tion. The parameters are estimated by generalized estimating equations
. The marginal effects are tested by means of Wald and score tests, us
ing the robust estimator of the variance. This approach avoids the pro
blems arising from assays leading to nonestimable individual titers. S
imulations were used to compare the Wald and score tests to Wilcoxon a
nd Student t-tests. This method is applied to the comparison of the an
tiviral efficiency of three treatments against HIV.We propose a genera
lized linear mixed model to estimate and test marginal effects on tite
rs repeatedly measured by serial dilution assays. The link is log-log
and the titer is assumed to follow a gamma distribution. The parameter
s are estimated by generalized estimating equations. The marginal effe
cts are tested by means of Wald and score tests, using the robust esti
mator of the variance. This approach avoids the problems arising from
assays leading to nonestimable individual titers. Simulations were use
d to compare the Wald and score tests to Wilcoxon and Student t-tests.
This method is applied to the comparison of the antiviral efficiency
of three treatments against HIV.