An important advantage of the embedded design would
be the ability to directly compare safety and other outcomes
in pregnant and nonpregnant women (and men),
since these participants would be enrolled in the same trial
and, therefore, would be directly comparable. It is likely
that stand-alone PhaseI trials in pregnancy would enroll
only pregnant women because of the expense of including
the nonpregnant controls. Therefore, comparison of data
from pregnant women in stand-alone Phase I trials with
nonpregnant women and men in the initial Phase I or II
trials would be equivalent to comparison with historical
controls. By contrast, embedded PhaseI trials in pregnant
women would potentially increase the statistical power of
the trial. If no differences were detected for a given outcome
between pregnant and nonpregnant participants,
then the data on these participants could be combined;
this would increase the statistical power.