However, some caution is necessary when interpreting results for the older cohort, where we do not have data on AD age of onset and thus direction of influence is unknown. If for many women onset of AD follows rather than precedes first reproduction, estimated effects of AD on timing of first childbirth may be biased. While comparison of models with time-invariant and -varying predictors suggest minimal impact in the young cohort, to the extent that onset of AD occurs earlier for younger women, such comparisons may be less informative.
