Analysis of time-to-pregnancy data

Statistical models for time-to-pregnancy data are usually formulated in dis crete time. Two approaches were surveyed, including methods of accounting f or both known heterogeneity (covariates, possibly time-dependent) and unkno wn heterogeneity (frailty). Delicate censoring and truncation patterns aros e for prospective and, particularly, retrospective sampling designs. The in clusion of several pregnancies per couple presents new challenges and possi bilities.