Model-based approaches to studying fertility and contraceptive efficacy.

Statistical methods recently developed to aid in identifying environmental exposures with reproductive toxicity can also be applied to trials of interventions undertaken specifically to impair fertility, i.e. methods of contraception. Although only applicable in a trial that includes a reliable benchmark for identifying the day of ovulation, the proposed measures of contraceptive efficacy derived from such a trial offer certain interpretive advantages over the more traditional approaches of evaluating contraceptives. Extensions of the same models also allow one to evaluate efficacy under any assumed pattern of imperfect use. One can also evaluate methods based on biomarkers for the fertile phase of the cycle, such as hydration of the cervical mucus, that may prove to be enormously helpful to couples who wish to use periodic abstinence as their method. In prospective studies of fertility, couples who occasionally use a barrier method should not be excluded from the study, but can be retained, without biasing the estimates for fertility parameters.