On adjusting for the outcome of previous pregnancies in epidemiologic reproductive studies.