Fetal loss, gravidity and pregnancy order: is the truncated cascade analysis valid?

Methods of studying birth order effect after controlling for final number of pregnancies or the elimination of women with a large number of pregnancies are shown to produce misleading results. The 'cascade' method of analysis is only valid in any population where birth limitation is practised if all pregnancy histories are included and there are no efforts made to control for number of pregnancies. It is suggested that in populations where any form of birth control is practised, so-called longitudinal analyses of perinatal risk should be treated with great caution.