Fetal loss rates and their relation to pregnancy order

Much of the recent controversy surrounding the relation between fetal death and pregnancy order has centred around the appropriateness of different types of analyses. In the present paper the interpretation of various methods are discussed with reference to "real" and "hypothetical" data. The pattern of results obtained when the fetal loss rates of a group of pregnancies are tabulated by pregnancy order was found to depend on the risk and parity distributions of the study population. These two parameters did not, however, appear to affect the within "sibship" or "gravidity" group patterns. These findings support the hypothesis that the frequently observed increase in fetal death rates in pregnancy orders above two could be largely artifactual. It is concluded that in any investigation of reproductive events women, and not their pregnancies, should form the prime unit of analyses.