How does variation in fetal loss affect the distribution of waiting times to conception?

Unmeasured variation has long been a concern in analyses of the waiting time to conception. Recent work by Heckman and Walker (1991) and Trussell and Rodriguez (1990) has underscored the fact that statistical considerations alone cannot discriminate among likely models describing the distribution. Here, we specify a single theoretically important source of heterogeneity, namely variability in intrauterine mortality, and assess its effects on the waiting times to a conception which results in a live birth. We find that the effects on the waiting times to a conception which results in a live birth. We find that the effects of variation in fetal loss are confined to the tail of the distribution. Unless variation in fetal loss is extremely large or a substantial proportion of observed waiting times are initiated at late ages, variation in fetal loss does not appear to explain much variation in conception waits. We conclude that heterogeneity in fetal loss does not explain the variation in fecundability that has been observed for first birth intervals. This conclusion supports the hypothesis that at early ages (below age 35) variation in the waiting time to a fertile conception may largely reflect the proportion of nonsusceptible couples in the population. The analyses suggest that for the purposes of testing theoretically motivated models, future efforts should be directed toward examining reproductive experience after age 35 and toward incorporating information on characteristics of the fertile period as it becomes known.