BACKGROUND: Reports of decreased semen quality over time have raised concerns about possible reductions in human fertility. Studies of couple fertility have produced conflicting results. We evaluate how changes in the availability and use of effective contraception and induced abortion might bias the direct study of time trends in couple fertility. METHODS: We assess the potential for bias in the context of 2 common study designs: (1) a study of time-to-pregnancy that estimates fecundability (excluding unintended pregnancies) and (2) a study of infertility rates that categorizes couples as fertile or infertile (including couples with unintended pregnancies as fertile). RESULTS: In time-to-pregnancy studies, bias alone could produce more than a 2-fold apparent increase in fecundability over recent decades. In studies of infertility rates, the bias works in the opposite direction: a 30% underestimation of infertility during earlier decades could produce an apparent decrease in fertility over time. CONCLUSIONS: Over the past 5 decades, changes in social factors that affect the rate and fate of unintended pregnancies could substantially bias time trends in fertility. These biases may explain the conflicting reports in the literature. Except in rare settings in which the factors affecting reproductive choices have not changed, it is probably impossible to identify biologic changes in fertility over recent decades.
BACKGROUND: Reports of decreased semen quality over time have raised concerns about possible reductions in human fertility. Studies of couple fertility have produced conflicting results. We evaluate how changes in the availability and use of effective contraception and induced abortion might bias the direct study of time trends in couple fertility. METHODS: We assess the potential for bias in the context of 2 common study designs: (1) a study of time-to-pregnancy that estimates fecundability (excluding unintended pregnancies) and (2) a study of infertility rates that categorizes couples as fertile or infertile (including couples with unintended pregnancies as fertile). RESULTS: In time-to-pregnancy studies, bias alone could produce more than a 2-fold apparent increase in fecundability over recent decades. In studies of infertility rates, the bias works in the opposite direction: a 30% underestimation of infertility during earlier decades could produce an apparent decrease in fertility over time. CONCLUSIONS: Over the past 5 decades, changes in social factors that affect the rate and fate of unintended pregnancies could substantially bias time trends in fertility. These biases may explain the conflicting reports in the literature. Except in rare settings in which the factors affecting reproductive choices have not changed, it is probably impossible to identify biologic changes in fertility over recent decades.
Authors: Melissa M Smarr; Katherine J Sapra; Alison Gemmill; Linda G Kahn; Lauren A Wise; Courtney D Lynch; Pam Factor-Litvak; Sunni L Mumford; Niels E Skakkebaek; Rémy Slama; Danelle T Lobdell; Joseph B Stanford; Tina Kold Jensen; Elizabeth Heger Boyle; Michael L Eisenberg; Paul J Turek; Rajeshwari Sundaram; Marie E Thoma; Germaine M Buck Louis Journal: Hum Reprod Date: 2017-03-01 Impact factor: 6.918
Authors: Enrique F Schisterman; Sunni L Mumford; Richard W Browne; Dana Boyd Barr; Zhen Chen; Germaine M Buck Louis Journal: J Clin Endocrinol Metab Date: 2014-05-20 Impact factor: 5.958