In Reply:

In Reply: We thank Drs. Sundermann and Alfonso for engaging with our article in the December 2024 issue.1 Methods for becoming pregnant, such as medically assisted reproduction, indeed play a role in neonatal outcomes; this is important to consider in sexual orientation–related disparities. However, there are major challenges reflected in our study that also plague the field. Almost no sufficiently sized data set includes detailed data on sexual orientation identity, pregnancy-level outcomes, and medically assisted reproduction use. The current data set—with some of the most robust data available on pregnancy outcomes by sexual orientation—assessed only ovulation induction, but not at the pregnancy level; therefore, we did restrict some analyses to individuals who reported never undergoing any ovulation induction (Appendix 7, https://links.lww.com/AOG/D865). Even if more detailed pregnancy-level, medically assisted reproduction use data were available, mediation analyses require advanced methods that have yet to be applied in this field. Because variables such as sexual orientation are upstream, factors affected by sexual orientation, such as access to health insurance, can affect both the mediator (eg, medically assisted reproduction) and outcome (eg, offspring with preterm birth or low birth weight). This exposure-induced mediator–outcome confounding violates key mediation assumptions, and simply adjusting for medically assisted reproduction would introduce new collider stratification bias. This challenge is common in health disparities research. Therefore, formal causal mediation analyses are required, which we are currently conducting. Furthermore, studies linking medically assisted reproduction use (particularly in vitro fertilization) with preterm birth and low birth weight have been conducted primarily in heterosexual populations experiencing physiologic infertility; thus, it is unclear whether medically assisted reproduction itself, the underlying causes of physiologic infertility, or both lead to these outcomes. Although more research is needed, we found that sexual orientation disparities in low birth weight were slightly wider when we restricted to individuals who never underwent ovulation induction compared with analyses in the full sample. We echo Drs. Sundermann's and Alfonso's call for additional research and extend it further by calling for new data collection that will enable analyses of the mechanisms of sexual orientation–related disparities in neonatal outcomes.