Environmental hazards, social inequality, and fetal loss: implications of live-birth bias for estimation of disparities in birth outcomes

Background: Restricting to live births can induce bias in studies of pregnancy and developmental outcomes, but whether this live-birth results underestimating disparities is unknown. Bias may arise from collider stratification due an unmeasured common cause fetal loss the outcome interest, or depletion susceptibles, where exposure differentially causes among those with underlying susceptibility. Methods: We conducted a simulation study examine magnitude population parameterized resemble one year conceptions California (N=625,000). simulated non-time-varying environmental hazard, risk spontaneous abortion, time birth using 1,000 Monte Carlo simulations. Our interest was preterm birth. included social vulnerability factor represent disadvantage, estimated overall differences for linear probability models stratified by factor. calculated how often confidence intervals true point estimate (CI coverage probabilities) illustrate effect estimates differed qualitatively truth. Results: Depletion susceptibles resulted larger compared stratification, socially vulnerable group. Coverage probabilities were not adversely affected stratification. reduced coverage, especially (coverage = 46%, non-socially 91% most extreme scenario). Conclusions: In simulations, hazardous exposures induced women. Keywords: bias; simulation; health disparities; justice;