Perinatal & Pediatric
Maternal residential mobility between pregnancies: a California statewide sibling study Giselle Bellia* Giselle Bellia Haoran Zhou Zeyan Liew
Introduction Geocoded residential address is often used in epidemiological research to indicate individual’s social and environmental exposures from their living environment, but large-scale studies that characterize maternal residential mobility between pregnancies are sparse. Using a California statewide sibling study, we investigated the pattern of residential mobility between pregnancies and its association with several major social-demographic and perinatal factors.
Methods We geocoded maternal residential addresses extracted from the birth records of 919,013 sibling pairs from 2007-2015 in California. We described the moving patterns between pregnancies, and we utilized logistic regression to estimate the odd ratios for maternal moving according to demographic variables (age at delivery, education level, and race/ethnicity), pregnancy factors (pregnancy interval, pre-pregnancy obesity, smoking, and preeclampsia), and delivery/birth outcomes (mode of delivery, preterm delivery, and infant low birth weight) of the prior pregnancy. We also investigated predictors of maternal moving to a more disadvantaged neighborhood, indicated by a higher census-tract level social vulnerability index.
Results Overall, 52% of mothers moved between pregnancies. The median distance moved was 5.87 km. An additional year of the interpregnancy interval was associated with 33% higher odds of having moved between pregnancies, while having a college degree was associated with 60% lower odds. Mothers who were younger, reported smoking during pregnancy, and self-identified as non-Hispanic Black were associated with having moved to a more disadvantaged neighborhood between pregnancies.
Conclusions Maternal residential mobility between pregnancies was more common within certain demographic subgroups in California. Our findings have important implications for the understanding of potential carryover effects, time-varying confounding, and generalizability issues in sibling-matched analyses.
