Primary Submission Category: Migration
Disparities in preterm birth by generation of immigration among Latines
Authors: Ashley Judge, Christina Ludema, Kelli Ryckman,
Presenting Author: Ashley Judge*
Preterm birth is associated with poor infant outcomes and long-term developmental consequences. Foreign-born Latines typically have better birth outcomes than US-born Latines, but this advantage diminishes with longer US residence. Few studies have separated generation of immigration among US-born Latines (2nd, or ≥3rd), due to limited data collected on parental country of birth.
We used data from a prospective cohort study of nulliparous individuals (2010-2013) among eight, clinical US sites, limited to 1581 individuals who self-identified as Hispanic. We used logistic regression models to estimate the odds of preterm birth (PTB) between cross-sectional 1st, 2nd, and ≥3rd generation Latines.
Among n=670 1st generation immigrants, 24% had resided in the US for <5 years and 40% were born in the Dominican Republic (DR). Around half of the 1st generation reported annual household incomes <$25,000 compared to 30% of the 3rd generation. The prevalence of smoking during pregnancy varied from 8% to 24% between the 1st and ≥3rd generation. Compared to the 1st generation, 2nd and ≥3rd generation Latines had 0.8 (0.8, 0.9) and 0.7 (0.7, 0.8) times the adjusted odds of PTB, respectively (prevalence: 1st=9%, 2nd=8%, and 3rd=7%). Associations differed by country of origin comparing the 1st and 2nd generations (Mexico: 1.4 (1.2, 1.7), (1st=4%, 2nd=8%); DR: 0.5 (0.4, 0.6), (8%, 5%)). Patterns of reported racial discrimination also differed by country of origin.
At first take, our results suggest a health advantage with increasing generation of immigration, but substantial effect modification by country of origin was observed. Our results suggest increasing social mobility and risky health behaviors in later generations, with more pronounced movement out of the <$25,000 category for 2nd generation individuals from DR. Our future work includes using decomposition methods to compute the percentage of disparities explained by individual-level and (possibly) neighborhood-level factors.