Neighborhood poverty and preterm birth revisited: Comparing cross-sectional measures to longitudinal poverty trajectories

Background. Neighborhood poverty is associated with increased risk of preterm birth (PTB); however, most of this research uses cross-sectional measures of neighborhood poverty. We hypothesize that the longitudinal poverty experience of a neighborhood may influence factors important to perinatal health such as psychosocial stress or access to health-promoting resources. Our objective was to compare the associations between neighborhood poverty and PTB using both cross-sectional and longitudinal measures of poverty.

Methods. We categorized California neighborhoods (i.e., census tracts) as having low (<5% poor), moderate (5-20% poor), or high (>20% poor) cross-sectional poverty based on data from the American Community Survey 2005-2009 (ACS). We then estimated longitudinal poverty trajectories from 1970-2009 using data from the ACS and the Neighborhood Change Database using three methods: a priori categorization, latent class growth modeling, and non-parametric clustering. We used logistic regression to estimate associations between neighborhood poverty measures and PTB for 24,394 women in the Maternal and Infant Health Assessment who gave birth between 2003-2009.

Results. High (compared to low) neighborhood poverty based on cross-sectional data was not associated with PTB after adjustment for individual-level demographic and socioeconomic factors (OR=1.03, 95% CI=0.88,1.21) . However, trajectories characterized by long-term high (compared to long-term low) neighborhood poverty were associated with 30-40 percent increases in odds of PTB in fully adjusted models (e.g., OR=1.39, 95% CI=1.17, 1.64), regardless of method used to estimate trajectories.

Conclusions. The longitudinal poverty experience of neighborhoods may be more strongly associated with factors influencing birth outcomes compared to cross-sectional measures of poverty.