The Proximity of Spatial Clusters of Low Birth Weight and Risk Factors: Defining a Neighborhood for Focused Interventions.

Background: Low birth weight (LBW) is associated with significant mortality and morbidity and remains a significant preventable problem. Risk factors include socioeconomic, demographics, and characteristics of the environment. Spatial analysis can uncover unusual frequencies of health problems in neighborhoods, eventually leading to insights for targeted interventions.

Objectives: This study's goals were to 1. Evaluate the geographic distribution of spatial clusters of LBW births and maternal risk factors. 2. Determine the spatial relationship between risk factors and LBW.

Methods: This study obtained data on LBW newborns and risk factors from 19,013 births over 5 years (2012-2016) for Escambia County Census Tracts, extracted from FloridaCharts.com. Software was used to detect significant spatial clusters; these clusters were then plotted on a map. Poisson regression determined the statistical relationship between Census Tract risk factors and LBW. A separate analysis of the LBW cluster controlling for risk factors was also performed.

Results: All risk factor clusters resided in similar locations as the LBW cluster. The multiple Poisson regression model containing all risk factors fully explained the LBW cluster. On bivariate Poisson regression all risk factors in the Census Tract were significantly related to LBW whereas in multivariable Poisson regression, the proportion of births to African American women in the Census Tract remained significant after adjusting for other risk factors (p < 0.001).

Conclusions For Practice: Clusters of LBW and risk factors were located in the same region of the county, with the proportion of births to African American women in the Census Tract remaining significant on multiple Poisson Regression. Targeted interventions should be directed at the geographic level.