BACKGROUND: Smoking during pregnancy increases the risk of adverse birth outcomes, including preterm birth, low birthweight, and infant death. Maine has the 6th highest rate of smoking during pregnancy, and in 2013, the 9th highest infant mortality rate. Traditionally, interventions around smoking during pregnancy focus on the individual. While smoking is an individual decision, it is impacted by neighborhood, socioeconomic status, and cultural norms. The purpose of this study is to understand how social and community context and geographic location influence smoking during pregnancy in Maine.
METHODS: We geocoded 2014-15 in-state births to Maine residents (24,557) using an address locator created in ArcGIS 10.1. Smoking during pregnancy was defined as smoking at least one cigarette during pregnancy according to the birth certificate. Generalized linear mixed effects models were used to predict smoking during pregnancy, accounting for both individual characteristics and census tract-level social determinants. Individual factors included maternal age, Medicaid status, and maternal education. Birth data were linked to three census tract-level variables: concentrated disadvantage (CD), educational attainment, and Rural-Urban-Commuting-Area (RUCA) codes. CD was calculated using five U.S. Census variables: poverty, unemployment, public assistance, female-headed households, and population under 18. Each census tract was assigned a CD Index Score and divided into quartiles. All analysis was conducted in SAS 9.3.
RESULTS: Of the 24,557 births, 99.6% (24,456) were successfully geocoded. Most births were to women living in areas of very high or high CD (54%) and 65% of births occurred to women living in rural areas. Nearly 8% of women who gave birth did not have a high school degree. Medicaid was the principal payer for 44% of births. The proportion of women who smoked during pregnancy was higher in areas of high/very high CD and in small/isolated rural areas. Census tract-level maps of birth outcomes and risk factors will be shown. The multilevel modeling is currently in progress and these results will also be presented.
CONCLUSIONS: Smoking during pregnancy is more prevalent in areas with fewer resources and lower socioeconomic status, and the results of the multilevel modeling should provide more information on the impact of these social determinants. The results of these analyses encourage a shift in focus away from individual characteristics to social determinants of health that may influence a pregnant woman’s behaviors. These results also identify areas of the state for focused prevention and intervention efforts directed at reducing smoking during pregnancy.