![[ Visit Client Website ]](images/banner.png)
BACKGROUND: Infant mortality is an important indicator of the health and well-being of a society because it is affected by many factors, including the community's poverty rate and education level, its economic development and the access to health care. The death of an infant is most often the result of complex interactions among factors, which presents challenges in identifying successful approaches to reducing infant mortality. In Virginia, we used multiple methods to inform policy and prevention efforts at the state and local levels.
METHODS: We used linked birth and infant death records as well as fetal death records from 2002-2011 for the state of Virginia. First, the Perinatal Periods of Risk (PPOR) approach was used to provide a summary of infant deaths categorized into periods with similar underlying causes and interventions. Next, logistic regression modeling of individual and community factors was conducted to identify risk and protective factors for infant mortality. For the regression modeling, deaths were stratified into three categories (birth defects, neonatal excluding birth defects, and postneonatal excluding births defects) because of the known differences in causes of death and associated prevention efforts in these groups. Additionally, the linked dataset was then geocoded to identify statistically significant clusters of infant mortality throughout the state using hotspot analysis.
RESULTS: PPOR revealed that while disparities in infant mortality persist between non-Hispanic Black and non-Hispanic White women in the state, the Maternal Health and Prematurity category has the highest rate of infant deaths for both racial/ethnic groups. Significant predictors of infant mortality across the three adjusted models were maternal race, maternal education, birthweight, interpregnancy interval, prenatal care, and condition of the infant at birth. Births that were normal birthweight or had longer interpregnancy intervals significantly reduced the odds of mortality by more than half for all three categories. Overall, affordability, educational level, deprivation, and environmental hazard accounted for about 30% of the variation in the infant mortality rate when using geographically weighted regression. The southeast region of Virginia was identified as a hotspot with almost half of the infant deaths from the time period occurring in this area.
CONCLUSIONS: Infant mortality can be disaggregated and targeted by using comprehensive analytic methods at the individual and community level to better understand the complex relationships involved. Addressing risk factors and areas of intervention identified by different methods can strengthen efforts to reduce infant mortality in Virginia.