PROXIMATE DETERMINANTS OF LBW AND INFANT MORTALITY

点击翻译按钮获取中文摘要

Closing the persistent black/white infant mortality differential is a high priority national objective. We hypothesize it to be a manifestation of how social inequality negatively affects the health of black mothers, and limits their access to pre- and perinatal health services. Our theoretical framework connects socioeconomic characteristics to infant outcomes through biologically-based proximate determinants. We propose a "weathering" hypothesis that the effects of social inequality on the health of populations compound with age. Our ongoing project confirms that black women have higher prevalence rates of health and behavioral characteristics that can complicate pregnancy than white, and that these differences increase with age. Black/white differentials in poor birth outcomes also increase with maternal age. The differences in women's health may be important mechanisms driving the black/white infant mortality differential. The release of state linked birth and infant death certificate tapes for the 1989 birth cohort, when the birth certificates were improved, makes testing these hypotheses possible for the first time. We will analyze tapes for California, New York, Michigan and North Carolina for the 1989 cohort. These states offer large numbers of births (including black and Hispanic), represent each region of the country, and vary in infant mortality and poverty rates. We will augment these with census data to provide measures of SES. We will supplement and validate the linked maternal health data with HANES data and do the same for the census data using data from the PSID. We will estimate multivariate logit models of the impact of maternal health, behaviors, prenatal care, and hospital level on birth outcomes; and of the relationship of maternal age, race, state, and other characteristics to maternal health. We will construct in-depth profiles of the states, as qualitative backdrops to the analyses to aid interpretation. Implemented by a multidisciplinary research team, this research goes beyond the demographic identification of high-risk populations, and beyond the biomedical focus on disease processes at the individual level, to connect disease processes to populations.