Multigenerational perinatal outcomes among Western Australian Aboriginal infants: a total population linked data study

Abstract

This thesis aims to identify reasons for the persistently poor perinatal health of Aboriginal infants using linked routinely-collected data. Multigenerational studies using such data are possible with Western Australia’s (WA) database of family relationships. Health records of Aboriginal infants born in WA from 1980 to 2011 and their relatives’ records were linked. Two methodological challenges are examined, followed by risks to perinatal health within a single generation, then over two generations. The first challenge followed a discovery that 18% of the children did not have a birth registration. Their mothers were more likely to be young and from remote areas. From a research perspective, this disadvantaged group are more likely to be excluded from studies using this dataset. The second challenge was the inconsistent recording of Aboriginal status in datasets and the poor understanding of the relative performance of algorithms used to identify who is Aboriginal. The consistency with which the algorithms assigned Aboriginal status to family members was assessed. The best-performing algorithm was then supplemented with relatives’ information to further improve consistency. Third, the contribution of maternal health behaviours to perinatal health was estimated. Of 28,119 births from 1998 to 2010, 27% of infants were small for gestational age (SGA), preterm, and/or died perinatally. Half (51%) of the infants were exposed in utero to maternal smoking, alcohol misuse, drug misuse, and/or assault and 37% of SGA births, 16% of preterm births, and 20% of perinatal deaths were attributable to these factors. Finally, fetal programming has been put forward as a contributor to poor health in indigenous populations. It refers to a fetus’ response (including growth restriction) to a hostile uterine environment, which has lifelong effects and, potentially, affects her offspring in turn. However, any causal relationship between maternal and offspring fetal growth is confounded by shared genetic and environmental factors. Two family-based approaches addressed this confounding; they provided little support for a hypothesis of heritable fetal disadvantage. In conclusion, significant improvements in fetal health can be expected when maternal risk factors shift, unconstrained by the mother’s own fetal history.