TESTING CAUSAL HYPOTHESES AND ASSOCIATIONS BETWEEN PERINATAL RISK FACTORS AND OFFSPRING MORBIDITY AND MORTALITY

Abstract

The Developmental Origins of Health and Disease (DOHaD) hypothesis is a broad theoretical framework that highlights how early risk factors have a causal influence on later physical and psychopathological outcomes. Numerous studies have documented such associations and underlying mechanisms have been hypothesized. Most existing studies, however, have not been able to rule out the possibility of environmental and genetic confounding. Thus, concerns exist about causal interpretations of the statistical associations identified between early risk factors and later outcomes. The six projects in my dissertation use quasi-experimental designs to rigorously test causal inferences across perinatal risk factors and offspring outcomes. The first two projects use a natural experiment approach to compare risk and hypothesized mechanisms across sensitive periods of development. In particular, the random occurrence of maternal bereavement stress across the preconception, prenatal, and postnatal periods was studied as a risk factor for offspring infant mortality and child and adult psychiatric problems. The next pair of projects examined birth weight and physical, psychiatric, educational, and socioeconomic problems using a sibling-comparison design. The final two projects used cousin-comparisons to explore the parental correlates and offspring psychiatric and education problems associated with interpregnancy interval, or the duration between the birth of an earlier born sibling and the conception of following sibling. Across these projects, findings both support and refute previous causal claims and important novel associations are identified. These studies allowed for a thorough examination of the nature of the associations between several perinatal risk factors and offspring physical, psychiatric, educational, and socioeconomic problems. The projects illustrate how combining several quasi-experimental designs can specifically test the DOHaD hypothesis by ruling out plausible alternative hypotheses. My findings also inform the direction future DOHaD-based studies should pursue.