Latent Variable Growth Curve and Mixture Models for Life Course Epidemiology

点击翻译按钮获取中文摘要

Preventing chronic illnesses such as cancer and cardiovascular disease requires a better understanding of their causes. The importance of adulthood risk factors such as smoking and diet are familiar to most people. However, research suggests that many of the diseases we face in adulthood may also be influenced by factors we experience in childhood, infancy, and even before we are born. To better understand these life course influences on disease risk, epidemiologists study cohorts of people that are followed up for a long period of time, often from birth. These kinds of studies have become more common over the past thirty years, but it is not entirely clear which statistical methods are best suited to analyze these data. My research aims to explore and demonstrate the usefulness of two advanced statistical methods for life course researchers. Using latent growth curve models and growth mixture models, I will investigate how infant and childhood growth patterns are related to obesity, using data from several existing studies. Research findings will be disseminated in a manner that facilitates the use of these methods by other researchers without advanced statistical training.