Grant: Travel Awards - Institute Community Support

Maternal weight gain is a potentially modifiable characteristic of pregnancy that has been associated with the health of both mothers and children. Traditional studies often explore the total or rate of maternal weight gain in pregnancy; however, these measures are imprecise, and their cumulative nature may bias results in studies of some health outcomes. Emerging research aims to leverage serial maternal weight measurements to more reliably investigate this association. However, research is often constrained by limited availability of information obtained at discrete gestational ages. Therefore, this research aimed to investigate and compare methods for estimating maternal weight between clinical visits. Two large clinical chart abstractions studies in Pittsburgh, Pennsylvania were leveraged to address this objective, although results are likely similarly applicable to pregnancies in Canada. Among the wealth of longitudinal data collected, we retained select maternal weight measurements (i.e. pre-pregnancy/delivery weight, as well as weight measured at first prenatal and glucose screening visits) in models to predict measurements at all other visits. We then calculated the difference between measured and predicted maternal weight, and summarized these differences (i.e. median, range, and mean squared error) to determine the relative performance of each model. Overall, 16 models were explored; 4 individual models that considered each mother's data separately, and 12 pooled models that borrowed information from all mothers to estimate each mother's pattern of weight change throughout pregnancy. We found that individual and pooled models that incorporated considerable variation in both pre-pregnancy weight and pattern of weight gain performed best. It is anticipated that these results will inform both ongoing and future studies investigating the relationship between maternal weight gain and pregnancy outcomes in North America and internationally.