Lifecourse epidemiology is the study of chronic disease risk associated with the long? term effects of physical, metabolic, behavioral, social and psychosocial exposures occurring throughout the life. Current statistical methods for risk factor trajectories fail with multiple outcomes, inconsistent timing or multiple risk factor trajectories, or can fail to replicate. A multidisciplinary team of biostatisticians and lifecourse epidemiologists proposes to address the gaps. We build on the strength of our previous R01, during which we published 22 peer?reviewed journal articles, gave 20 invited talks and 1 webinar, and built free, point?and? click, power and sample size software, used more than 500 times each month, by more than 350 unique scientists. The examples in the proposal come from a funded study of the primordial causes of diabetes and obesity, which will use maternal gestational blood pressure trajectory as a predictor of future childhood obesity. Better design and analysis tools will allow lifecourse epidemiologists to better answer crucial questions about the importance of risk factor trajectories in the etiology of diseases. We have four aims. 1) Develop novel data analysis methods for lifecourse epidemiology studies that assess the association between one or more outcomes and trajectories of risk factors, examine the direct and mediated effects of risk factor trajectories, and determine critical periods within risk factor trajectories. 2) Derive new sample size methods for lifecourse studies involving risk factor trajectories. 3) Extend our widely used free, open?source, point?and?click sample size software for studies of risk factor trajectories to allow scientists to use the new sample size methods. 4) Widely disseminate the new techniques and free software through articles and seminars. Leverage our NIH funding to extend our short courses and massive open online course (1R25GM111901?01), add chapters to a book?in?progress (NLM 1G13LM011879?01), and produce more tutorials for our power and sample size website, currently visited by over 350 scientists each month (NIDCR 1 R01 DE02083201A1). While the research was inspired by a study of fetal programming, the results apply broadly to any study in which hypotheses about relating risk factor trajectories to outcomes are of interest. Risk factor trajectories are increasingly being examined for diseases as diverse as obesity, cardiovascular disease, stroke, diabetes, and cancer. Finally, extending our widely used power and sample size software to most classes of mixed models will bring accurate, easy?to?use, point?and?click study design to hundreds more NIH supported scientists.

R01GM121081

2010-12-09

2021-06-30