Multivariate growth charts and robust quantile estimation

Growth charts have been widely used in clinics and medical centers to monitor an individual subject's growth or health status in context of population values. Current growth charts consider only one measurement at a time. The investigators propose to develop a framework for multivariate growth charts. More specifically, the objectives of the proposed research include (1) to provide a mathematically formal but clinically sensible definition of multivariate growth charts, and study their properties; (2) to propose methodology for estimating and incorporating time effects and other potential covariate effects (e.g. past growth) into those growth charts; (3) to develop statistical inference and model assessment tools; (4) to compare the proposed methodology with existing methods and explore alternative devices for multivariate growth chart construction; (5) to propose robust quantile estimators to limit the impact of outliers. The multivariate growth charts can provide much more informative tools than the currently available univariate growth charts in many areas of public health and biomedical applications. For example, pediatrics will be able to do better screening for obesity with a multivariate approach using longitudinal information on weight and height than with a single body-mass-index (BMI) chart. The methodologies to be developed by the investigators (e.g. estimation of covariate-adjusted multivariate quantile functions) are of general interest to statistical research. The proposed research will be widely disseminated through publications, presentations in domestic and international conferences, and collaborations with clinical researchers.