Comparing Reference Charts for Cross-Section and Longitudinal Data

Reference charts are valuable tools for clinicians in their daily work on pediatric clinics. Reference charts are often constructed by smoothing techniques, and in this paper we present a newly developed non-parametric test for comparing these charts. We illustrate the method by two examples. The rst example compares crosssectional data on height in children from two Danish studies from 1970 and 1990, respectively. A second example shows how longitudinal data on growth for two types of skeletal dysplasia may be compared. In the cross-sectional setting the test compares the average height over the range of ages for the two groups, and in the longitudinal setting the test compares, similarly, the average height given a particular history of development.