Multi-level repeated measures growth modelling using extended spline functions.

This paper explores the fitting of multi-level models to growth data over a wide age range using a new class of extended spline models. These extend conventional splines based on a '+' function representation by allowing variable order functions and by including fractional polynomial terms. The work focuses on modelling human growth in height and head circumference with example data sets. The procedures can be used with covariates and for comparing population parameters.