Title: Optimal designs for fractional polynomial models
Authors: Victor Casero-Alonso - University of Castilla-La Mancha (Spain) [presenting]
Jesus Lopez-Fidalgo - University of Castilla-La Mancha (Spain)
Weng Kee Wong - UCLA (United States)
Abstract: Fractional polynomials have been shown to be much more flexible than polynomials for fitting continuous outcomes in the biological and health sciences. Despite their increasing popularity, design issues for fractional polynomials models have never been addressed. D- and I-optimal experimental designs for prediction using fractional polynomial models are provided, their properties are evaluated and a catalogue of design points useful for fractional polynomial models is provided. As applications, we re-design two studies using optimal designs and show they can produce substantial gains in terms of cost and quality of the statistical inference. We also provide a user friendly applet for generating optimal designs for fractional polynomials up to degree 3.