Mandibular growth prediction: mean growth increments versus mathematical models.

The aim of this study was to compare growth predictions obtained by adding mean annual velocities with predictions derived from a polynomial model of the population's growth curve. Given the child's previous measures at 11, 12 and/or 13 years of age, the cephalometric distance sella-gnathion at 15 years was estimated. Based on a sample of 223 boys and girls, the root mean square error decreased from 0.28 cm (males) and 0.18 cm (females) at 11 years, to 0.19 cm (males) and 0.12 cm (females) at 13 years. Root mean square errors were similar between methods, which was due to high correlations between measures across ages. Significantly, predictions based on mean increments were biased. They often over or underestimate growth for children who are larger and smaller than average. The observed bias was due to expected changes of variance associated with growth, which unconditional methods of prediction cannot control for. Predictions derived from growth models are conditional upon the child's size and are, therefore, unbiased.