Comparing predictive abilities of longitudinal child growth models

Abstract

The Bill and Melinda Gates Foundation's Healthy Birth, Growth and Development knowledge integration project aims to improve the overall health and well-being of children across the world. The project aims to integrate information from multiple child growth studies to allow health professionals and policy makers to make informed decisions about interventions in lower and middle income countries. To achieve this goal, we must first understand the conditions that impact on the growth and development of children, and this requires sensible models for characterising different growth patterns. The contribution of this paper is to provide a quantitative comparison of the predictive abilities of various statistical growth modelling techniques based on a novel leave-one-out validation approach. The majority of existing studies have used raw growth data for modelling, but we show that fitting models to standardised data provide more accurate estimation and prediction. Our work is illustrated with an example from a study into child development in a middle income country in South America.

Keywords: child development; growth curve; trajectory.

Figure 1

Fitted growth trajectory of a single child based on fitting each of our 6 models on the raw scale. The blue dots are the data, the black dot is the held out datapoint, and the red line is the model fit. FACE, fast covariance estimation; SITAR, Superimposition by Translation and Rotation [Colour figure can be viewed at wileyonlinelibrary.com]

Figure 2

Fitted growth trajectory of a single child based on fitting 5 of our models on the Z scale (SITAR was not fitted on this scale). The blue dots are the data, the black dot is the held out datapoint, and the red line is the model fit. FACE, fast covariance estimation; HAZ, height‐for‐age Z score; SITAR, Superimposition by Translation and Rotation [Colour figure can be viewed at wileyonlinelibrary.com]

Figure 3

Comparison of results from fitting the brokenstick model on raw data and on Z‐transformed data. Each point represents the MSE values obtained from random holdout on 1 dataset. HAZ, height‐for‐age Z score; MSE, mean squared error [Colour figure can be viewed at wileyonlinelibrary.com]

Figure 4

Comparison of results from fitting the brokenstick model on raw data and on Z‐transformed data. Each point represents the MSE values obtained from last value holdout on 1 dataset. HAZ, height‐for‐age Z score; MSE, mean squared error [Colour figure can be viewed at wileyonlinelibrary.com]