Nonlinear Growth Curve Modeling Using Penalized Spline Models: A Gentle Introduction

초록

This didactic article aims to provide a gentle introduction to penalized splines as a way of estimating nonlinear growth curves in which many observations are collected over time on a single or multiple individuals. We begin by presenting piecewise linear models in which the time domain of the data is divided into consecutive phases and a separate linear regression line is fitted in each phase. Linear splines add the feature that the regression lines fitted in adjacent phases are always joined at the boundary so there is no discontinuity in level between phases. Splines are highly flexible raising the fundamental tradeoff between model fit and smoothness of the curve. Penalized spline models address this tradeoff by introducing a penalty term to achieve balance between fit and smoothness. The linear mixed-effects model, familiar from multilevel analysis, is introduced as a method for estimating penalized spline models. Higher order spline models using quadratic or cubic functions which further enhance a smooth fit are introduced. Technical issues in estimation, hypothesis testing, and constructing confidence intervals for higher order penalized spline models are considered. We then use data from the Early Childhood Longitudinal Study to illustrate each step in fitting a higher order penalized spline model, and to illustrate hypothesis testing, the construction of confidence intervals, and the comparison of the functions in 2 groups (boys and girls). Extensive graphical illustrations are provided throughout. Annotated computer scripts using the R package nlme are provided in online supplemental materials. Translational Abstract Methods such as daily diaries and technological advances such as wearable monitoring devices make it possible to repeatedly measure people during their daily lives. These advances produce rich longitudinal data sets that offer the possibility of providing a close representation of the trajectory of change over time. Traditional approaches such as regression with polynomial terms (e.g., X-2, X-3) may not be flexible enough to represent a complex pattern of change over time. Alternatives such as spline models that divide the data into adjacent time segments and fit a different polynomial function in each segment may be too flexible, producing a highly fluctuating trajectory that is too rough. In this didactic presentation we introduce penalized spline models that produce nonlinear trajectories that can balance the 2 goals of having a trajectory that fits complex patterns of change yet maintains a relatively smooth form. We show step-by-step the procedure of fitting penalized splines to represent the growth trajectory of a single individual or the mean trajectory of change in a group of individuals. We show how to perform hypothesis tests, construct confidence intervals, and compare the growth trajectories in 2 groups. We provide extensive graphical illustrations of each step of the procedure using data on the development of reading ability in children from kindergarten through 8th grade. We provide extensive annotated R computer scripts in online supplemental materials permitting researchers to apply these approaches to their own longitudinal data sets.

키워드