Back to Annual Meeting Page		American Public Health Association 133rd Annual Meeting & Exposition December 10-14, 2005 Philadelphia, PA

Abstract #108544

Patrick Crockett, PhD¹, Eric Harvey, PhD¹, Xuguang Guo, PhD¹, David Francis, PhD², and Pim Brouwers, PhD³. (1) Constella Health Sciences, 2605 Meridian Parkway, Durham, NC 27713, 919-544-8500, pcrockett@constellagroup.com, (2) Department of Psychology, University of Houston, 4800 Calhoun Rd, Houston, TX 77204, (3) National Institute of Mental Health, 6001 Executive Boulevard, Rockville, MD 20852

For some health effects, one anticipates that an affected population will begin to diverge from normal responses some time after exposure or infection, and it is of interest to be able to characterize the starting time of divergence. As an example, for children born with HIV, scores on neurocognitive tests may be consistent with normal children until a certain age, and then the HIV children may not develop at the same rate after that point. Such data may be modeled with mixed effects models including splines with unknown knots. Others have shown how to include splines into mixed effects models [Ruppert, Wand, and Carroll, 2003], but the locations of knots (the point of divergence in our example) are specified in advance. We present a method for incorporating a linear spline into a mixed effects model, allowing the knot location to be a random effect. We demonstrate the use of the method for longitudinal data (Bayley II scores [Lindsey & Brouwers, 1999]) on children born with (or without) HIV.

Learning Objectives:

• Learn when a mixed effects model is appropriate.
• Learn the basic nature of linear splines.
• Learn the form of a mixed effects model with linear splines with random knots.
• Learn how to fit such a model to data, including complications and how to deal with them.

Presenting author's disclosure statement:

I wish to disclose that I have NO financial interests or other relationship with the manufactures of commercial products, suppliers of commercial services or commercial supporters.

Recorded presentation