Long-term body weight trajectories and health in older adults: hierarchical clustering of functional curves

Anna Zajacova, University of Wyoming
Huong Nguyen, Ohio State University
Snehalata Huzurbazar, University of Wyoming

The relationship between body mass index (BMI) and health develops over the life course. There is increasing interest among researchers in modeling long-term changes in BMI and indentifying distinct BMI trajectory types in the population. Traditionally, researchers have used fully parametric (regression) or semi-parametric (latent class) models, which required difficult-to-justify decisions that sometimes yielded conflicting findings. The aim is to identify clusters of long-term BMI curves among older adults and associated health correlates, using a novel nonparametric functional-data approach. Data are from the Health and Retirement Study (N=9,893), a nationally representative panel survey of adults born in 1931-41. BMI was collected in up to 10 waves between 1992 and 2010. We utilize a cutting-edge functional data analysis for sparse longitudinal data, specifically hierarchical clustering of BMI functions estimated via the PACE algorithm. Three BMI trajectory clusters emerged for each gender: normal stable, overweight gaining, and overweight losing. The initial health of the overweight gaining group in both genders was poorer than that of the normal stable group but their mortality was comparable. The overweight losing cluster experienced significantly poorer health at baseline and higher risk of mortality. The BMI trajectories among older adults cluster into distinct types, with differing health risks. The study highlights the potential of functional data analysis for BMI trajectories, as well as many other developmental and age-dependent processes relevant to obesity and health.

  See paper

Presented in Poster Session 3