Bayesian Mixed-Effects Location Scale Models for the Analysis of Objectively Measured Physical Activity Data from a Lifestyle Intervention Trial

• Activity Number: 335
• Type: Topic Contributed
• Date/Time: Tuesday, August 6, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Health Policy Statistics Section
• Author(s): Juned Siddique*+ and Donald Hedeker and Bonnie Spring
• Companies: Northwestern University and University of Illinois-Chicago and Northwestern University Feinberg School of Medicine
• Keywords: physical activity ; accelerometer ; location-scale ; random effects

Abstract

Objective measurement of physical activity using wearable accelerometers is now used in large-scale epidemiological studies and clinical trials of lifestyle and exercise interventions. These devices measure the frequency, intensity, and duration of physical activity at the momentary level, often using measurement epochs of 1 minute or less yielding a large number of observations per subject. In this talk, we describe a Bayesian mixed-effects location scale model for the analyses of physical activity as measured by accelerometer using data from a randomized lifestyle intervention trial of 204 men and women. We model both the mean and variance over time as a function of subject-specific random effects and time varying covariates. Our model allows us to measure how covariates can influence both the mean and the variance of physical activity over time. It also allows us to measure whether changes in variability over the course of the study are predictive of treatment relapse.