Event Date
SPEAKER: Jodi Lapidus; Oregon Health and Science University
TITLE: “Intensive Longitudinal Models for Wearable Device Data”
ABSTRACT: The availability and use of wearable technologies -- body worn sensors that collect vital signs, activity, and sleep -- has exploded in recent years. Wearable devices are increasingly incorporated into cohort and intervention studies to investigate longitudinal profiles of health indicators. The rich longitudinal data produced facilitates investigation of more complex relationships. For example, researchers may be interested in characterizing variation in resting heart rate, in addition to the average, as a function of covariates over time. Heterogeneous mixed effects regression models are an ideal strategy for wearable device data as they, not only facilitate comparison of longitudinal profiles across groups of subjects, but can also be parametrized to estimate effects of covariates on intra- and inter-subject variation. Based on data from an occupational cohort of ~300 individuals, we examine how differences in the amount nightly sleep influences daily resting heart rate over 4-6 months. We demonstrate how to parameterize fixed and random components of the model to estimate sleep-related associations with intra- and inter-individual heart rate variation. We show the usefulness of this strategy for elucidating complex associations in wearable device and other intensive longitudinal data streams.
DATE: Thursday, April 18th, 4:10pm
LOCATION: MSB 2112, Math Dept (2nd floor, Mathematical Sciences Building)
REFRESHMENTS: 3:30pm MSB 4110 (4th floor lounge)
STA 290 Seminar List: https://statistics.ucdavis.edu/seminars