An Improved Version of Activity Intensity and Its Comparison with Activity Count, with the Application to Women Health Initiative

• Activity Number: 270
• Type: Contributed
• Date/Time: Monday, August 10, 2015 : 2:00 PM to 2:45 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Jiawei Bai* and Chongzhi Di and Ciprian Crainiceanu and Luo Xiao
• Companies: The Johns Hopkins University and Fred Hutchinson Cancer Research Center and The Johns Hopkins University and The Johns Hopkins University
• Keywords: accelerometer ; accelerometry ; activity intensity ; physical activity ; activity count

Abstract

Summarizing densely sampled raw accelerometry signals into low frequency meaningful and interpretable metrics is extremely important for public health and medical research. The main purpose of the paper is to formally compare two metrics: Activity Count (AC) from ActiGraph GT3X+ and the improved version of Activity Intensity (AI) from Bai et al (2014). The new AI was shown to possess favorable properties such as easy deployment, additivitiy and rotational invariance. The data used in the paper were from a subgroup of 10 elderly women who wore an ActiGraph GT3X+ on the hip and performed certain activities in the lab, while the starting and ending time of each activity was recored. Indirect calorimeters were used to measure the oxygen consumption of the subjects, which were then converted into metabolic equivalent of task (MET). Through exploratory analysis as well as receiver operating characteristic analysis, AI was found 1) being able to much better distinguish between low intensity activities than AC; 2) better associating with MET than AC.