The feasibility of using SenseCams to measure the type and context of daily sedentary behaviors

  • Authors:

  • Catherine Marinac;Gina Merchant;Suneeta Godbole;Jacqueline Chen;Jacqueline Kerr;Bronwyn Clark;Simon Marshall

  • Affiliations:

  • San Diego State University, San Diego, California and University of California San Diego, San Diego, California;San Diego State University, San Diego, California and University of California San Diego, San Diego, California;University of California San Diego, San Diego, California;University of California San Diego, San Diego, California;University of California San Diego, San Diego, California;University of Queensland, Brisbane, Australia;University of California San Diego, San Diego, California

  • Venue:
  • Proceedings of the 4th International SenseCam & Pervasive Imaging Conference
  • Year:
  • 2013

Quantified Score

Hi-index 0.00

Visualization

Abstract

The SenseCam data can be used to estimate time spent in specific episodes of sedentary behaviors, as well as some dimensions of sedentary behaviors. However, it is unknown whether SenseCam data can be aggregated to provide an objective estimate of total sedentary time accumulated during a single day. We compared SenseCam-derived day-level estimates to self-report estimates of time spent in sedentary behaviors using 39 days of concurrent SenseCam and self-report data from a sample of university employed adults (age 18--70 years). We also examined whether SenseCam data can be used to compute day-level estimates of specific dimensions of sedentary behavior (e.g., co-occurring sedentary behaviors and social context). Twenty-four percent of the days of SenseCam image data collected did not have enough image data (i.e., ≥8 hours of data) to generate day-level estimates. Further, the day-level agreement between the SenseCam and self-report estimates of time spent in sedentary behaviors varied considerably by device wear time. In terms of dimensions of sedentary behaviors measured by the SenseCam, over one-third of the total sedentary time involved a social interaction and the majority (71%) of the estimated sedentary time was spent in one behavior. Overall, SenseCam data can be used to compute day-level estimates of time spent in specific episodes of sedentary behaviors and the images provide data on critical dimensions of these behaviors; however, device wear-time significantly influences the accuracy of day-level estimates.