Smartphone pervasive sensing of physical activity of overweight adults in a long-running randomized controlled trial (Preprint)

Abstract

BACKGROUND Clinical trials are expensive why it should be a priority to acquire as much data as possible during the trial. The burden on participants and staff is often the limiting factor on the amount of data feasibly acquired, which is why trials may benefit from incorporating the readily available small sensor-packed ubiquitous device; the smartphone. OBJECTIVE The aim of this study was to assess whether a smartphone can assist or replace existing practices in evaluating a physical activity intervention study in overweight sedentary adults. METHODS We introduce the smartphone as an additional sensing device in a physical activity intervention study that investigates the effects of active commuting and leisure-time exercise on a range of biological measures. RESULTS We find the smartphone is able to measure multiple modalities ubiquitously over a long duration and a hierarchical Bayesian analysis reveal estimates that are well in line with an independent analysis of the biological measures. CONCLUSIONS The smartphone has in this study shown that it, while not being without limitations, is able to augment current research methodologies and add value in historically infeasible ways. We can now ask questions that factorizes temporally on a minute-scale resolution and conceptually over the domains of everyday life. CLINICALTRIAL Clinicaltrials.gov NCT01962259