Outputs Available from Objective Monitors

Abstract

Physical activity epidemiologists are focused upon the prevalence and patterns of physical activity and sedentary behavior in a population, as well as the causes and effects of these behaviours. Of utmost importance to such research is the careful and precise measurement of physical activity, historically reliant on self-reported behaviours. However, with the introduction, development, and enhance-ment of various body-worn accelerometer devices, the objective monitoring of physical activity has progressed from the simple recording of step counts obtained via pedometry to the objective assessment of a much broader range of time-stamped free-living behaviors. Most public health physical activity guidelines have been framed in terms of the recommended minimum weekly time to be spent in moderate-to-vigorous physical activity (MVPA), and researchers thus focused initially on translating manufacturers' activity count outputs into a comparable measure. Initial calibration research sought to establish instrument-specific activity count cut-points that were associated with absolutely defined intensity levels (indicated by metabolic equivalents or METs). However, it became increasingly apparent that any established cut-points were sample-and instrument-specific, hampering the generalization of findings and limiting epidemiological studies to an assessment of relative rankings of physical activity within a data set. Potential solutions to this issue advanced to date have considered data obtained from additional physiological sensors (there is little evidence as yet that these yield improved measurements), and the use of increasingly more complex statistical models, focused on additional features of the raw accelerometry signals. Success realized in the laboratory has been somewhat illusive under free-living conditions, and the scale of technical data storage was initially daunting. The steady emergence of a wide variety of consumer wearable monitors has opened up new opportunities for wireless data collection, but it has also complicated the landscape of measure-ment choices for epidemiologists and other investigators. We may thus need to enter a phase of reflection, followed by a right-sizing of epidemiological