Examination of accelerometer reactivity among children, adolescents, and adults: Implications for influencing physical activity estimates, proportion meeting physical activity guidelines, and associations with health outcomes

The health benefits of regular Physical activity (PA) have been rigorously documented and continue to be investigated from the standpoint of prevention, as well as, treatment of a multitude of adverse health outcomes including cardiovascular disease, obesity, sleep disorders, depressive symptomatology and countless others. In order to minimize subjectivity and more accurately measure PA behavior, researchers often use accelerometers. Although this method is considered valid and reliable a methodological issue that has rarely been addressed is the potential “reactivity” effect of measuring PA with an objective measure, such as accelerometry. Reactivity is a behavior change by the participant due to the fact that he/she is aware that they are being monitored. The implications of accelerometer reactivity are such that, if present, then accelerometer-determined PA estimates as well as associations between PA and health outcomes may be biased. These estimates (i.e., PA estimates and their associations with health) inform the development and implementation of PA-related surveillance systems and intervention studies; thus, identification of whether accelerometry reactivity occurs not only has implications for the validity of an accelerometry study, but also has far reaching implications at the community and policy level. As a result, the purpose of the current study was to examine whether accelerometry reactivity is present in a nationally representative sample of U.S. children, adolescents, and adults. Three specific aims will be addressed in this study. The first aim of our study will examine whether accelerometer reactivity is indeed present, with evaluations considered across a nationally representative U.S. sample of children, adolescents, and adults, as well as, in various demographic and morbidity characteristics. Second, if reactivity is detected, the proportion meeting physical activity guidelines will be calculated in the original data, then recalculated upon removal of ‘reactive’ data, in order to determine the extent to which reactivity biased proportional estimates meeting PA guidelines, as well as, weekly aggregative estimates of total physical activity and total activity counts. Similar to the second aim, the last aim will examine the extent to which potential reactivity may influence the association between PA and health