Maximizing the Utility and Comparability of Accelerometer Data from Large-Scale Epidemiologic Studies.

Abstract

There is much evidence showing that physical activity is related to optimal health, including physical and mental function, and quality of life. Additionally, data are accumulating with regards to the detrimental health impacts of sedentary behavior. Much of the evidence related to long-term health outcomes, such as cardiovascular disease and cancer - the two leading causes of death in the United States and worldwide, comes from observational epidemiologic studies and, in particular, prospective cohort studies. Few data on these outcomes are derived from randomized controlled trials, conventionally regarded as the "gold standard" of research designs. Why is there a paucity of data from randomized trials on physical activity or sedentary behavior and long-term health outcomes? A further issue to consider is that prospective cohort studies investigating these outcomes can take a long time to accrue sufficient numbers of endpoints for robust and meaningful findings. This contrasts with the rapid pace at which technology advances. Thus, while the use of devices for measuring physical behaviors has been an important development in large-scale epidemiologic studies over the past decade, cohorts that are now publishing results on health outcomes related to accelerometer-assessed physical activity and sedentary behavior may have been initiated years ago, using "dated" technology. This paper, based on a keynote presentation at ICAMPAM 2022, discusses the issues of study design and slow pace of discovery in prospective cohort studies and suggests some possible ways to maximize the utility and comparability of "dated" device data from prospective cohort studies for research investigations, using the Women's Health Study as an example.