Synthesis and Validity of Accelerometer Devices and Methods Used in Epidemiological Studies of Physical Activity Bout Duration and Health Outcomes: A Systematic Review.

Abstract

Background: A paradigm shift in physical activity (PA) research and recommendations emerged in the most recent guidelines, which now emphasize that all physical activity bouts count,irrespective of their duration. Nevertheless, the validity of accelerometer devices and methods for detecting PA bouts of any duration, as considered in studies underpinning the PA guidelines on bout duration and health, remains unclear. The aim of this systematic review was to systematically analyze the validity level of accelerometer devices and methods that have been used in epidemiological studies on the association between PA bout duration and health outcomes.

Methods: This systematic review followed the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines and employed a two-step search strategy. First, we identified the existing body of evidence on PA bout duration and health by updating the systematic search originally performed for the 2018 Physical Activity Guidelines Advisory Committee scientific report and subsequently expanded by Jakicic et al. Three databases (PubMed, CINAHL, and Cochrane) were searched up to 14 November 2024. The second and final step, which constituted the primary focus of this review, was to retrieve and systematically evaluate the underlying validity studies cited in these epidemiological investigations, with risk of bias assessed using a modified version of the QUADAS-2 tool.

Results: A total of 48 epidemiological studies were included. Analysis of device type, position, epoch duration, drop time, bout duration, and start year of study data collection revealed 113 methodological instances, representing 68 unique combinations, underscoring the heterogeneity in practice. Among the 48 epidemiological studies, 26 validation studies were referenced, and only four were deemed suitable for addressing bout performance detection, accounting for less than 20% of epidemiological studies. For those four studies, the overall study risk of bias was ranked as "High" or with "some concerns".

Conclusions: Better accuracy and harmonization of accelerometer devices and methods are crucial for broader adoption in national health surveillance systems, ensuring comparability of results and facilitating global knowledge.

Registration: This systematic review was registered on PROSPERO (CRD42023394705).