Comparing national device-based physical activity surveillance systems: A systematic Review

Background: Physical activity surveillance systems are important for public health monitoring but rely
mostly on self-report measurement of physical activity. Integration of device-based measurements
in such systems can improve population estimates, however this is still relatively uncommon in existing surveillance systems.

Purpose: This systematic review aims to create an overview of the methodology used in existing device-based national PA surveillance systems.

Methods: Four literature databases (PubMed, Embase.com, SPORTDiscus and Web of Science) were
searched, supplemented with backward tracking. Articles were included if they reported on
population-based (inter)national surveillance systems measuring PA, sedentary time and/or
adherence to PA guidelines. When available and in English, the methodological reports of the identified surveillance studies were also included for data extraction.

Results: This systematic literature search followed the PRISMA guidelines and yielded articles on one or multiple waves of 15 different national and one international surveillance system. The included studies showed substantial variation between (waves of) systems in number of participants, response rates, population representativeness and recruitment. In contrast, the methods were similar on data reduction definitions (e.g. minimal number of valid days, non-wear time and necessary wear time for a valid day).

Conclusions: The results of this review indicate that few countries use device-based PA measurement
in their surveillance system. The employed methodology is diverse, which hampers comparability
between countries and calls for more standardized methods as well as standardized reporting on these methods.

Practical implications: The results from this review can help inform the integration of device-based PA measurement in (inter)national surveillance systems.

Funding: This work was supported by Amsterdam Public Health research institute and Statistics Netherlands.