Back

Comparison of sedentary time calculated from count- and raw accelerometer data: The Tromso Study

Weitz, M.; Morseth, B.; Hopstock, L. A.; Brage, S.; Horsch, A.

2025-06-03 epidemiology
10.1101/2025.06.02.25328840 medRxiv
Show abstract

Accelerometers have become a popular and indispensable tool to objectively measure various aspects of daily behavior. While physical activity can be estimated relatively easy, sedentary behavior requires reliable Time in Bed and non-wear time algorithms. Over the years, various algorithms have been proposed, but little is known about how these algorithms quantitatively influence the estimation of sedentary behavior. In this study, we systematically compared different data processing strategies and their effect on sedentary behavior in 6155 participants asked to wear an accelerometer for seven consecutive days. We found considerable variation in the estimates ranging from 183 min to 807 min per day, mainly due to considerable influence by time-in-bed algorithms and sedentary thresholds. Non-wear time algorithms, on the other hand, only had a minor influence on the estimation of sedentary time, despite substantial variation between the algorithms. Our results suggest that sedentary behavior estimates from accelerometers must be interpreted with respect to the methods used to obtain them. When results from different studies are compared, the data should be reprocessed in a harmonized way.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.