Validation of the Fitbit Zip for monitoring physical activity among free-living adolescents

Margaret Schneider, Larissa Chau

The widespread availability of affordable consumer-oriented devices for monitoring physical activity offers an appealing option to physical activity researchers, but studies are needed to demonstrate the validity and reliability of these products. To examine the validity of the Fitbit Zip, we recruited three cohorts (N's = 25, 35, and 27) of middle-school students to wear the Fitbit and the ActiGraph simultaneously for a week. Adolescents were healthy volunteers representing a range of activity levels. Mean daily minutes of MVPA and mean steps per day were compared between the Fitbit Zip and the Actigraph. The step data for the Fitbit Zip correlated highly with the step data yielded by the ActiGraph (r's = .72, .92, .96), and the MVPA data for the Fitbit Zip correlated highly with the MVPA data from the ActiGraph (r's = .67, .79, .94). Bland-Altman plots revealed that the Fitbit Zip overestimated activity in comparison to the ActiGraph, especially for Cohort One, which completed the study before Fitbit modified their algorithms to count as activity only bouts that continued for at least 10 min. Our data suggest that the Fitbit Zip is a reasonable alternative to the ActiGraph for estimating activity among free-living adolescents. However, data from the Fitbit should not be used interchangeably with data from the ActiGraph, as there is a consistent tendency for the Fitbit to overestimate steps in comparison to the ActiGraph. Also, the findings confirm concern about using for research a consumer-oriented device that does not make public their algorithms.