Measuring steps with the Fitbit activity tracker: an inter-device reliability study

Area of Science:

• Sports Science
• Human-Computer Interaction
• Biomedical Engineering

Background:

• Activity trackers like Fitbit are increasingly used for self-monitoring physical activity.
• Social comparison of activity data can motivate behavioral changes.
• High inter-device reliability is crucial for meaningful social comparisons.

Purpose of the Study:

• To assess the inter-device reliability of Fitbit Ultra activity trackers in step measurement.
• To evaluate reliability across different time aggregation levels (minutes, hours, days).

Main Methods:

• Ten Fitbit Ultra devices were worn by one individual for eight consecutive days.
• Inter-device reliability was analyzed using intra-class correlation coefficient (ICC), Bland-Altman plots, limits of agreement (LOA), and Mixed Model Analysis.
• Data was aggregated at minute, hour, and day levels.

Main Results:

• Fitbit Ultra devices demonstrated good inter-device reliability for step measurement at all aggregation levels.
• Reliability was particularly strong when steps were measured on a daily basis.
• Statistical analyses confirmed consistent performance across multiple devices.

Conclusions:

• Fitbit activity trackers offer reliable step-counting, supporting accurate peer comparisons.
• Daily aggregation of step data provides the most dependable basis for social comparison.
• The findings support the use of Fitbit for social-based physical activity interventions.