A meta-analysis of Fitbit devices: same company, different models, different validity evidence

Area of Science:

• Physical Activity Measurement
• Wearable Technology Validation
• Biomedical Data Analysis

Background:

• Fitbit devices are widely used for tracking physical activity and energy expenditure.
• Existing research on Fitbit validity is fragmented, lacking a comprehensive summary.
• Objective validity evidence across various Fitbit models is needed.

Purpose of the Study:

• To systematically examine the validity evidence of different Fitbit models for estimating energy expenditure.
• To investigate how device-specific factors influence the accuracy of Fitbit energy expenditure estimates.
• To provide a summarized validity evidence for Fitbit devices through meta-analysis.

Main Methods:

• A meta-analysis was conducted on 29 studies identified from five databases, analyzing 402 articles.
• Seven moderator variables were examined: Fitbit model, device placement, device type, heart rate capability, release year, activity types, and sedentary activity.
• The summarized validity coefficient (r) was calculated for energy expenditure estimation against criterion references.

Main Results:

• The overall summarized validity coefficient for Fitbit devices estimating energy expenditure was r=.64.
• Fitbit model itself was not a significant factor, but heart rate capability, activity types, and sedentary activity significantly impacted validity.
• Not all Fitbit models demonstrated similar accuracy in estimating energy expenditure.

Conclusions:

• Fitbit devices offer moderate validity in estimating energy expenditure during physical activity.
• Device features like heart rate monitoring and the type of activity being tracked are crucial for accuracy.
• Further research is necessary to understand and improve the validity of Fitbit devices, considering influencing factors.