Physical Activity Trend eXtraction: A Framework for Extracting Moderate-Vigorous Physical Activity Trends From...

Area of Science:

• Health Sciences
• Behavioral Science
• Biomedical Engineering

Background:

• Moderate-vigorous physical activity (MVPA) provides significant health benefits but is often neglected.
• Research in physical activity behavior change is transitioning to objective data collection using fitness trackers.
• Extracting meaningful insights from wearable fitness tracker data presents unique challenges.

Purpose of the Study:

• To develop a framework for preprocessing and extracting MVPA trends from wearable fitness tracker data.
• To support studies focused on behavior change related to physical activity.
• To address challenges in analyzing objective MVPA data from fitness trackers.

Main Methods:

• Developed the Physical Activity Trend eXtraction (PATX) framework using heart rate data from fitness trackers.
• PATX framework includes data imputation, personalized heart rate zone recalculation, and MVPA trend extraction.
• Framework validated on 123 college participants over 18 months using Fitbit Charge HRs, with two case studies.

Main Results:

• PATX identified significant MVPA increases in 41 participants and decreases in 44 participants (P<.05).
• Case study 1 corroborated existing research on MVPA and mental health associations.
• Case study 2 revealed individuals are less likely to notice personal MVPA changes when social network ties mirror their own changes.

Conclusions:

• The PATX framework offers meaningful and flexible outputs, addressing data concerns associated with fitness trackers.
• PATX supports MVPA behavior change studies by enabling more objective assessments of physical activity.
• The framework facilitates deeper understanding of MVPA trends and their relationship with behavior and social context.