ASPO Abstracts
Measuring physical activity in breast cancer survivors: Comparison of Actigraph, Fitbit, and International Physical Activity Questionnaire data
Category: Lifestyles Behavior, Energy Balance & Chemoprevention
Conference Year: 2023
Abstract Body:
Aim: To evaluate agreement of physical activity (PA) measures from Actigraph GT3Xs, the International Physical Activity Questionnaire (IPAQ), and Fitbits in breast cancer survivors (BCS). Methods: Cook and Move for Your Life was a feasibility study testing a 6-month nutrition and PA behavioral intervention in female BCS. In the study, PA was collected using 3 methods: 1) wearing an Actigraph GT3X accelerometer for 7 days; 2) self-reporting PA over the previous 7 days (IPAQ); and 3) using a Fitbit Inspire HR to self-monitor PA. Actigraph and IPAQ data were collected at baseline and 6 months, while Fitbit data were collected following randomization to 6 months. Measures of light, moderate-to-vigorous (MVPA), and total PA were estimated for each method: validated cut-points from non-sleeping wear time for Actigraph, a standardized scoring procedure for IPAQ, and an undisclosed proprietary algorithm for Fitbit. MVPA of 150 min/week indicated meeting PA guidelines. Correlations of PA levels and agreement on meeting the MVPA target were estimated using Pearson correlation coefficients and kappa statistics, respectively. Results: Seventy-three BCS provided baseline data for the analysis; 48 also provided 6-month data. Baseline mean MVPA in min/week was 158 for Actigraph and 298 for IPAQ. Similarly, 6-month mean MVPA from the Actigraph, IPAQ, and Fitbit was 179, 279, and 382 min/week, respectively. At baseline and 6 months, Actigraph and IPAQ had very weak correlation for all PA measures (|r|<0.12). At 6 months, correlation between Fitbit and Actigraph was weak for MVPA (r=0.23) and strong for light (r=0.67) and total PA (r=0.65), but correlation between Fitbit and IPAQ was weak to very weak for all PA measures (|r<0.38). The proportion of participants meeting the MVPA target at 6 months measured by Actigraph, Fitbit, and IPAQ was 50%, 79%, and 63%, respectively, which translated to moderate agreement between Actigraph and Fitbit (kappa=0.6), and slight agreement between Actigraph and IPAQ (kappa=0.1). Conclusion: Agreement of subjective and objective PA measures ranged from poor to moderate, highlighting the potential variability among these measures in BCS. Future studies of PA in BCS must balance potential measurement biases with cost and ease of data collection.
Keywords: physical activity assessment, breast cancer