Measuring and understanding adherence in a home-based exercise intervention during chemotherapy for early breast cancer

K. A. Nyrop, A. M. Deal, S. K. Choi, C. W. Wagoner, J. T. Lee, A. Wood, C. Anders, L. A. Carey, E. C. Dees, T. A. Jolly, K. E. Reeder-Hayes, H. B. Muss

Ensuring and measuring adherence to prescribed exercise regimens are fundamental challenges in intervention studies to promote exercise in adults with cancer. This study reports exercise adherence in women who were asked to walk 150 min/week throughout chemotherapy treatment for early breast cancer. Participants were asked to wear a Fitbit(TM) throughout their waking hours, and Fitbit steps were uploaded directly into study computers. Descriptive statistics are reported, and both unadjusted and multivariable linear regression models were used to assess associations between participant characteristics, breast cancer diagnosis, treatment, chemotherapy toxicities, and patient-reported symptoms with average Fitbit steps/week. Of 127 women consented to the study, 100 had analyzable Fitbit data (79%); mean age was 48 and 31% were non-white. Mean walking steps were 3956 per day. Nineteen percent were fully adherent with the target of 6686 steps/day and an additional 24% were moderately adherent. In unadjusted analysis, baseline variables associated with fewer Fitbit steps were: non-white race (p = 0.012), high school education or less (p = 0.0005), higher body mass index (p = 0.0024), and never/almost never drinking alcohol (p = 0.0048). Physical activity variables associated with greater Fitbit steps were: pre-chemotherapy history of vigorous physical activity (p = 0.0091) and higher self-reported walking minutes/week (p < 0.001), and higher outcome expectations from exercise (p = 0.014). Higher baseline anxiety (p = 0.03) and higher number of chemotherapy-related symptoms rates "severe/very severe" (p = 0.012) were associated with fewer steps. In multivariable analysis, white race was associated with 12,146 greater Fitbit steps per week (p = 0.004), as was self-reported walking minutes prior to start of chemotherapy (p < 0.0001). Inexpensive commercial-grade activity trackers, with data uploaded directly into research computers, enable objective monitoring of home-based exercise interventions in adults diagnosed with cancer. Analysis of the association of walking steps with participant characteristics at baseline and toxicities during chemotherapy can identify reasons for low/non-adherence with prescribed exercise regimens.