Mediators of physical activity outcomes in a peer-led intervention for breast cancer survivors

Physical activity (PA) adoption can reduce treatment-related sequelae of breast cancer. Peer-led PA interventions are a promising and relatively inexpensive approach to scaling up interventions. The current study seeks to identify mediators of PA change amongst cancer survivors enrolled in a peer-led behavior change intervention. The study team partnered with the American Cancer Society's Reach to Recovery program (RTR) whose volunteers' provided information and support to breast cancer survivors.

Participants were 76 survivors (mean age 55.6 years, 1.1 years since diagnosis) who were randomized to PA Plus RTR (12-week PA telephone counseling delivered by RTR volunteers) or contact control. Data was collected on PA (self-reported and objectively measured) and potential mediators (self-efficacy, exercise decisional balance [ratio of the pros/advantages to the cons/disadvantages] and social support) at baseline and 12 weeks. Using a multiple mediation approach with bootstrapped standard errors, we examined mediators of the intervention effect on PA outcomes.

Compared to control, PA Plus RTR participants had higher mean self-efficacy, lower decisional balance cons and social support at 12 weeks controlling for baseline. Higher mean self-efficacy was associated with greater minutes of self-reported PA, whereas higher decisional balance pros was associated with higher objectively measured PA at 12 weeks. There were significant indirect effects of self-efficacy on self-reported PA and decisional balance on objectively measured PA.

PA Plus RTR increased self-reported and objectively measured PA by changing theoretical constructs hypothesized to be associated with behavior change. Peers delivering a PA intervention should focus on increasing survivors' self-efficacy for exercise especially in challenging circumstances such as being on vacation, and also help to overcome disadvantages of exercise such as taking time away from family.