Examining sociodemographic and health-related characteristics as moderators of an mHealth intervention on physical activity outcomes in young adult cancer survivors

This study explored whether sociodemographic and health-related characteristics moderated mHealth PA intervention effects on total and moderate-to-vigorous physical activity (MVPA) at 6 months, relative to a self-help condition among young adult cancer survivors (YACS). We conducted exploratory secondary analyses of data from a randomized controlled trial among 280 YACS. All participants received digital tools; intervention participants also received lessons, adaptive goals, tailored feedback, text messages, and Facebook prompts. Potential moderators were assessed in baseline questionnaires. PA was measured at baseline and 6 months with accelerometers. Linear model repeated measures analyses examined within- and between-group PA changes stratified by levels of potential moderator variables. Over 6 months, the intervention produced MVPA increases that were ≥ 30 min/week compared with the self-help among participants who were males (28.1 vs. -7.7, p = .0243), identified with racial/ethnic minority groups (35.2 vs. -8.0, p = .0006), had baseline BMI of 25–30 (25.4 vs. -7.2, p = .0034), or stage III/IV cancer diagnosis (26.0 vs. -6.8, p = .0041). Intervention participants who were ages 26–35, college graduates, married/living with a partner, had a solid tumor, or no baseline comorbidities had modest MVPA increases over 6 months compared to the self-help (ps = .0163-.0492). Baseline characteristics did not moderate intervention effects on total PA. The mHealth intervention was more effective than a self-help group at improving MVPA among subgroups of YACS defined by characteristics (sex, race, BMI, cancer stage) that may be useful for tailoring PA interventions. These potential moderators can guide future optimization of PA interventions for YACS. NCT03569605.