Barriers to physical activity as moderators of intervention effects.

The impact of interventions to increase physical activity (PA) may vary as a function of participants' barriers to PA. The aim of this paper is to determine whether individual barriers (demographic, physical health, psychological health, neighborhood factors, perceived barriers to PA, social support for PA) moderate treatment effects on increases in PA. Three treatment conditions tested the relative efficacy of a group-based PA intervention alone or supplemented by either personal or automated phone calls made between group meetings. From 2010 to 2012, 284 African American women (ages 40–65) living in the Chicago, IL, area were randomized to one of the three treatment conditions. Data collection occurred at baseline as well as 24 and 48 weeks after baseline. Moderation of intervention effects by barriers to PA were tested across four outcome measures (self-reported moderate-vigorous PA, self-reported walking, accelerometer steps, and aerobic fitness) using multilevel mixed-effects analyses. Significant condition by barrier interaction effects for the accelerometer steps outcome were found for material hardships, general health, depressive symptoms, neighborhood crime rate, and perceived barriers to PA. For aerobic fitness, intervention effects were moderated by material hardships and perceived pain. Increases in the outcome variables were greater for the conditions in which group sessions were supplemented with personal and/or automated calls. Among participants with greater barriers to PA, supplementing the intervention group meetings with between-session personal and/or automated phone calls may be an effective way to strengthen intervention effects. These results may inform the use of treatment supplements in the context of adaptive interventions.