Moderators of health behavior initiation and maintenance in a randomized telephone counseling trial.

OBJECTIVE This study compares moderators of initiation and maintenance of health behavior changes. METHODS Data come from a cluster-randomized, 12-month telephone counseling intervention for physical activity and diet, targeting type 2 diabetes or hypertension patients (n=434, Australia,2005-2007). Demographic and health-related characteristics, theoretical constructs, and baseline behavioral outcomes were considered as moderators. Mixed models, adjusting for baseline values, assessed moderation of intervention effects for trial outcomes (physical activity, intakes of fat, saturated fat, fiber, fruit, vegetables) at end-of-intervention (12 months/initiation) and maintenance follow-up (18 months), and compared moderation between these periods. RESULTS Social support for physical activity and baseline physical activity were significant (p<0.05) moderators of physical activity at 12 months. Gender, marital status, social support for healthy eating, BMI, and number of chronic conditions were significant moderators of dietary changes at 12- and/or 18 months. Instances of moderation differing significantly between 12- and 18 months were: baseline physical activity for physical activity (initiation) and marital status for fat intake (maintenance). CONCLUSIONS This exploratory study showed that moderation of physical activity and diet effects sometimes differed between initiation and maintenance. To identify unique moderators for initiation and/or maintenance of behavior changes, future studies need to report on and statistically test for such differences.