Testing motivational and self-regulatory mechanisms of action on device-measured physical activity in the context of a weight loss maintenance digital intervention: a secondary analysis of the NoHoW trial

Abstract

Background: To date, few digital behavior change interventions for weight loss maintenance focusing on long- term physical activity promotion have used a sound intervention design grounded on a logic model under- pinned by behavior change theories. The current study is a secondary analysis of the weight loss maintenance NoHoW trial and investigated putative mediators of device-measured long-term physical activity levels (six to 12 months) in the context of a digital intervention. Methods: A subsample of 766 participants (Age = 46.2 ± 11.4 years; 69.1% female; original NoHoW sample: 1627 participants) completed all questionnaires on motivational and self-regulatory variables and had all device- measured physical activity data available for zero, six and 12 months. We examined the direct and indirect effects of Virtual Care Climate on post intervention changes in moderate-to-vigorous physical activity and number of steps (six to 12 months) through changes in the theory-driven motivational and self-regulatory mechanisms of action during the intervention period (zero to six months), as conceptualized in the logic model. Results: Model 1 tested the mediation processes on Steps and presented a poor fit to the data. Model 2 tested mediation processes on moderate-to-vigorous physical activity and presented poor fit to the data. Simplified models were also tested considering the autonomous motivation and the controlled motivation variables inde- pendently. These changes yielded good results and both models presented very good fit to the data for both outcome variables. Percentage of explained variance was negligible for all models. No direct or indirect effects were found from Virtual Care Climate to long term change in outcomes. Indirect effects occurred only between the sequential paths of the theory-driven mediators. Conclusion: This was one of the first attempts to test a serial mediation model considering psychological mechanisms of change and device-measured physical activity in a 12-month longitudinal trial. The model explained a small proportion of variance in post intervention changes in physical activity. We found different pathways of influence on theory-driven motivational and self-regulatory mechanisms but limited evidence that these constructs impacted on actual behavior change. New approaches to test these relationships are needed. Challenges and several alternatives are discussed.