The main aim of this study was to test whether the effectiveness of a worksite physical activity intervention delivered in five work organizations varied as a function of intervention fidelity. We conducted a fidelity analysis as part of a large matched-pair cluster randomized controlled trial of a worksite physical activity intervention (AME for Activity). Participants (N = 1260) were employees from five organizations in the UK. The primary trial outcome was physical activity at 9 months post intervention. Adherence, exposure, quality of delivery and participant responsiveness/engagement were measured to assess fidelity. Qualitative data about the context in which the intervention was delivered were collected via focus groups, interviews and field notes. Multi-level modelling was used to provide a comparison of the effect of the intervention on increases in physical activity for worksites where intervention fidelity was good, compared with those where intervention fidelity was poor or moderate. Intervention fidelity was poor in two organizations, moderate in two organizations and good in one organization (local council). Re-analysis of the trial data comparing employees in the local council (N = 443) with employees in all other worksites (N = 611) revealed a significant effect of the intervention on physical activity levels among council employees only. These findings suggest that the measurement of fidelity and the testing of the effects of intervention fidelity on outcomes, as part of the evaluation of complex interventions, are essential to understand the context and conditions in which interventions are most effective.
Keywords: cluster randomized controlled trial; fidelity; physical activity intervention; worksite.
© The Author 2014. Published by Oxford University Press. All rights reserved. For Permissions, please email: journals.permissions@oup.com.