Fidelity of delivery of a physical activity intervention: Predictors and consequences

Assessing fidelity of behavioural interventions is important, but demanding and rarely done. This study assessed adherence to behaviour change techniques used in an intervention to increase physical activity among sedentary adults (ProActive; N = 365). Transcripts of 108 sessions with a sub-sample of 27 participants were assessed. An independent assessor coded adherence of four 'facilitators' who delivered the intervention to 208 protocol-specified facilitator behaviours (e.g. 'elicit perceived advantages of becoming more active') in four key sessions. Four raters classified the 208 behaviours under 14 techniques (e.g., goal setting, use of rewards) to enable calculation of adherence to techniques. Observed adherence to techniques across participants was modest (median 44%, IQR 35-62%), and lower than that reported by facilitators. Adherence differed between facilitators (range: 26-63%) and decreased across the four sessions (mean drop 9% per session, 95% confidence interval 7-11%). In this small sample facilitator adherence was unrelated to (change in) participants' physical activity or its cognitive predictors: Attitudes, subjective norm, perceived behavioural control and intention. Future research should investigate causal pathways between fidelity indicators and outcomes in larger samples and develop and test less intensive measures of fidelity.