How Has Intervention Fidelity Been Assessed in Smoking Cessation Interventions? A Systematic Review

Introduction. Intervention fidelity concerns the degree to which interventions are implemented as intended. Fidelity frameworks propose fidelity is a multidimensional concept relevant at intervention designer, provider, and recipient levels; yet the extent to which it is assessed multidimensionally is unclear. Smoking cessation interventions are complex, including multiple components, often delivered over multiple sessions and/or at scale in clinical practice; this increases susceptibility variation in the fidelity with which they are delivered. This review examined the extent to which five dimensions from the Behaviour Change Consortium fidelity framework (design, training, delivery, receipt, and enactment) were assessed in fidelity assessments of smoking cessation interventions (randomised control trials (RCTs)). Methods. Five electronic databases were searched using terms "smoking cessation," "interventions," "fidelity," and "randomised control trials." Eligible studies included RCTs of smoking cessation behavioural interventions, published post 2006 after publication of the framework, reporting assessment of fidelity. The data extraction form was structured around the framework, which specifies a number of items regarding assessment and reporting of each dimension. Data extraction included study characteristics, dimensions assessed, data collection, and analysis strategies. A score per dimension was calculated, indicating its presence. Results. 55 studies were reviewed. There was a wide variability in data collection approaches used to assess fidelity. Fidelity of delivery was the most commonly assessed and linked to the intervention outcomes (73% of the studies). Fidelity of enactment scored the highest according to the framework (average of 92.7%), and fidelity of training scored the lowest (average of 37.1%). Only a quarter of studies linked fidelity data to outcomes (27%). Conclusion. There is wide variability in methodological and analytical approaches that precludes comparison and synthesis. In order to realise the potential of fidelity investigations to increase scientific confidence in the interpretation of observed trial outcomes, studies should include analyses of the association between fidelity data and outcomes. Findings have highlighted recommendations for improving fidelity evaluations and reporting practices.