br Methods br Results br Discussion

Methods
Results
Discussion In this paper we have drawn on Extended Normalisation Process Theory to evaluate the delivery of a complex intervention – the Strengthening Families Programme 10–14. We have used the theory as a framework to understand how the interplay between the intervention and local delivery systems shaped implementation. This was achieved by considering how practitioner agency (capability and contribution) and delivery systems (capacity and potential) interacted with the intervention (May, 2013), the extent to which it was delivered as intended, and how this varied over space and time. The findings provide valuable contextual information for the RCT which is assessing the effectiveness of SFP 10–14 on behavioural outcomes, and important insights into the extent to which the programme can be delivered as intended within a UK setting, following Paprotrain from the original US version. Overall SFP 10–14 was delivered with good fidelity, and families received the intended intervention. Adherence to programme content by facilitators was high, but with some variation across delivery sites. These findings mirror those of previous evaluations of SFP 10–14 in the USA (Cantu et al., 2010; Spoth, Guyll, Trudeau & Goldberg-Lillehoj, 2002), and provide evidence that its implementation functions in broadly similar ways across contrasting national contexts, with key activities which are hypothesized to produce behaviour change being delivered in line with the programme\'s logic model. In our study, high potential among practitioners towards SFP 10–14 appeared to have a positive influence on implementation. Practitioners valued the intervention (and were committed to delivering it) because they believed SFP 10–14 could help families and fill gaps in existing services. Through facilitator training, delivery staff gained an understanding of what they were required to deliver and how programme components were intended to work and interact, thus enhancing their capability to deliver it. Variation in adherence to planned programme content appeared to be influenced by levels of capacity to achieve coordination across multi-agency networks, though coordinator potential was also important. In particular, the organisation and quality of facilitators’ preparatory meetings – which were designed to optimise delivery processes and group dynamics - varied significantly. Programme adherence appeared to be higher in counties where the meetings were held regularly and fulfilled their intended purpose. Nonetheless, it is possible that the quality of meeting arrangements was associated with other aspects of implementation which also affected adherence. Targets for staffing numbers and consistency were met, but involvement of facilitators from multi-agency networks became increasingly difficult as levels of capacity reduced – partly due to cuts in public-sector jobs and funding. However, coordinators were sometimes able to overcome these challenges, for example by drawing more on staff from the programme\'s host agency or organising multiple meetings. Family recruitment into the trial was challenging, and group size and composition targets were not always achieved. Practitioners’ concerns about random allocation of participants to intervention/control arms, and universal provision of the intervention (as opposed to focusing on families with support needs) impacted levels of potential and their willingness to refer families. However short-term funding, job losses in partner agencies, and competition from other programmes also played a part in reducing capacity to support recruitment – a situation many interventions face. Despite practitioners’ concerns about the use of an RCT design, some aspects of the trial may have had a positive impact on implementation capacity and capability, particularly comprehensive funding for staffing, multi-agency practitioner training, and resources to enable family attendance – levels of funding which were not always sustained after the end of the study. Good rates of fidelity could also have been shaped by the heightened levels of monitoring which the trial introduced, such as observation of programme sessions, reinforcing practitioner commitment and enhancing intervention contribution.