Assessing and ensuring fidelity of the nationally implemented English NHS diabetes prevention programme: lessons learned for the implementation of large-scale behaviour change programmes.

Background: Health services interventions are typically more effective in randomised controlled trials than in routine healthcare. One explanation for this 'voltage drop', i.e. reduction in effectiveness, is a reduction in intervention fidelity, i.e. the extent to which a programme is implemented as intended. This article discusses how to optimise intervention fidelity in nationally implemented behaviour change programmes, using as an exemplar the National Health Service Diabetes Prevention Programme (NHS-DPP); a behaviour change intervention for adults in England at increased risk of developing Type 2 diabetes, delivered by four independent provider organisations. We summarise key findings from a thorough fidelity evaluation of the NHS-DPP assessing design (whether programme plans were in accordance with the evidence base), training (of staff to deliver key intervention components), delivery (of key intervention components), receipt (participant understanding of intervention content), and highlight lessons learned for the implementation of other large-scale programmes. Results: NHS-DPP providers delivered the majority of behaviour change content specified in their programme designs. However, a drift in fidelity was apparent at multiple points: from the evidence base, during programme commissioning, and on to providers' programme designs. A lack of clear theoretical rationale for the intervention contents was apparent in design, training, and delivery. Our evaluation suggests that many fidelity issues may have been less prevalent if there was a clear underpinning theory from the outset. Conclusion: We provide recommendations to enhance fidelity of nationally implemented behaviour change programmes. The involvement of a behaviour change specialist in clarifying the theory of change would minimise drift of key intervention content. Further, as loss of fidelity appears notable at the design stage, this should be given particular attention. Based on these recommendations, we describe examples of how we have worked with commissioners of the NHS-DPP to enhance fidelity of the next roll-out of the programme.