Background Local authorities, planning departments and housing developers are responsible for delivering environmental and individual-level interventions with high reach and potential impact on health. The implementation of such interventions is outside of a researcher’s control and natural experimental methods are often used. However, collaborating to implement a randomised controlled trial (RCT) of current policy offers an opportunity to learn about the challenges for intervention evaluation. We aim to reflect on the collaboration required to run an RCT.

Methods Northstowe, in the East of England, is a large new housing development and residents can claim financial incentives to support the use of alternative travel modes to the car, delivered by the local authority. ICAN-Northstowe was a mixed-method study and a collaboration between local and district authorities in Cambridgeshire and researchers to assess the feasibility of manipulating financial incentive exposure, collaboratively developing and implementing a RCT and the intervention to affect travel behaviour. The local authority defined the intervention, identified eligible households (n=475), and delivered the interventions. Researchers recruited participants in 2021 (n=99), randomised participants, conducted surveys and semi-structured interviews at baseline, three and six month follow up and analysed data. All stakeholders interpreted the data through regular meetings and a half-day workshop.

Results Overall, both parties were working towards a common overarching vision to evaluate and understand the potential of incentives. The collaboration offered the opportunity to evaluate a live intervention with results that were immediately valuable to a captive audience with capability to implement findings. The process identified different perspectives that required management and compromise to balance the competing priorities. These included: (1) adhering to necessary academic and ethical processes while minimising additional workload on practitioners and exercising flexibility to capture the dynamic practice landscape; (2) incompatibility of practical processes, e.g., data sharing that required adaptation from both parties; (3) differing dissemination priorities which included timescales, target audiences and format. Practitioners prioritised understanding the intervention effect on modal shift and implementable practice changes, whereas researchers also wanted to develop an understanding of why incentives were successful (or not).

Conclusion This study demonstrated that with early involvement of practice partners it is feasible to co-deliver a randomised controlled trial based on current practice. Advantages include setting objectives which are practice-based, therefore accelerating the impact of the research. Acknowledging the challenges and compromises required for collaborative working is important to inform future research to evaluate policy-led interventions.