Background Ensuring academic research leads to research that is useful for end users is a key challenge in the health research arena. Stakeholder engagement is being increasingly recognised as an important way to achieving impact. The workHORSE project was designed to continuously engage with stakeholders, via four iterative workshops and an e-platform, to inform the development of an open source/open access modelling tool to enable commissioners to quantify the potential cost-effectiveness and equity of the NHS Health Check Programme. An objective of the project is to evaluate the involvement of stakeholders in the process of building the workHORSE computer modelling tool.

Methods The design of the workshop programme was theory-based using the Cairney/Oliver key co-production principles. We identified stakeholders using our extensive networks and snowballing techniques. Iterative development of the decision support modelling tool was informed through engaging with stakeholders during three workshops (to date). We used detailed scripts facilitating open discussion and opportunities for stakeholders to provide additional feedback subsequently. At the end of each workshop, stakeholders completed stakeholder engagement questionnaires to explore their views and experiences throughout the process. The research team also completed questionnaires to explore their expectations prior to the workshops and their experiences thereafter.

Results A total of 25 stakeholders have participated, of which 11 attended two or more workshops. They spanned all levels: local (NHS commissioners, GPs, local authorities and academics), third sector and national organisations (including Public Health England).

Stakeholders experiences were positive overall. They felt valued and commended the involvement of practitioners. Major reasons for attending included being able to influence development and having insight and understanding of what the tool could include and how it would work in practice. They appreciated the iterative process involving a series of workshops which provided opportunities for them to learn about and reflect upon the model’s capacity, usage and usefulness. Researchers saw the process as an opportunity for developing a common language and trust in the end product and ensuring the support tool was transparent. The workshops have acted as a reality check ensuring model scenarios and outputs are relevant and fit for purpose.

Conclusion Computational modellers rarely consult with end users when developing tools to inform decision-making. The added value of co-production (collaboration and iteration with stakeholders) potentially enables modellers to produce a ‘real-world’ operational tool. Likewise, stakeholders have increased confidence in the decision support tool’s development and applicability in practice.