Engaging stakeholders in realist programme theory building: insights from the prospective phase of a primary care dementia support study

‘Dementia - Personalised Care Team’ (D-PACT) is a five-year NIHR funded programme, using realist methods to develop and evaluate a complex, person-centred intervention for people with dementia and their carers. During the early project stages, we engaged with multiple stakeholders, including people with dementia and their carers, to develop an initial programme theory (IPT) – into an elaborated programme theory (EPT), by helping to uncover intervention mechanisms leading to outcomes in specific contexts. Realist research methods for developing programme theories are under-reported. In addition, there is a paucity of practical guidance on how to engage underserved and vulnerable populations in complex interventions programme theory development. We attend to these gaps, providing a worked example of how we meaningfully engaged people living with dementia and carers, alongside field experts, as stakeholders in this process. Our IPT theory building included multi-stakeholder primary research exercises and meetings with PPI contributors and an Expert Reference Group. We adapted interview schedules, and used visual resources and scenario-based activities, to support stakeholders to think in a ‘realist’ way. Using realist and thematic analyses led to hypothesis-building of causal mechanisms. Sharing findings with stakeholders led to further refinement of the intervention design, ready for testing in a subsequent feasibility study. We found that, despite the cognitive challenges associated with dementia, innovative methods of engagement can enable this stakeholder group to understand the realist approach and provide a platform through which to share their experiences. Taking a highly flexible and unhurried approach, led to novel insights into the complexities of person-centred dementia support. We argue for more detailed methodological guidance, based on realist principles, on how to collaborate with underrepresented populations to rigorously gain insights as to what is lik