Background ‘Integrated care’ for chronic conditions is considered central to international health system reform. However, models of integrated care work differently in different circumstances. In Ireland, the National Diabetes Programme aimed to integrate diabetes care across primary, secondary and tertiary settings based on patient complexity through the introduction of new clinical posts and guidance for diabetes care. We conducted a realist evaluation to determine how and why the implementation of the programme worked (or not) across the country.

Methods Through documentary analysis and qualitative interviews (n=19) with a purposive sample of national stakeholders, we developed an initial theory on how the programme was expected to work. We then refined this theory in semi-structured interviews (n=39) with professionals purposively sampled to represent different clinical disciplines involved in implementation. We applied a realist logic of analysis and synthesis to iteratively build CMO configurations.

Results National stakeholders assumed that: 1) introducing guidance would formalise and standardise how care was provided, 2) that professionals would ‘buy in’ and align their work with new ways of working, and 3) that the new clinical posts would become catalysts for service changes at local level. At a national level, important contexts included varying levels of awareness about the programme, no plan for communicating service changes, and no established approach to implementation or professional oversight. Locally, experience delivering diabetes care, resource demands and familiarity with the intended purpose of the new clinical posts were important contextual factors. The extent to which integrated care was adopted and implemented depended on judgements made by health professionals (GPs, nurses, specialists and podiatrists) working in these contexts, specifically; judging the relative advantage of the programme and whether to engage in negotiations to legitimize their roles in diabetes care.

Conclusion Theory-based evaluations are better equipped to deal with the complexity of introducing multi-component interventions into dynamic health systems. This study suggests that, given a disconnect between responsibility for programme design and implementation, in the absence of systematic communication about the nature of changes and lack of clarity around governance and reporting structures, professionals used their judgment to adopt, implement and adapt interventions to match their priorities and circumstances.