There are few published examples of realist syntheses and those that exist do not include a detailed account of the approach used because authors tend to focus on the dissemination of findings within publications [6, 7, 24–27]. This lack of information about application of the realist approach is unhelpful to a novice realist reviewer. To fill this gap, this paper presents in some detail the approach we took to conduct a realist synthesis of evidence about the effect of change agency on evidence-informed healthcare. Change agency is a complex implementation intervention, which made realist synthesis an appropriate approach for unpacking its effects within different contexts and groups. However, undertaking this review was not without its challenges, not least because of the practicalities of working as an unfunded and geographically dispersed group.
Our approach deviated in some respects from the approach recommended by Pawson et al. For example, Pawson et al. do not advocate a comprehensive literature search, or double reviews and data extraction. In this sense we developed a hybrid approach that was fundamentally rooted in realist synthesis philosophy and principles (i.e., theory led, purposive, iterative, stakeholder involvement), but which also drew on some of the practices of traditional systematic reviewing.
One of the strengths of realist review is the approach’s firm roots in philosophy and social sciences [2, 9]. Rather than being a method or formula, it is a ‘logic of enquiry’ , which enables a flexible, all-embracing approach to explanation of what works for whom in what circumstances and in what respects. Rather than controlling for real life events, realist synthesis provides a framework for working with and untangling the complexity of real-life implementation. This allows for an equal focus on what works, as much as what does not work, in an attempt to learn from failures and maximise learning across policy, disciplinary and organisational boundaries. Furthermore, realist synthesis is inherently stakeholder driven, which facilitates engagement and the inclusion of multiple perspectives.
The strengths of realist synthesis underpin its limitations. Realist synthesis is premised on a set of principles rather than a formula, and whilst this allows for flexibility and inclusivity, it means that the findings from a review are theoretically transferable. For example, it follows that if the appraisal and data extraction needs to be bespoke to the particular review questions that arise from the theoretical framework, these will be different for each review. Furthermore, given that the fundamental interest in realist synthesis is about finding out what works in what contexts, the recommendations one can make will not be generalisable. Rather a realist review results in findings that are theoretically transferable; ideas (‘theories’) that can be tested in different contexts, with different stakeholders.
Pawson et al.  suggest that realist syntheses are not for novices. Unlike a Cochrane review, for example, which relies on standardised protocols and tools, the demands on a realist synthesiser are different. For example, quality assurance within realist synthesis is dependent on reviewers’ explicitness and reflexivity. During this review, we kept a log of the process and decisions made throughout the process, which we developed into a technical report. In addition, we undertook a more formal reflective process during the review because members of the group had varying experiences of realist review. This involved reflecting on questions about what was going well, what was going less well, as well as engaging in group learning activities. Throughout the review process, we had large and small group discussions that provided the opportunity for building in checks and balances, and for explicating processes. In turn, this requires a high level of expertise in reasoning, research methods and quality appraisal, and expertise in the subject area. The complexity of the realist synthesis approach means that it is time-consuming and human resource intensive, and for those reasons a potentially expensive endeavour.