Can theoretically informed realist approaches be successfully employed to synthesis multiple data sets and develop and evaluate a complex intervention? The 'Engager' synthesis saga

Realist informed approaches have been cited as being suitable for integrating multiple data sources and as appropriate for developing complex interventions. What, precisely, is meant by ‘realist informed’ approaches and... [ view full abstract ]

Realist informed approaches have been cited as being suitable for integrating multiple data sources and as appropriate for developing complex interventions. What, precisely, is meant by ‘realist informed’ approaches and how data from apparently incommensurable ontological approaches should be combined, remains less clear. We suggest that the development of theoretically informed intervention development and evaluation approaches, which move beyond inhabiting an a-theoretical middle ground, remain under theorised.
The Engager intervention supports prisoners with common mental health problems. The intervention was designed to engage with individuals with common mental health problems who are close to release, to set up a pathway of care in preparation for discharge, and to carry on for up to 8-16 weeks in the community. Within the wider context of a lack of suitable services for this population the Engager research project developed a theoretically informed intervention from a realist literature review, focus groups, case studies and the expertise of people with lived experience. The intervention was then tested in a pilot trial, which was evaluated by a realist formative process evaluation, additional focus groups, case studies and those with lived experience continued to contribute. The resultant intervention and it’s delivery platform are currently being evaluated in a realist process evaluation in a randomised control trial (RCT).

The presentation will interrogate the key challenges for theoretically informed realist methodologies that were experienced, reflected on and (hopefully) resolved within the first two phases of the synthesis. The issues addressed will include:
• Integrating a qualitative informed appreciation of absences and silences; realist informed methods lean towards prioritizing things that have happened. Learning to ‘listen’ for things that had not happened, or which we had not ‘heard’ was particularly beneficial.
• Allowing a range of data collection techniques to inform the overall realist focus. This allowed us to both reduce expansive approaches while also actively working against linear and reductive thinking.

The solutions arrived at included developing the following:
• A multi-stage iterative intervention development process.
• Rejecting realism=pragmatism and committing to the most appropriate method for the question being asked.
• Becoming comfortable with shifting from exploratory qualitative to applied realist approaches.
• Developing our understanding of how a critical realist approach, of what works for who, where, why, how and when, could be applied across data sets, regardless of the original ontological position from which they were produced; whilst appreciating that they might have additional findings of interest within their own paradigms.
• Recognising that our decision making processes needed to be explicit and include not only the research team, but also those who may receive the intervention and those who may deliver it. The titular reference to ‘saga’ is intended to highlight the role that an understanding of power dynamics contributes to developing complex interventions within real world settings.

We believe that, against the challenges of Marchal et al., we have produced a theoretically coherent Realist RCT of a complex intervention. We invite you to listen to our defence of this position and challenge us.