Since January 2019 I have been a Research Fellow on the Dementia-Person Aligned Care Team (D-PACT) project at the University of Plymouth; a project developing and evaluating a Dementia Support Worker (DSW) intervention based in primary care. D-PACT is using a realist informed approach to developing programme theory. My background was in Speech and Language Therapy (SLT) and lecturing. I had experience of qualitative research, but until D-PACT, did not know much about realist methodology. I am about to move on to a new post on another dementia project, and have been reflecting that over the last two and a half years I have learnt so much about this area. In this blog I hope to offer up just a flavour of this learning to researchers who are unfamiliar with realist enquiry and might want to explore its potential for their own research.All Posts
Historically, intervention evaluations have asked the question ‘Does this intervention work?’ i.e., if an intervention is implemented, is there a positive result? (A to B binary approach). Realist methodology is a theory-driven approach to developing and evaluating interventions, examining what happens in between A and B, recognising that people, not interventions, create change. An intervention is not going to work for all people, so the question is: ‘What works, for whom, how, and in what circumstances?’ Ray Pawson and Nick Tilley, influenced by the philosophy of critical realism, were originators of Realist Evaluation in the 1990s (see reading material at the end of this blog). Theories are usually constructed around Context/Mechanism/Outcome (CMO) configurations. Contexts are elements that pre-exist the intervention (e.g., environmental and cultural factors, policies, demographics, peoples’ existing belief systems, etc.). Mechanisms are both opportunities offered by the intervention (mechanism resources) and the psychological or behavioural responses to these resources (mechanism responses/reasoning). Outcomes are the result of an interaction between contexts and mechanisms. For example, Handley et al (2017) developed theory on how to improve dementia-friendly hospital care, expressed through a series of CMO statements, such as:
‘Where behaviours that challenge* are understood as communication of an unmet need (Context), through training, resources and support from experts in dementia care (Mechanism resource), staff will feel they have improved capacity and capability to influence the situation (Mechanism response/reasoning), making it more likely they will identify and address the need (Outcome).’
Developing realist theory involves synthesising a wide range of information, obtained usually through mixed methods; at the same time, placing a high value on qualitative approaches. Theory is developed iteratively. For instance in D-PACT, we developed an early version of our intervention theory: the mechanism resources offered by a Dementia Support Worker intervention, the mechanism responses these might lead to and the hypothesised outcomes. This was informed by literature and discussions with Patient and Public Involvement (PPI) contributors, as well as interviews and focus groups with various stakeholders (e.g., people with dementia, carers and field experts). This early version was articulated as a set of CMO type statements. Initial theory informed the development of a prototype written manual and practitioner training.
Later, initial theory was tested out in practice, through a feasibility study, with DSWs based in a small number of GP practices. Several further iterations of theory were then based on analysis of new data sources e.g., interviews with people with dementia and carers receiving the intervention, interviews and informal discussions with DSWs and their supervisors, and video/audio recordings of support interactions. ‘Realist interviewing’ techniques (Pawson, 1996; Manzano, 2016) can be used, whereby a researcher presents aspects of current intervention theory to interviewees for them to consider, relate to their own experiences and feedback whether the theory is valid or whether alternative explanations should be considered. This can be achieved with some people with dementia and carers where questions are not too abstract, for instance ‘We are thinking that the Dementia Support Worker could support people to keep doing the things they like to do. Has this happened for you?…What did this lead to?… How did it make you feel?’ etc. With practitioners, it is sometimes possible to be even more explicit about testing out theories, perhaps by showing them parts of CMO statements and asking for their feedback. Each time we refined theory, this informed refinements to the manual and training.
What I like about realist methodology
I like how realist approaches sit in the middle ground between two extremes on a spectrum of research paradigms regarding how ontology (reality) and epistemology (knowledge) are construed. At one extreme, positivists believe reality is out there to be discovered and can be measured. Deductive thinking is key; research starts from theory and tests whether observations match expectations. At the other end of the spectrum, for constructivists there is no single reality. People construct their own realities. Inductive reasoning is key; theories are built from the data. Realists sit somewhere between the two, believing that there is a reality out there, but how people construct their own realities is also of great interest. Realists accept that we may only ever partially uncover reality, but through testable theory and eclectic methods of enquiry, the aim is to reach the closest possible, good enough explanation.
I also like how, especially in the initial theory development stages, researchers can use their own hunches and creative imagination, combined with evidence, in order to create theories, which will later be tested out (and hunches either thrown out or evidenced).
Finally, I like how you can look back at early versions of theory and see how your knowledge and understanding of how an intervention might work/not work has grown, become more nuanced and more useful for informing real world practice, with all its complexities and ever changing contexts. Most healthcare interventions are complex, with many interacting components which don’t necessarily act in a linear fashion. Realist informed methods are ideal for developing new interventions and testing out how they might or might not work in a variety of real-world settings.
What I have found challenging
Although there are some papers which provide a guide to analysing data to inform realist theory building, this is an emerging field. My colleagues and I have had to explore how best to use data analysis tools like NVivo, to code data deductively and inductively against CMO-type theory statements. This process has been a bit trial and error however we are currently writing a journal article that will explain our analytic process and how we refined it over time, hopefully adding to the growing literature on realist analysis.
Whilst drawing on multiple data sources and other information, it can be difficult to keep track of where ideas have come from. Sometimes, a sudden clarity over something that was uncertain can arise from a discussion with a stakeholder or another researcher, or a piece of literature, and then this is incorporated into the intervention theory. It is important to keep a track of the reasons for all these little refinements at every stage, which is not always easy.
Throughout my post on the D-PACT project I have been on a realist adventure. I’ve attended some courses run by the Centre for Advancement in Realist Evaluation and Synthesis, which I have found really helpful and enjoyable. Some informative webinars as well as details on courses can be found on the CARES website: https://realistmethodology-cares.org/. I also joined a realist reading group, then got asked to step in and run the group when the usual facilitator went on maternity leave. Even though at that stage I felt a little out of my comfort zone, I said yes and didn’t regret it. Preparing for group meetings accelerated my learning no end, so I can highly recommend getting involved with a realist reading group or setting one up. I think I have now become a ‘realist thinker’ (I’m hoping someone will give me a badge) and will continue to develop my skills in applying realist principles as a dementia researcher involved in developing and evaluating complex interventions. This blog has covered only a snippet of the realist research world, but I hope it has inspired other researchers to find out more.
Dr Sarah Griffiths is a Research Fellow at the University of Plymouth. A former senior lecturer on a BSc Speech and Language Therapy course at Plymouth Marjon University, and with a background in speech and language therapy . Sarah now researching Primary care based dementia support at the University of Plymouth.
You can follow Sarah on Twitter Follow @sgriffiths1966
Emmel, N., Greenhalgh, J., Manzano, A., Monaghan, M., & Dalkin, S. (Eds.). (2018). Doing realist research. Sage.
Handley, M., Bunn, F., & Goodman, C. (2017). Dementia-friendly interventions to improve the care of people living with dementia admitted to hospitals: a realist review. BMJ open, 7(7), e015257.
Manzano, A. (2016). The craft of interviewing in realist evaluation. Evaluation, 22(3), 342-360.
Pawson, R. (1996). Theorizing the interview. British Journal of Sociology, 295-314.
Pawson, R. (2013). The science of evaluation: a realist manifesto. Sage.
Pawson, R., & Tilley, N. Realistic evaluation. 1997. London, California and New Delhi: Sage.
*Challenging behaviours is an outdated term. See https://onlinelibrary.wiley.com/doi/epdf/10.1111/jan.14787
This study is funded by the National Institute for Health Research (NIHR) Programme Grants for Applied Research (RP-PG-0217-20004). The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care.
What have been your experiences of engaging with patients using online communication tools? Reply in the box below.