Realist approaches to data generation

In spite of the inherent complexity of public health practice, practitioners are faced with both a requirement to demonstrate effectiveness and a dearth of tools to do so. The authors have been working on the development of realist approaches to evaluation for a number of years. This has proved a particularly successful methodological framework (Carr et al 2008a). The authors have used realist approaches in conjunction with appreciative enquiry (Carr & Lhussier (2008), with health impact assessment (Lhussier et al. 2008) and with soft system methodology (Carr et al. 2008b). This presentation aims to expose and encourage discussion about these methodological developments.Realist approaches encourage practitioners to populate a ‘context-mechanism outcome’ framework. This enables a surfacing of often implicit forms of knowledge, so that details of intervention mechanisms can be contextualised. From there, practitioners are encouraged to identify short, medium and long outcomes which they expect the intervention might achieve. This enables practitioners to establish evidence generating evaluation strategies to support intervention development. – In conjunction with principles of health impact assessment, a realist approach enabled practitioners to become explicit about why their project worked in the particular circumstances (Lhussier et al. 2008). – In conjunction with soft system methodology, realist methodology enabled practitioners to highlight the complexity of their combined intervention and potential synergies of action. (Carr et al. 2008b) – In conjunction with appreciative inquiry, realist synthesis enables a co-creation of understanding that builds on past success and frames them in a way that will facilitate replication (Carr & Lhussier 2008). Methodologically, these advances offer huge potential to practitioners faced with the necessity to generate evidence of effectiveness. They offer the possibility to frame past successes, to acknowledge local specificities and the often impossibility of achieving long term epidemiologically measurable impacts, and to frame and assess progress.