Systematic Reviews: Beyond Cochrane to Complexity

Abstract

From the perspective of complexity science the deterministic approach of Cochrane systematic reviews appears very restricted, seemingly missing the complex interaction between social and environmental context, individuals and social groups, health care providers and specific health interventions. However, the Cochrane Collaboration has established a rigour for evidence review of the effect of interventions. This rigour is applied to research questions that use peer reviewed published empirical evidence as their data source. This paper considers the role of Cochrane style reviews for our complex world and compares this with a realist approach to the review of research literature, which seeks to understand how and why an intervention works in particular contexts. Both recognise complexity, both include the identification of what the intervention is intended to do, both recognise the importance of context, and rigour is equally important in both. However, Cochrane reviews limit the reach of the review to specific contexts whereas realist reviews seek to understand how interventions work in differing contexts. Cochrane reviews seek to include studies of interventions with similar components. These components are context specific manifestations of the theory on which the intervention is based. Realist reviews seek to understand how theoretical mechanisms are manifest in different contexts. For collaborative international research a Cochrane review may need to be undertaken separately for contexts that are very different such as high and low income countries. However, the realist review itself seeks to understand how context makes a difference to outcome, and is therefore strengthened by comparison of different contexts. Cochrane style and realist reviews can be complimentary in advancing our understanding of our complex world and how complex interventions have their effect. Further research will be stimulated through comparative realist reviews that identify the variation in how interventions work or not across very different contexts.