Investigating complexity in systematic reviews of interventions by using a spectrum of methods.

Systematic reviews framed by PICOS (Populations, Interventions, Comparisons, Outcomes, and Study designs) have been valuable for synthesizing evidence about the effects of interventions. However, this framework is limited in its utility for exploring the influence of variations within populations or interventions, or about the mechanisms of action or causal pathways thought to mediate outcomes, other contextual factors that might similarly moderate outcomes, or how and when these mechanisms and elements interact. Valuable insights into these issues come from configurative as well as aggregative methods of synthesis. This article considers the range of evidence that can be used in systematic reviews of interventions to investigate complexity in terms of potential sources of variation in interventions and their effects, and presents a continuum of purposes for, and approaches to, evidence synthesis. Choosing an appropriate synthesis method takes into account whether the purpose of the synthesis is to generate, explore, or test theories. Taking complexity into account in a synthesis of economic evidence similarly shifts emphasis from evidence synthesis strategies focused on aggregation toward configurative strategies that aim to develop, explore, and refine (in advance of testing) theories or explanations of how and why interventions are more or less resource intensive, costly or cost-effective in different settings, or when implemented in different ways.