Aim: The case is made for conceptualising guidelines as more flexible, interpretative pieces of information that contain a mixture of factual evidence, evidence that requires interpretation and evidence that will be shaped and moulded by the particular context into which it will be introduced. By acknowledging this way of viewing guidelines, practitioners and researchers are more able to understand what works in terms of successful implementation. This perspective differs from rating the type of evidence within guidelines according to the nature of the evidence.
Methods: The approach used was based on a standard process of formulating theoretical arguments: presenting the background, making and justifying the case by presenting evidence on successful guideline implementation strategies. Having presented the reasons why we should relook at the way we conceptualise and use guidelines, the paper then contrasts the evidence derived from research with examples of practitioners' accounts of managing the introduction of guidelines in practice. From this a number of conceptualisations are presented that offer new ways of constructing multifaceted interventions for guideline implementation.
Results: A set of arguments has been presented that provide background evidence, draw broad conclusions, address the ongoing issues around non-spread of new information and offer different ways of constructing research studies. The arguments are built on empirical and experiential evidence and refer to several major studies.
Conclusions: Seven conclusions are drawn from the paper: guidelines are not 'literal' objects - they are complex communication vehicles that are used as catalysts to stimulate discussion, learning and debate across knowledge boundaries; current evidence on guideline implementation reinforces the need to look at complex, multifaceted interventions based on specific diagnosis of barriers to implementation; this process poses conceptual, theoretical and methodological challenges to the research community; research designs ought to consider the contribution of process models as well as causal models; the implication of such a move would be investment in more process studies before the setting up of expensive causal or intervention (RCT type) studies; refocusing of implementation research away from theories of behaviour change to more consideration of knowledge management is to be encouraged and organisational theory and theory formulation and testing ought to be heterogeneous rather than narrowly focused.