Guidance on guidelines: Understanding the evidence on the uptake of health care guidelines

Regardless of health issue, health sector, patient condition, or treatment modality, the chances are that provision is supported by "a guideline" making professionally endorsed recommendations on best practice. Against this background, research has proliferated seeking to evaluate how effectively such guidance is followed. These investigations paint a gloomy picture with many a guideline prompting lip service, inattention, and even opposition. This predicament has prompted a further literature on how to improve the uptake of guidelines, and this paper considers how to draw together lessons from these inquiries.

This huge body of material presents a considerable challenge for research synthesis, and this paper produces a critical, methodological comparison of 2 types of review attempting to meet that task. Firstly, it provides an overview of the current orthodoxy, namely, "thematic reviews," which aggregate and enumerate the "barriers and facilitators" to guideline implementation. It then outlines a "realist synthesis," focussing on testing the "programme theories" that practitioners have devised to improve guideline uptake.

Thematic reviews aim to provide a definitive, comprehensive catalogue of the facilitators and barriers to guideline implementation. As such, they present a restatement of the underlying problems rather than an improvement strategy. The realist approach assumes that the incorporation of any guideline into current practice will produce unintended system strains as different stakeholders wrestle over responsibilities. These distortions will prompt supplementary revisions to guidelines, which in turn beget further strains. Realist reviews follow this dynamic understanding of organisational change.

Health care decision makers operate in systems that are awash with guidelines. But guidelines only have paper authority. Managers do not need a checklist of their pros and cons, because the fate of guidelines depends on their reception rather than their production. They do need decision support on how to engineer and reengineer guidelines so they dovetail with evolving systems of health care delivery.