Objectives The use of laboratory tests is known to vary widely in many settings, including in primary care. Patterns of variation and an overall upward trend in testing have been reported in many observational studies. We conducted a realist review of the literature with the objective of seeking explanations for these observed patterns, in relation to laboratory testing in primary care settings. Our aim was to use the extensive existing literature on testing to develop evidenced theoretical explanations for the underlying causes of variation in testing, to improve our understanding of how and why differences in clinical practice in testing come about.

Method Realist review is a theory-driven, interpretive style of literature review that seeks to gather together and configure and re-interpret the existing evidence on a topic. An ‘initial programme theory’ describing potential explanations for clinicians’ decision making about test ordering was developed via scoping the literature and drawing on the experience of a wide range of stakeholders, including patients, clinicians and policymakers. Diverse evidence was then gathered via a systematic literature search, including quantitative and qualitative studies, and grey literature. This data was synthesised to refine the programme theory, and provide a series of evidenced causal explanations for variation in clinicians’ decision making, presented in terms of the important contexts that affect testing decisions, and the underlying mechanisms that cause them.

Results The evidence-based programme theory explains how test ordering can fulfil many roles in primary care settings, including as part of heuristics (rules of thumb) employed by clinicians to manage uncertainty, and as a social and strategic tool deployed to manage interactions with patients. The programme theory also offers explanations for how wider structures shape decisions, such that ordering tests may become a default ‘path of least resistance’ for clinicians. The literature points to effects that may emerge over the longer term, shifting norms, and tending to increase expectations of, and reliance upon, laboratory tests. In an overall context where time and cognitive resources are limited, there is a sense that decisions about laboratory tests are relatively trivial and inconsequential. Clinicians’ decision-making efforts are directed elsewhere, prioritising decisions about treatment and the maintenance of relationships with patients.

Conclusions When clinicians face high workloads and numerous competing priorities, they must decide where to direct their cognitive resources, and which clinical decisions require most time and effort. In designing interventions to change test ordering behaviour, researchers could adopt two opposing tactics. Strategies to disrupt and replace ‘bad’ testing habits with better ones by making changes to test ordering systems may minimise additional cognitive load, but the high-quality evidence needed to underpin such system changes is lacking. An alternative tactic is to attempt to prioritise a focus on laboratory testing and its potential negative consequences, aiming to shift testing norms, but this requires clinicians’ attention and will necessarily divert cognitive resources from other areas of practice. This review also demonstrates the value of the realist approach to generate useful and evidence-based findings that take into account the complexity of topics relating to overdiagnosis.