In a recent thought-provoking opinion piece in Emergency Medicine Australasia, Dubbo Emergency Physician, Dr Daniel Stewart reflects on rational decision making in current emergency medicine practice.
Drawing on decades of decision-making science, Daniel describes two contrasting approaches. The key difference is whether clinicians use Bayesian thinking—that is, whether they factor in the prior probability of disease when making diagnostic and management decisions.
Bayesian thinking underpins the threshold decision-making approach which enables emergency clinicians to use information gleaned from patient assessment to inform test and treatment thresholds. Bayesian thinking is best illustrated by the approach to a patient with suspected pulmonary embolism. Pre-test probability of pulmonary embolism is best estimated using a combination of clinical gestalt and clinical decision tools (like PERC and Wells’ scores). This probability then guides clinicians to the next appropriate step which could involve imaging (high-probability, D-dimer (intermediate-low probability) or no test (very low probability).
Daniel contrasts this threshold approach with relief-seeking decision making where tests are driven by clinician inability to tolerate the discomfort associated with diagnostic uncertainty. Daniel laments that this relief-seeking approach has infiltrated the practice of Emergency Medicine manifesting as front-loading (tests ordered before adequate assessment) and order sets (tests grouped for convenience rather than relevance). Perceived benefits to efficiency and patient flow have further embedded these strategies into practice jeopardizing our ability to nurture Bayesian thinking in the next generation of Emergency Physicians.
What this means for rural emergency medicine researchers
These decision-making frameworks apply across all emergency settings, but rural contexts may shape how they are used. Limited access to on-site or after-hours testing may shift clinicians towards more deliberate, probability-based decision-making—or, conversely, towards other, less rational approaches. This raises testable questions for rural research, such as whether clinicians in settings with fewer onsite diagnostic resources use fewer low-value investigations, and how does this affect outcomes? Comparing rural and urban decision-making patterns may help clarify when threshold-based approaches are naturally supported, and when system pressures drive alternative behaviours.
Daniel concludes by exhorting the Emergency Medicine community to transform Emergency Departments into learning environments where threshold-based decision making can flourish. This transformational change towards a diagnostic stewardship culture, where clinicians order the right test for the right patient at the right time, will be challenging. Emergency Physicians must overcome four interlinked environmental barriers to threshold-based decision-making—efficiency, culture, complexity, resources. How do we embed the threshold approach in emergency clinicians who are routinely called upon to provide contemporaneous, timely, emergency care to multiple undifferentiated patients amidst staff and bed deficits whilst being constantly interrupted by phone calls ?
Front-loading and order sets can be seen as cultural responses to an overwhelmed system trying to maintain efficiency. As Daniel rightly points out, the solution lies in redesigning ED physical and social environments to make threshold-based decision making the most obvious choice that is simultaneously easy, attractive, and satisfying. Making this work will require clinicians to design, test, and refine context-specific interventions, drawing on design thinking, systems science, behavioural science, and implementation science.