Martin Than and colleagues provide a potential protocol for the most common and risk-filled issue in emergency medicine: the safe and appropriate discharge of emergency department patients with chest pain. Two obstacles are of note. First, a one-size-fits-all approach, despite being common, ignores Bayesian considerations and in low-risk groups results in poor yield. Than and colleagues address this by identifying a group whose post-test probability of an adverse event is low enough to consider foregoing further testing. The second obstacle is a pervasive beliefthattheacceptable "miss rate"for acute coronary syndromes is 0%.This belief ignores both the impossibility of such accuracy and the associated costs of pursuing such a miss rate. The solution to both problems is the "test threshold"-a prevalence of disease above which the potential benefits of testing outweigh the potential harms. For the assessment of potential cardiac ischaemia, this threshold has been estimated at about 2%.Thus if either the pretest or post-test probability of disease is less than 2%, testing will be more likely to yield harm than benefit. Than and colleagues have identified a clinically relevant decision aid that identifies a group below this threshold.
展开▼
