Division of Hematology and Oncology, Beth Israel Deaconess Medical Center, Boston, Massachusetts, USA.
Acad Med. 2010 Mar;85(3):556-63. doi: 10.1097/ACM.0b013e3181ccd59b.
The concept of thresholds plays a vital role in decisions involving the initiation, continuation, and completion of diagnostic testing. Much research has focused on the development of explicit thresholds, in the form of practice guidelines and decision analyses. However, these tools are used infrequently; most medical decisions are made at the bedside, using implicit thresholds. Study of these thresholds can lead to a deeper understanding of clinical decision making. The authors examine some factors constituting individual clinicians' implicit thresholds. They propose a model for static thresholds using the concept of situational gravity to explain why some thresholds are high, and some low. Next, they consider the hypothetical effects of incorrect placement of thresholds (miscalibration) and changes to thresholds during diagnosis (manipulation). They demonstrate these concepts using common clinical scenarios. Through analysis of miscalibration of thresholds, the authors demonstrate some common maladaptive clinical behaviors, which are nevertheless internally consistent. They then explain how manipulation of thresholds gives rise to common cognitive heuristics including premature closure and anchoring. They also discuss the case where no threshold has been exceeded despite exhaustive collection of data, which commonly leads to application of the availability or representativeness heuristics. Awareness of implicit thresholds allows for a more effective understanding of the processes of medical decision making and, possibly, to the avoidance of detrimental heuristics and their associated medical errors. Research toward accurately defining these thresholds for individual physicians and toward determining their dynamic properties during the diagnostic process may yield valuable insights.
阈值的概念在涉及诊断测试的启动、持续和完成的决策中起着至关重要的作用。许多研究都集中在明确阈值的开发上，形式是实践指南和决策分析。然而，这些工具很少使用；大多数医疗决策都是在床边做出的，使用的是隐含阈值。对这些阈值的研究可以深入了解临床决策。作者研究了构成个体临床医生隐含阈值的一些因素。他们提出了一个使用情境引力概念的静态阈值模型，以解释为什么有些阈值高，有些阈值低。接下来，他们考虑了阈值错误放置（校准错误）和诊断过程中阈值变化（操纵）的假设效果。他们使用常见的临床场景来演示这些概念。通过对阈值的校准错误进行分析，作者展示了一些常见的适应不良的临床行为，尽管这些行为在内部是一致的。然后，他们解释了如何通过操纵阈值来产生常见的认知启发式，包括过早结束和锚定。他们还讨论了尽管穷尽了数据收集，但仍未超过阈值的情况，这通常会导致可用性或代表性启发式的应用。对隐含阈值的认识可以更有效地理解医疗决策的过程，并可能避免有害的启发式及其相关的医疗错误。针对个体医生准确定义这些阈值并确定其在诊断过程中的动态特性的研究可能会产生有价值的见解。
