DERIVED THRESHOLDS IN MEDICAL DECISION MAKING

我们已经开发出一种新的方法来研究定量方面的 医疗决策。 概率是疾病的概率， 临床医生将决定使用诊断测试或启动 疗法 阈值概念以前曾被提出作为一种手段， 计算理论上的最优概率以指导决策。 这些阈值概率可以使用决策分析来获得。 技术. 相比之下，我们描述了一种推导医生的 (or管理临床决策的实际阈值， 描述性而非规范性方法。 我们的初步研究表明 显示了在使用诊断测试的患者与 疑似冠状动脉疾病的专业类型的医生。 的 阈值模型也应该有助于审查的重要性， 不同类型的病人。 我们现在有一个独特的机会将阈值模型应用于决策 涉及疑似肺栓塞的患者。 指导下 在尤金·桑格博士的指导下，核医学学会完成了一项 研究肺部扫描的临床实用性。 数据库包括 超过2000名患者，并包含所需的信息， 应用阈值模型。 因为桑格博士愿意 通过合作，我们可以将一个有用的模型应用于一个重要的问题， 临床数据已经收集。 我们将使用决策 研究患者特征的作用的阈值， 医生对疑似肺栓塞患者的管理。 具体来说，我们将测试医生有更高的假设， 测试和治疗的阈值： 1.老年患者与年轻患者相比， 2.已知患有癌症的患者与未患癌症的患者， 3.轻度与重度肺栓塞症状的患者， 4.没有保险的人比有保险的人多。 这项研究将提供一个更好的了解医生的决定 对疑似结核病患者进行诊断试验的方法和使用 肺栓塞

We have developed a new method for investigating quantitative aspects of medical decision-making. Thresholds are the probabilities of disease at which a clinician will decide to use a diagnostic test or to initiate therapy. The threshold concept has previously been advanced as a means of calculating theoretically optimal probabilities to guide decision-making. These threshold probabilities can be obtained using decision analytic techniques. In contrast, we describe a method of deriving a physician's (or group of physicians) actual thresholds governing clinical decisions, a descriptive rather than normative approach. Our preliminary studies have shown differences in the use of diagnostic tests for patients with suspected coronary artery disease by specialty type of physicians. The threshold model should also be useful for examining the importance of different types of patients. We now have a unique opportunity to apply the threshold model to decisions involving patients with suspected pulmonary embolism. Under the direction of Dr. Eugene Saenger, the Society for Nuclear Medicine has completed a study of the clinical usefulness of lung scans. The data base includes over two thousand patients and contains the information that is needed for applying the threshold model. Because of the willingness of Dr. Saenger to collaborate, we can apply a useful model to an important problem for which the clinical data has already been collected. We will use decision making thresholds to investigate the role of patient characteristics in physicians' management of patients with suspected pulmonary embolism. Specifically, we will test the hypotheses that physicians have higher thresholds for testing and treating: 1. elderly patients compared to younger patients, 2. patients with known cancer versus those who do not have cancer, 3. patients with mild versus severe symptoms of pulmonary embolism, 4. patients without insurance than for those with insurance. This research will provide a greater understanding of physicians' decision making and their use of diagnostic tests for patients with suspected pulmonary embolism.