Thornton J G
Baillieres Clin Obstet Gynaecol. 1990 Dec;4(4):857-66. doi: 10.1016/s0950-3552(05)80349-9.
Decision analysis has its weaknesses: it may oversimplify problems, and probability and utility estimates may be biased. However, the alternative--decision-making by intuition--is likely to oversimplify problems even more and to be subject to even greater bias. Clinicians and patients often ignore major items of data, they handle probabilistic information badly, and their decisions are subject to a number of well-documented biases (Tversky and Kahnemann, 1974). Concern over these problems is not a reason to avoid decision analysis. Probably the greatest virtue of the decision tree and the other components of decision analysis is that the method forces decision-makers to make the bases for their decisions explicit. Generating the information often clarifies problems, and the source of difference of opinion can then be elucidated. Most of the benefits of decision analysis come to the decision-maker in the early part of the analysis as the problem is formulated and probabilities and values are estimated. Only occasionally is it necessary to perform new research, formally measure experts' opinions by the Delphi method, or do elaborate sensitivity analysis.
它可能会过度简化问题，而且概率和效用估计可能存在偏差。然而，另一种方式——凭直觉做决策——可能会使问题简化得更厉害，并且偏差可能更大。临床医生和患者常常忽略主要的数据项，他们处理概率性信息的能力很差，而且他们的决策存在许多有充分文献记载的偏差（特沃斯基和卡尼曼，1974年）。对这些问题的担忧并非是回避决策分析的理由。决策树及决策分析的其他组成部分最大的优点可能在于，该方法迫使决策者明确其决策依据。生成信息往往能澄清问题，进而可以阐明意见分歧的根源。决策分析的大部分益处是在分析早期，当问题被提出、概率和价值被估计时，就为决策者带来了。只有偶尔才需要开展新的研究、通过德尔菲法正式衡量专家意见或进行详尽的敏感性分析。
