Festa T, Foco A, Garbarini A, Enrichens F, Caponi R, Mao P, Grande L, Sciascia C, Pessione E, Brusa A, Marforio P, Jayme A, Giordano A
Minerva Med. 1983 Mar 17;74(11):539-46.
Bayes theorem; multivariance analysis, and the discriminating function of Mahalanobis applied to sixteen laboratory examinations employed in the diagnosis of diseases of the extrahepatic bile ducts and pancreas gave the following results: 1) the weight of each test was first measured in general terms, and then with regard to different levels of differentiation, with a classification in decreasing order in each case; 2) little redundancy was noted, and few examinations were of relatively negligible importance; furthermore, combinations of several tests did not lead to positive results; 3) reclassification according to the formula of Bayes and by means of the discriminating function proved sound supports for clinical decision, since they provided 70% and 80% of exact classification respectively; 4) the importance of the minicomputer in the construction of experimental diagnostic models was substantiated.
贝叶斯定理、多变量分析以及马氏判别函数应用于用于肝外胆管和胰腺疾病诊断的16项实验室检查,得出了以下结果:1)首先从总体上衡量每项检查的权重,然后针对不同的分化水平进行衡量,每种情况下均按降序分类;2)发现冗余较少,几乎没有检查的重要性相对可忽略不计;此外,多项检查的组合并未产生阳性结果;3)根据贝叶斯公式并借助判别函数进行重新分类,证明对临床决策有可靠的支持,因为它们分别提供了70%和80%的准确分类;4)证实了小型计算机在构建实验诊断模型中的重要性。