Schewe S, Krüger K, Herzer P, Schattenkirchner M
Medizinische Poliklinik der Universität München.
Z Rheumatol. 1991 Jul-Aug;50(4):211-5.
409 unselected outpatients were asked standardized expert system questions on the medical history of joint complaints. An average number of 30.4 +/- 6.6 out of 60 questions with 346 different answers were asked in a mean time of 20.1 +/- 6.6 min. This time, required for the questions to be displayed on the computer screen and input assistance from a staff-member, proved to be comparable to the time required for a medical history to be taken on the same patients by a physician. In 75% of the cases the computer was able to predict the correct diagnosis on behalf of information of the medical history only. Using only medical history as a source of information, computer and physician had the same rate of wrong diagnoses. Before and after the standardized medical history on the computer screen all patients were asked about their opinion of the computer as a diagnostic aid. Male and younger patients more often appreciated computerized diagnosing. Just a few patients felt that their relation to the physician was disturbed by the computer or were inconvenienced themselves. Most patients stated they did not believe that physicians could ever be replaced by computers. Only a few of them had resignations about storage and statistical utilization of personal data.
409名未经挑选的门诊患者被问及有关关节疾病病史的标准化专家系统问题。在平均20.1 +/- 6.6分钟的时间内,从60个问题中平均提出了30.4 +/- 6.6个问题,这些问题有346种不同答案。事实证明,将问题显示在电脑屏幕上以及工作人员提供输入协助所需的时间,与医生为同一患者采集病史所需的时间相当。在75%的病例中,仅根据病史信息,计算机就能预测出正确诊断。仅将病史作为信息来源时,计算机和医生的误诊率相同。在电脑屏幕上显示标准化病史之前和之后,所有患者都被问及他们对计算机作为诊断辅助工具的看法。男性和年轻患者更常认可计算机诊断。只有少数患者觉得计算机干扰了他们与医生的关系或给自己带来不便。大多数患者表示,他们不相信医生会被计算机取代。只有少数人对个人数据的存储和统计利用表示不满。