沃森具有理解人类语言、处理信息和找到准确答案的能力，拥有理解病人信息、给出可能的诊断或治疗等答案的潜力。其中，医疗保健面临的是医疗过程中不可避免的错误诊断问题（Graber，富兰克林&戈登，2005；Croskerry，2009）。生前诊断差异进行尸检率大约是20%到40%，第三都有准确的诊断是可以避免的（Gawande，2002）。诊断错误分为无过失误差、系统误差和错误的认知（Graber，戈登和富兰克林，2002）。最大限度地减少医生的认知错误引起的错误，计算机支持系统中的应用已被视为一个理想的解决方案（Graber et al.，2002）。几个计算机支持工具已经发展了几十年的dxplain（巴内特et al.，1987）和伊莎贝尔（Ramnarayan et al.，2007）。然而，看来，误诊率没有提高技术显著降低（Croskerry，2009）。有限的知识库，长期的咨询和诊断信息的限制这些系统的广泛使用要求高（Ramnarayan et al.，2007）。
沃森也许能够克服传统诊断或其他支持系统的一些缺陷，并帮助医护专业人员作出正确的诊断，甚至治疗和其他相关细节。与上述系统相比，沃森能够处理以人类语言表示的知识。再加上较高的处理速度，咨询时间将大大节省。沃森可以直接分析患者的病史，所以新的问题，这将有助于医生明确的情况下，症状可以提高和诊断不确定不考虑所有的可能性（Harvie &麦科夫，2012）。除了应用于诊断，沃森还可以产生治疗方案。
然而，Watson赢得了冠军的危险！实现上述目标还远远不够。第一个挑战是沃森是否能真正了解病史。病史包括病人投诉的描述、目前疾病的历史、身体系统的回顾和家庭史（卫生和人类服务部，2010），这比危险的内容复杂得多！其次，沃森的输出呈现给医生需要改进（Ferrucci et al.，2012）。由于沃森是用来帮助医疗保健专业人士避免认知错误，它需要解释为什么和如何产生某种诊断或治疗选项；如果不是这样，Watson得出结论，但医生不这样做，这可能会导致更多的错误
墨尔本代写assignment:沃森IBM计算机系统
Watson, a computer system developed by IBM scientists, is able to answer questions posed in human language with high speed and accuracy. In 2011, it won the championship in Jeopardy!, which is an American quiz show featuring multifarious questions in natural language. In its first appearance on Jeopardy!, Watson dealt with the similar amount of 1 million books to provide the most accurate result to a certain question (Harvie & McGoff, 2012). The success hints the extensive application of Watson in dramatically distinct and excessively specialized fields like business and healthcare. Healthcare is one of the most import aspects that Watson can be applied to.
With the ability of comprehending human language, processing information and finding exact answers, Watson owns the potential to understand information from patients and give out answers such as probable diagnosis or treatments. One of the problems that healthcare faces with is evitable diagnostic mistakes in healthcare process (Graber, Franklin & Gordon, 2005; Croskerry, 2009). The rate of antemortem diagnostic discrepancy detected through autopsy is about 20% to 40%, a third of which would have been avoided by accurate diagnosis (Gawande, 2002). Diagnostic errors are classified into no-fault errors, system errors and cognitive errors (Graber, Gordon & Franklin, 2002). To minimize errors caused by mistakes of physicians’ cognition, the application of computer-based support systems has been regarded as an ideal solution (Graber et al., 2002). Several computer-based support tools have been developed for decades including DXplain (Barnett et al., 1987) and Isabel (Ramnarayan et al., 2007). However, it seems that the misdiagnosis rate has not decreased significantly with improved technology (Croskerry, 2009). Limited knowledge base, prolonged consultation, and high requirement for diagnostic information limit the wide use of these systems (Ramnarayan et al., 2007).
Watson may be able to overcome some of the shortcomings of traditional diagnosis or other support systems, and help healthcare professionals give out correct diagnosis, even treatment and other related details. Compared with above mentioned systems, Watson is able to process knowledge which is presented in human language. Coupled with superior process velocity, consultation time will be highly saved. Watson can directly analyze medical history of patients, so new questions which will help physicians clarify conditions, symptoms and outcomes can be raised and a diagnosis will not be confirmed without considering all probability (Harvie & McGoff, 2012). Besides the application in diagnosis, Watson can also generate treatment options.
However, Watson which won the champion in Jeopardy! is far from enough to fulfill the above goal. The first challenge is whether Watson can really understand the medical history. Medical history is consisted of description of patient complaints, history of present illness, review of body systems and family history (Department of Health and Human Services, 2010), which are much more complex than the contents in Jeopardy!. Secondly, the outputs of Watson presented to physicians need to be improved (Ferrucci et al., 2012). As Watson is used to help healthcare professionals avoid cognitive errors, it needs to explain why and how a certain diagnosis or treatment option is generated; if not so, Watson makes the conclusions but physicians do not, which might lead to more errors.
