INTELLIGENT SOFTWARE AGENTS THAT ACCESS PDQ

一位肿瘤学家正在考虑治疗一位病人的腹泻， 紫杉醇，可能会发现报告诱导伤寒有用。 智能 软件代理或中介可以跨越护理点之间的差距， 和相关报告。 参考紫杉醇和腹泻 基于计算机的患者记录（CPR），一种调解人可以 预测对这些报告的需求并在PDQ中搜索它们， 癌症，和其他知识来源。 调解员执行这些任务， 其他任务，通过使用域模型， 与用于获取知识来源的资源相比较。 第一阶段将侧重于开发和测试调解员原型 以及识别将被 需要让他们工作。 第二阶段将实施这些和其他调解员和所需的域 模型 域模型的维护将通过连接到 统一医学语言系统（UMLS） 临床信息系统（CIS）和CPR供应商现在认识到， 将知识带到护理点的商业价值：调解人 提供一种标准的、渐进的、可推广的方法， 这个目标。

An oncologist contemplating treatment for diarrhea in a patient on paclitaxel, may find reports of induced typhilitis useful. Intelligent software agents, or mediators, can span the gap between the point of care and the relevant reports. Given reference to paclitaxel and diarrhea in a Computer-based Patient Record (CPR), one kind of mediator can anticipate the need for these reports and search for them in PDQ, CANCERLIT, and other knowledge sources. Mediators perform these and other tasks by making use of a domain model that relates the attributes found in the CPR to those used to access the knowledge sources. Phase I will focus on the development and testing of prototype mediators and the identification of the features of the domain model that will be required to make them work. Phase II will implement these and other mediators and the required domain model. Maintenance of the domain model will be aided by connections to the Unified Medical Language System (UMLS). Clinical Information System (CIS) and CPR vendors now recognize the commercial value of bringing knowledge to the point of care: mediators offer a standard, incremental, and generalizable method of accomplishing this objective.