Essin D J, Lincoln T L
LAC+USC Medical Center.
Proc Annu Symp Comput Appl Med Care. 1994:509-13.
Information gathered during the healthcare process is lost when forced into rigidly structured record-oriented databases. By contrast, content can be difficult to manipulate if stored as unstructured text. Spurred by the requirements of electronic publishing, military procurement and the Internet, new robust standards for structuring documents have been developed and deployed. These standards can provide a foundation for a document-based Electronic Medical Record System. In order to fully exploit this added flexibility, an information model is necessary to define both the direct and contextual content of documents. Once context, as well as fact, are recorded in formal structures, inferential techniques can either selectively extract knowledge and data from documents or aggregate data to create summaries so that all interested and authorized parties have a better chance of meeting their information needs from a single, permanent data source.
在医疗过程中收集到的信息,若被强行录入结构严格的面向记录的数据库,就会丢失。相比之下,如果将内容存储为非结构化文本,又会难以处理。受电子出版、军事采购和互联网需求的推动,已制定并部署了用于构建文档的强大新标准。这些标准可为基于文档的电子病历系统奠定基础。为了充分利用这种额外的灵活性,需要一个信息模型来定义文档的直接内容和上下文内容。一旦上下文以及事实被记录在正式结构中,推理技术就可以从文档中选择性地提取知识和数据,或者汇总数据以创建摘要,这样所有感兴趣且获得授权的各方就更有可能从单一的永久数据源满足其信息需求。