Study And Practice Of EHR Standardization And Structuralization

Today digital healthcare has become the main direction in healthcare development. The Electronic Health Record (EHR) related studies and applications which concern create, syncretize, storage, transform, manage and utilize of clinical information have rapidly developed in the recently several years.In the typical healthcare provider, patient's vital medical information is scattered across medica records kept by many different caregivers in many different location. These separate, imprecise and unrelated informations may result in billions of dollars spent in health care that does not improve patient outcomes and even medical errors. EHR provide an answer for these questions.EHR provide easy approach to patient information, enhance decision support and reference data, decrease the chances of enos, and improve patient-provider communications. They can be a major tool in dealing with the quality of care and cost/time issues in medical practice. Furthermore they may enable a more patient-centered apprach to healthcare and patients will more responsibility for managing their own health and healthcare. Medical study, education, public health safety and policy development will benefit from EHR also.Standardization and structuralization is one of the key studies in EHR. Structured data is the requirement of computer processable, and standardization is essential for interoperability of systems and the sharing of data between organizations. As the clinical document is the main part of medical records, this dissertation analysed the three core elements (representation, data capture, manage and uitlize) of standard and structured clinical document.Firstly, this dissertation developed an clinical document representation architecture to fulfill the machine processable, human readable and allow authentication reqirements from different domains. A new representation was designed in the architecture, CDA R2 offers implementers the ability to combine structured entries and narrative block, and IETF/W3C XML Signatures add authentication, data integrity, and support for nonrepudiation to the clinical documents, and two standards are both satifiied.Then, a method which could create structured entries and reference narrative at same time was developed to create standard and structured clinical document. To solve the problem that standard terminology could not precisely reconstruct narrative context, this dissertation proposed an automatic context creation method which could create context for standard encoded structured data and reconstruct good narrative. Further introduced UMLS knowledge base to support standardization and proposed a novel normal method to create SDE which separated the knowledge modeling from SDE system implementation. Seperated medical knowledge model could be reused and service for other knowledge based application. Online terminology service cooperated with dynamic concept description SDE form could realize the dynamic encode clinical information with stardard terminology at runtime.In addition, to efficent storage, retrieval, management and analysis of EHR information, this dissertation designs a clinical documents repository based on XML DataBase. In this manage and storage schema, standardization and structured clinical data could be fully utilized. To take full advantage of narrative and unstructured format clinical information, this dissertation proposed two novel algorithms for term exract and negation detection in Chinese clinical documents to create terms and concepts index on documents. These works build a foundation for computer process and analysis of narrative clinical documents.Last, A Clinical Document Management System based on these solutions was implemented by the author and collegues in a live clinical environment. Integrated with other systems and modules, a full electronic medical record system for large healthcare providers was developed. For sharing documents in regional cooperative healthcare, the dissertation proposed a regional document sharing architecture base on IHE XDS. This implementation experience has fed back into the development of the approach and has provided "proof-of-concept" verification of our solutions' completeness and practical utility.The methods and practice in this dissertation draw a whole blueprint for clinical document representation, data creation, management and uitilization. These works will also provide significant reference for the further stardard and structured EHR construction.