Developing Clinical Guideline Ontology And Electronic Documents For Hypertension

Clinical decision support systems that integrate medical knowledge,healthcare information and reasoning engine can provide clinicalrecommendations for specific patients and be employed as important tools inperforming evidence-based medicine. Developing and deploying intelligentchronic disease management systems with clinical decision support functions willenable the accurate assessment and safe, effective intervention. Clinical practiceguidelines are the main knowledge bases of clinical decision support, and recordsof health related information of individuals are fundamental to make informeddecisions. Healthcare related information of patients, which collected byelectronic health record systems during care delivering must be retrieved andanalyzed in decision making of computer systems. Therefore, sharable electronichealth records for clinical decision support are critical play very important rolesin computerized clinical practice guidelines. The core issue is the way of healthinformation standardization during EHR and CDSS development. The objectives of this research is to investigate the standardization issues indevelopment of computer interpretable guidelines and electronic healthdocuments by building clinical guideline ontologies and health record documentmodels with standard modeling tools. From such investigation, this research triesto harmonize the standardization approaches of both CIGs and health records. Inaddition, by reviewing, analyzing, modeling and specifying of an exampleclinical guideline, we expect to identify the information standard needs indeveloping clinical guidelines and to provide suggestions from informationstandard perspectives for representation and organizing of electronic healthrecords.In this study, we took the management of hypertension for an example,based on the Hypertension Prevention and Intervention Guideline, referred to themethodology of SAGE guideline ontology development, and adopted Protégé asthe ontology development environment to create the guideline ontology. As aresult, we display the rules of Hypertension Prevention Guide with a set offlowchart. The data items were extracted from the flowchart node and grouped bythe category of HL7vMR. With the information in vMR, we establisheddocument for the patients with hypertension by referencing to HL7CDAtemplates, and produced the style sheet using XML editor. Using related recordforms in the "National Public Health Service Specification" as a health recordinstance, we made comparison to the electronic health documents developedthrough above process to find differences and potential gaps in informationstandardization.The main results of this research are as follows:(1) Taking "Hypertension Prevention Guideline (2009Edition)"as an object,32flowcharts were produced by refining and coding the rules in guideline and extract82data items from the flowchart nodes, which were organized into fourclasses of vMR, including7data items in class of Goals,43in Observations,10in Problem and22data items in class of SubstanceAdministrations.(2) The structure of electronic health documents for patients receivinghypertension related healthcare was designed in accordance with HL7CDA andits templates. The XML Schema and XML document of electronic healthdocuments for patients with hypertension were built by XML editor and stylesheets for the documents formed at last.(3) By comparing the style sheet of electronic health documents producedfrom hypertension guideline modeling to the related sheet for data collection inNational Public Health Service Specification, differences in scope of informationcollected, definition and representation of data recorded were identified.Suggestions of addressing synchronously data standardization in both guidelinesand health record documents developments were proposed to enable theimplementation of decision support systems.The main conclusions of this research are as follows:（1）To support computer implementation of clinical guidelines, we can use aclinical guideline as starting point and build a guideline ontology by reviewingand coding of clinical guideline. The related clinical information can beabstracted, modeled and standardized from the ontology, and then electronichealth documents built from the information, ensuring the content of electronichealth documents to meet the data needs of the clinical decision making and thesemantic and representation keep consistent with those in clinical guideline.（2）Modeling and expressing of the patient data in both CIGs and clinicaldocument can be made by a same standard system such as HL7, enabling theretrieve and share of patient data in clinical decision support systems. This will lay foundations for interoperability between EHR systems and CDSS.（3）To ensure computers to interpret clinical guidelines correctly, thesemantics and representations of clinical concepts must be defined accurately.Therefore, the descriptions in guideline texts need to be specified from bothdomain and information views. Meanwhile, the standardization of patients’documents should be harmonized with clinical guidelines.Focusing on computerized application of clinical practice guidelines, thisstudy conducted the modeling of guideline ontology and health record documents,providing a possibly feasible way for consistent expression of patient informationin computer interpretable guideline and electronic health document. Additionally,this research addressed the standard issues of health data in development of bothclinical practice guidelines and health record documents, providingrecommendations on coping with those issues. This work is expected to behelpful in integrating data from heterogeneous information systems in CDSSimplementation.