Increasing the Meaningful Use of Electronic Medical Records: A ...

Increasing the Meaningful Use of Electronic Medical Records: A Localized Health Level 7 Clinical Document Architecture System Jun Liang1, Mei Fang Xu1,*, Lan Juan Li2, Sheng Li Yang2, Bao Luo Li3, De Ren Cheng4, Ou Jin5, Li Zhong Zhang6, Long Wei Yang7, and Jun Xiang Sun8 1

Second Affiliated Hospital of Zhejiang University College of Medicine, Hangzhou 310000, China 2 Chinese Academy of Engineering, Beijin 100000, China 3 China Hospital Information Management Association, Beijin 100000, China 4 Zhejiang University, Hangzhou 310000, China 5 Hangzhou State Software Industry Base Co., Ltd., Hangzhou 310000, China 6 Hangzhou Normal University, Hangzhou 310000, China 7 The First Affiliated Hospital of Zhejiang University College of Medicine, Hangzhou 310000, China 8 Sir Run Run Shaw Hospital of Zhejiang University College of Medicine, Hangzhou 310000, China [email protected], [email protected]

Abstract. The health information systems of most medical institutions in China are isolated. Communications across these systems are generally realized through point-to-point interfaces at the database level, which tend to lack interoperability, extensibility, and security. In the resent study, we developed localized, document-oriented and data-focused clinical document architecture (CDA) templates based on health level 7 (HL7) CDA. Then, by combining these templates with the Service-oriented architecture for HL7 middleware, we accomplished interoperability across multiple heterogeneous systems. Modules of our system have been put into trial use in six medical institutions, including the Second University Hospital (main campus and regional campuses) of the School of Medicine, Zhejiang University and other collaborating hospitals. Keywords: HL7 CDA, interoperability, localization experience, SOA, structured data.

1 Introduction The development of hospital information systems (HIS) has been recognized as an imporatnt requirement in the reform of the health care system in China [1]. Although *

Corresponding author.

L. Cao, J. Zhong, and Y. Feng (Eds.): ADMA 2010, Part II, LNCS 6441, pp. 491–499, 2010. © Springer-Verlag Berlin Heidelberg 2010

492

J. Liang et al.

84% of the major hospitals in the country are now equipped with locally-developed online HIS, severe limitations remain, including low levels of system integration, high heterogeneity, and low application levels [2]. Trials with internationally-developed technically mature systems, however, have been unsuccessful due to differences in culture, work flow, and user habits [2]. In 2009, the Ministry of Health of China set the goal for HIS in Chinese hospitals as “sharing, not only processing and collection, of information using computers”. However, the chief information officers (CIOs) of many hospitals have not yet fully understood the essential role of standardization in the sharing of data across hospitals. Instead, some continue to maintain a closed-shop environment as a convenient shortcut around the establishment of an effective HIS [2]. Past experience in the United States has revealed that this route invariably results in hospital-specific solutions and highly complicated interfacing scenarios. As such, this approach can be considered non-viable overall [3]. At the same time, the experience from Marshfield Clinic Center in United States has shown that structured data form is going to be useful for future exploitation of the documentation [4]. The source of structured clinic data where those come from which will be two different ways. The first type of data is from clinic information data source which has already been structuring. Oppositely, the second type of data is from different data source which has not been structured. Those data have not been structured data would have need reclassification and post process in order to unified data from to be used. This part of the research work is though Localized Health Level 7 CDA template and minimized cost to structuring all clinic data. It is going to be shown as an undeniable truth that unified structuring clinic data is going to be a foundation for the secondary use which includes data mining for digital clinic record and the Interoperability across medical institution. The China Hospital Information Management Association (CHIMA) [5] has defined three steps for achieving successful sharing of health information, among which the standardization of interoperability is the fundamental step [6]. In this paper, we demonstrate the development of effective exchange and, thus, maximum utilization of health information stored in accord with HL7 CDA R2 [7]. In addition to standardization, the software/service architecture is also a major factor in determining the interoperability of across-institution medical software [8]. Service-oriented architecture (SOA) is an architecture style consisting of consistent but loose coupled components for the building of technology-driven, service-oriented software applications [9]. Web Services provide important methods of realizing SOA, and define mechanisms for providing interoperability between different software in a distributed heterogeneous environment. Section 2 analyzes the HL7 reference information model (RIM) and its vocabulary system. Section 3 describes the methodology of customizing CDA templates, and the development of an XML schema following the HL7 CDA standard. The schema can be used to create or verify CDA documents consisting of essential data elements and vocabularies. Section 4 discusses the transmission of documents consisting of electronic medical records (EMR) based on Web Services. Section 5 discusses implementation of the system and some experience. The last Section summarizes this work and recommends directions for future studies.

Increasing the Meaningful Use of Electronic Medical Records

493

2 RIM and Vocabulary HL7 RIM is a general, objective-oriented static model of health information that is not limited to the requirements of a specific subject area [10]. RIM includes six important base classes: Act, Participation, Entity, Role, Relationships, and Rolelink. Their interactions are illustrated in Fig. 1.

Fig. 1. A schematic showing the structure of RIM

Fig. 2. R-MIM for HL7 CDA Diagnostic Imaging Report; R-MIM is a subset of RIM, and includes class clones for the creation of CDA documents, and fully extended sets of attributes and relations

The data types in RIM and HL7 provides a powerful mechanism that allows CDA to incorporate concepts in other standard coding systems, such as the Systemized Nomenclature of Medicine Clinical Terms (SNOMED CT), Logical Observation Indenters Names and Codes (LOINC), and the International Classification of Diseases (ICD 9-CM). For applications in the health care system in China, the relevant codes must be translated, matched, and extended to meet the requirements of this system. Fig. 2 shows a part of a refined message information model (R-MIM), which describes the HL7 CDA Diagnostic Imaging Report.

3 Locally Customized CDA Templates and Information Architecture In the development of a locally customized CDA template, each field in the currently used target system was confirmed to follow the standards published by the Ministry of Health of China by checking against four standards set by the Ministry (“Rules for Health Information Dataset Classifying and Coding” [11], Metadata Specification of Health Information Dataset [12], the Rules for Data Element Standardization of Health Information [13], and the Guidelines for Data Schema Description of Health Information [14]), which were published as a framework for

494

J. Liang et al.

regulating the format of metadata and datasets in HIS. It is recommended that these guidelines be followed at all levels of health information subsystems. The metadata were thus updated and reviewed by a panel of experts on hospital computerization. General CDA templates that meet the requirement of the hospitals were generated following two of the methods recommended in HL7 [6]: Constraint method. The R-MIM model for the business process of the target subject area and the R-MIM model of CDA were combined to form a general CDA R-MIM model for the particular subject area. Or, if a suitable model was available in the HL7 template library, it was directly adopted. The work flow of the methodology of the constraint method is shown in Fig. 3.

Fig. 3. The methodology of the Constraint method

Transformation method. Clinical report documents observing international standards (e.g. RIS & PACS image reports following the DICOM SR Basic Diagnostic Imaging Reports standards) were transformed into CDA Diagnostic Imaging Reports, compatible with the CDA standards. The work-flow of this method is shown in Fig. 4.

Fig. 4. The methodology of the Transformation method


495

Subsequently, the relation between general CDA templates for various subject areas and the corresponding metadata in the original systems were mapped. This was a critical and difficult step, and we experimented with the two methods described below. 1. Each element in the template was matched to a metadatum. If a metadatum could not be matched, the template was revised by adding a corresponding CDA section. If a surplus element was found in the template, the match was determined by its frequency of occurrence. It was removed if it was considered to be “optional”. And it was unchanged if it was considered to be “required”. Then, a CDA schema for the subject area was generated using the HL7 V3 generator [15] based on the R-MIM of that area. Fig. 5 shows an example of element-to-metadatum matching. The resulting CDA Schema and XML documents are shown in Fig. 6.

Fig. 5. An example of the mapping between CDA elements and the corresponding metadata

Fig. 6. An example of CDA schema and XML document

2. The metadata were matched to the legacy system data structures. A scheme was extracted from the Legacy System and then matched to the CDA schema. This work primarily used first method as the basis for realizing the form structure of the clinical documents. Since HL7 CDA Level 2 had already have the identity of completed structuring [7], the compatible-HL7 CDA XML files which follow the methodology of HL7 would meet the requirement of “meaningful use” and Secondary use” for clinical EMR.

4 Transmission of EMR Documents Based on Web Service Exchange of information across heterogeneous health information systems was realized in a standard style based on SOA. The CDA documents were packaged using

496

J. Liang et al.

XDS to allow each heterogeneous system to continue functioning autonomously and independently. As a mature technique, Web Service provides a complete framework of services and has been commonly adopted in the development of SOA for distributed software [9]. In the SOA developed in this work, each heterogeneous system communicated with other systems via Web Services interfaces; the CDA documents were packaged following the XDS standard and exchanged following the HL7 specification. Based on these techniques, a framework of across-institution interoperability was developed corresponding to the current status of HISs in China see Fig. 7. The framework of the integrated SOA is shown in Fig. 8.

Fig. 7. A schematic showing the framework of interoperability across medical institutions

Fig. 8. A schematic showing the unified SOA framework; the design of the HL7 CDA Web Services adapter is similar to a simple Web Service interface, but the configuration file of HL7 Web Services and the HL7 methodology (RIMÆD-MIMÆR-MIMÆHMDÆMT) requires a top-down approach, in which the WDSL protocol is defined first and then the proxy and bottom codes are generated


497

The system was trialed in several hospitals, and the business processes and operation conditions in each heterogeneous system, including a subset of relevant data from other systems, were evaluated. The “engineer” mode was used, and a HL7 CDA adopter layer was added to the system without changing its internal structure or workflow. After a medical report was verified by the heterogeneous system, the HL7 CDA adopter then retrieved the relevant data from the business database and created a digitally signed CDA document by combining with an XSLT according to the specific requirement of the hospital. The CDA document was then packaged, submitted to a document database and registered.

5 Analyses of Experimental Applications In this work, we demonstrated the development of an EHR that is able to cover the life span of a patient and maintain and exchange medical documents in a patientcentered fashion. Our experiences in this work suggest that other types of CDA can also be successfully locally customized. In the future, we hope to extend the work to all clinical observation documents in 18 heterogeneous health information subsystems. Fig. 9 shows an example document. In terms of practical applications, the system is able to search the XDS-based registry for the ID code of the patient to obtain an index number. Based on this index number, the clinical document can then be retrieved in the document database. This document was created on XML. It follows HL7 CDA R2 standards, and is directly machine-processable [16] without the need for semantic transformation through additional interfaces. The complexity of system integration is shown below, with the original P2P mode for comparison: P2P Model: T(n) = n (n – 1)

SOA Model: T(n) = n – 1 .

(1)

where n is the number of the heterogeneous systems.

Fig. 9. An example showing the generation of CDA and applied XSLT based on a document from a hospital in China; the CDA Schema was mapped to relevant metadata and combined with XSLT to generate a document that meets the target clinical requirements (CDA version: CDA R2)

498

J. Liang et al.

6 Conclusion and Future Directions In this work, we realized the mapping between CDA Schemes of some clinical documents and the corresponding metadata using a flow process-based, document oriented, top-down approach and following the methodology “RIMÆ domain message information model (D-MIM) Æ R-MIM Æ Hierarchical message descriptions (HMD) Æ message type (MT)” of the HL7 standard. In future, we plan to extend our system to all clinical observation documents in 18 heterogeneous health information subsystems, improve the security and privacy of the CDA documents, and develop a unified patient identifier cross-referencing profile (PIX) control center based on the WS/PIDS standard. Our ultimate goal is to achieve a system of effective transmission of clinical documents across medical institutions based on structured CDA templates and to realize “Secondary Use” of clinical medical data. Acknowledgments. This research was supported by the key projects of the National Science and Technology program during the eleventh five-year plan period (Contract No.2008BAH27B01; Contract No.2008BAH27B03).

References 1. Chen, Z.: Overall Strengthening of Health Care Quality Management. Ministry of Health of the People’s Republic Of China (2008) 2. Li, B.L.: China Health White Paper. In: China Health Information Technology Conference & International Forum (2008) 3. Kraft, G.: Healthcare IT Standardization: Will the New Stimulus Package Help. IT Professional 11, 4–5 (2009) 4. Kate, K.: Marshfield Clinic Data Warehouse/Analytics. In: China Health Information Technology Conference & International Forum, pp. 16–20 (2009) 5. China Hospital Information Management Association, http://www.chima.org.cn 6. Li, B.L.: EHR Meaningful Use. China Hospital Information Management Forum, Hang Zhou (2010) 7. Dolin, R.H., Alschuler, L., Boyer, S., Beebe, C., Behlen, F.M., Biron, P.V.: HL7 Clinical Document Architecture, Release 2. Journal of the American Medical Informatics Association 13, 9–20 (2005) 8. Lewis, G.A., Morris, E., Simanta, S., Wrage, L.: Why Standards Are Not Enough to Guarantee End-to-End Interoperability. In: 7th IEEE International Conference on CompositionBased Software Systems, pp. 164–173. IEEE Press, New York (2008) 9. Newcomer, E., Lomow, G.: Understanding SOA with Web Services. Pearon Education, Upper Saddle River (2005) 10. Beeler, G.W.: HL7 Version 3 - An object-oriented methodology for collaborative standards development. International Journal of Medical Informatics 48, 151–161 (1998) 11. Ministry of Health of the People’s Republic of China: WS/T 306-2009. Rules for Health Information Dataset Classifying and Coding (2009) 12. Ministry of Health of the People’s Republic of China: WS/T 305-2009. Metadata Specification of Health Information Dataset (2009)


499

13. Ministry of Health of the People’s Republic of China: WS/T 303-2009. Rules for Data Element Standardization of Health Information (2009) 14. Ministry of Health of the People’s Republic of China: WS/T 304-2009. Guidelines for Data Schema Description of Health Information (2009) 15. Bonney, W.: HL7 Tools: The Comprehensive Guide. HL7 Tooling Committee (2005) 16. Dogac, A., Laleci, G.B., Aden, T., Eichelberg, M.: Enhancing IHE XDS for Federated Clinical Affinity Domain Support. IEEE Transactions on Information Technology in Biomedicine. 213–221 (2007)