XML in Healthcare

Universities, governments, and industry organizations have been using XML for document storage and data integration for a decade. Database and software vendors, including IBM, recognized the potential of XML early on, so they helped develop XML-based standards and industry-specific markup languages, while working to integrate XML and SQL in their database platforms. IBM and Informix first released the IBM DB2 XML Extender, Informix Web DataBlade, and Informix XSLT DataBlade, and then integrated XML support more deeply into RDBMS servers, greatly advancing the integration of XML and SQL database processing.

XML is important to many industries, especially those with complex integration, archiving, and compliance requirements. The healthcare industry can use service-oriented architecture (SOA) and SQL/XML-enabled databases to build new applications while addressing compliance and standards issues to support electronic medical record systems.

The Rise of XML in Healthcare

XML's rise in the healthcare industry is due in part to laws such as the Health Insurance Portability and Accountability Act (HIPAA) that protect patient security and privacy. Enacted by the U.S. Congress to protect insurance coverage, HIPAA includes standards for electronic transactions and requirements for data privacy and security for claims, payments, benefits investigations, claim status, and other transactions. HIPAA also requires the U.S. Department of Health and Human Services to establish rules for the distribution of healthcare information.

Translating these legal requirements into applicable standards is often the work of standards development organizations; one of the most important standards development organizations in the healthcare industry is Health Level Seven (HL7). HL7 develops standards for the exchange of administrative and clinical data in healthcare, including claims processing, medical imaging, and pharmaceuticals (see the sidebar "HIPAA and HL7" at the bottom of this article). The healthcare industry has also developed specifications for operational data models and tabular models for research data (Clinical Data Interchange Standards Consortium). These standards apply to communications between internal systems and external entities such as the U.S. Food and Drug Administration.
New standards in a variety of industries, including healthcare, have promoted the development of XML-enabled applications. This new wave of technology means that we can build applications (often composite applications using SOA) that access medical data through a combination of interoperable services and rich database support.

However, managing large amounts of XML data efficiently is a challenge. IBM introduced pureXML technology in DB2 9, which allows documents to be stored, indexed, and queried in native XML format. Some healthcare organizations have used the native XML capabilities of IBM DB2 to build systems that not only comply with healthcare industry standards but also improve data access and performance.

Difficulties with Electronic Medical Records

Because medical data is often subject to frequent modification, large healthcare organizations rely heavily on XML. One such organization is the UCLA Health System, a multi-hospital healthcare provider that provides a variety of clinical and healthcare services. It includes Ronald Reagan UCLA Medical Center, Santa Monica UCLA Medical Center and Orthopaedic Hospital, Mattel Children's Hospital UCLA, Resnick Neuropsychiatric Hospital at UCLA, and UCLA Medical Group (including primary care and special care departments). More than 2,000 physicians perform more than 1 million outpatient visits and 80,000 inpatient admissions each year.

[page]

A healthcare system of this size must process a large volume of data every day, including medical record updates, lab results, MRI/CT/EBCT angiography images, admission/discharge/transfer data, and pharmacy orders. All of the data must be stored securely and reliably (and ultimately archived appropriately), and it must be accessible as needed. In addition, the data must be easily searchable, transportable, and organized by a variety of staff in multiple locations. Information also needs to be entered and updated easily and quickly, so the required IT infrastructure is very complex.

UCLA Health System meets these needs with the Patient Oriented Document System (PODS), an electronic medical record (EMR) repository capable of storing and retrieving more than 20 million documents. PODS is the source of patient information for UCLA's document management system, extended SOA (xSOA). Working with PODS, xSOA provides viewing interfaces for the GE BDM drug information system, the CliniComp Essentris emergency system, the Orion Soprano outpatient data system, and the forms portal. The xSOA Central Document Bus connects to the GE Picture Archiving and Communication System (PACS), outpatient applications and images, and the PODS repository. The Image Bus enables access to patient diagnostic images, and the Forms Bus handles the various electronic forms used by UCLA Health Services. The HL7 Message Bus provides HL7-compliant communications.
Using SOA in conjunction with the DB2 database, PODS enables 2,000 physicians and 3,000 nurses to access patient records. The system supports more than 400 electronic forms; these forms replaced 1,000 pages of paper forms previously used, helping to eliminate misunderstandings and omissions caused by handwriting. The database stores information on approximately 2 million patients, with 12,000 documents added every day, including new test results, doctor's prescriptions, and other patient-related data. PODS includes a document repository and metadata repository that can associate medical record files with a DB2 database. PODS stores medical record image files on file servers; it uses an IBM DB2 database to store the corresponding metadata and a network-attached storage array to store image files, including PDF and text.

When a document enters the system, it is stored on the file server and indexed in the DB2 database. Documents are not deleted because the PODS repository also serves as an archive of patient data. To ensure serviceability and 24x7 high availability, the PODS architecture includes redundant servers and databases, synchronized between database servers through data replication.

The DB2 database metadata store currently contains 30 million rows of information. Metadata is stored in XML format using the DB2 9 native XML engine. Dr. Charles Wang, architecture manager for UCLA Medical Center Computing Services, noted that PODS uses more than 400 schemas that conform to the W3C XML Schema language. The PODS software maps these schemas into a virtual schema for use throughout the system. The system uses a combination of four keys to create unique identifiers for documents and associated metadata.

To protect patient privacy, PODS is designed with a multi-level security model. In addition to DB2's built-in security features, the software architecture also includes document and metadata processors that integrate with security services. Role-based security and single sign-on capabilities are also provided. When using the file system to manage XML documents and schemas, issues such as security, concurrency control, parallel processing, and versioning must be specifically addressed, and using DB2 to store XML schemas and documents can easily address these issues.

UCLA Health System Based on SOA

[page]

The PODS implementation shows how SOA helps applications use underlying services; in this case, services for accessing patient information. Documents enter the system through the document service interface and are placed in a queue. The PODS architecture uses IBM WebSphere MQ for asynchronous messaging and queuing. For HL7 messaging, UCLA Health System uses Sun SeeBeyond eGate Integrator, which, according to Dr. Wang, is "an enterprise-wide standard used by all application interfaces."

The PODS queue manager controls the input queue, exception queue, and replication queue. Standard services for managing metadata and image files provide application programming interfaces (APIs) for uploading, downloading, querying, and updating documents. The system supports auditing and can generate reports for all database activities except uploads.

The Evolution of PODS

UCLA Health System PODS implementation uses DB2 pureXML capabilities to support access to patient documents, including loading, querying, and updating data. It provides a set of Web service interfaces to support the outpatient system to upload and query data.

The latest versions of this system are PODS3 and PODS4. Both versions are based on SOA, but use different DB2 capabilities for XML processing. For processing XML messages in the form of SOAP-based Web services, PODS3 and PODS4 use a combination of Systinet and IBM WebSphere software. However, the PODS4 implementation marks the transition from DB2 XML Extender to the pureXML capabilities of DB2 9. For example, DB2 9 introduces support for a feature defined by the SQL:2003 standard that allows XML column types as first-class data types. XML types can be used in Data Definition Language (DDL) statements, functions, and stored procedures.

Another benefit of migrating to DB2 9 is that the hybrid storage engine and query optimizer "understand" XML and can map queries involving XML into relational syntax. Migrating to PODS4 did not change the functional requirements of PODS, but pureXML technology simplified metadata processing and helped meet the response time and scalability goals of the system.

The upload process demonstrates the differences between PODS3 and PODS4. For example, when uploading a PDF document, the PODS3 upload process stores the document and XML metadata files on the EMC file server. It decomposes the metadata used by the DB2 SQL storage engine into an XCollection, which is a type implemented by the DB2 XML Extender. The upload process also uses the IBM WebSphere Application Server parser to validate the schema and index the paths to the PDF and XML files in the DB2 database. The DB2 transaction associated with the upload includes generating a unique document ID, recording the upload in the activity history table, and performing SQL INSERTs to 18 tables. The PODS4 upload process handles XML document metadata in a different way. Instead of storing metadata in 18 tables, the PODS4 upload process uses DB2 columns of type XML to hold the metadata, and only requires INSERTs to four tables.

Simplified processing and standards compliance

[page]

Migrating from PODS3 to PODS4 and DB2 9 has greatly simplified database management and replication tasks for UCLA Health System. The PODS3 architecture used DB2 8 and XML Extender, and the database contained 28 tables. The PODS4 database design uses XML columns, requires only 10 tables, and eliminates the 20 stored procedures used by PODS3.

Because UCLA Health System uses XML to store patient metadata, supporting new electronic forms in the PODS collection requires creating data definitions, or schemas, for the form content. DBAs who frequently work with new forms want to be able to efficiently define new data types, such as test results. Migrating to DB2 9 reduces the time required to add new forms and schemas to the system: supporting a new form in PODS4 typically takes only two hours, compared to two weeks in PODS3.

In any industry, standards are important for interoperability and efficient data exchange. XML has become an important tool for healthcare providers because it is an effective markup tool that defines the vocabulary required for data exchange and archiving. However, a robust application requires a reliable data management infrastructure. PODS shows how healthcare providers can solve the challenges of creating, storing, and exchanging electronic medical records. Through PODS, UCLA Health System and other healthcare organizations have begun to leverage XML technology to handle complex electronic medical records, meet HIPAA and HL7 standards, and improve productivity.

HIPAA and HL7

In 1996, HL7 established the Claims Attachment Working Group, whose mission was to standardize the information needed to process insurance claims, and subsequently developed a series of standards to meet HIPAA requirements. In the same year, HL7 began actively studying XML through its SGML/XML Special Interest Group. The initial results were six recommended attachments for claims processing.

Since then, HL7 has developed the messaging specifications Clinical Document Architecture (CDA) and Reference Information Model (RIM). HL7 has also developed standards for the electronic submission of CDA Public Health Case Reporting (PHCR) to state and local public health departments. As the standard has matured, XML has become an increasingly important technology. For example, the first version of CDA defined an XML architecture for the exchange of outpatient documents based on the XML Document Type Definition (DTD) included in the specification, using the HL7 RIM and HL7 registered encoding vocabularies to define its semantics. The upcoming third version of CDA uses only XML encoding.

"IBM DB2's native support for XML allows for storage of content in the healthcare industry standard HL7 CDA format," said Karla Norsworthy, IBM vice president of software standards. "IBM is committed to promoting interoperability and innovation in the healthcare industry through open standards. We have seen the positive impact that broad adoption of open standards, such as Java and XML, and healthcare standards developed by organizations like HL7, have on agility, time to market and innovation."