By Amnon Shabo
November 15, 2003
| 'Medical record' commonly refers to the legally attested data in electronic clinical documents that are maintained by a healthcare provider. Typically, the provider uses its own medical archiving system in some proprietary format, which creates two challenges:
· Sharing and exchanging medical records between providers is made difficult because of different formats.
· Medical professionals lack a "longitudinal view" of patients' medical histories because information is maintained in independent silos.
The result: Patients can be placed at risk because medical professionals may have to rely on poor record-keeping or incomplete information when diagnosing ailments or prescribing treatments.
Attempts to address these challenges have led the medical industry to unite in an effort to create an electronic health record (EHR) standard — real-time, paperless files that provide a comprehensive patient history, including medications dispensed, diagnostic test results, medical images, and recent vital signs. EHRs offer three critical advantages over traditional medical records:
· A longitudinal view, letting medical professionals gain a complete picture of a patient's medical history
· Cross-organizational access, enabling different doctor offices, hospitals, laboratories,and other medical institutions to share information equally
· Patient data not necessarily logged during official medical treatments, such as self-documentation and personal genome information
Consequently, in many ways EHRs represent the final frontier for healthcare informatics. Much work has to be done, though, before the life sciences industry can realize the exciting new possibilities for standardized EHRs.
Partly because of competing models, standards-development organizations haven't yet produced acceptable specifications for EHRs. Health Level Seven (HL7), the most dominant healthcare standards organization in the United States, has created a specification for a clinical document format called Clinical Document Architecture (CDA). A collection of patient CDA documents could serve as a good basis for EHRs, but a problem is the redundancy and lack of coherency that would be prevalent in such an aggregation of records.
In April 2003, several U.S. agencies, including the Centers for Medicare & Medicaid Services, the Veterans Health Administration, and the Agency for Healthcare Research and Quality, sought to further this initiative by asking HL7 to produce an "EHR functional model and standard" aimed at improving patient care. In response, a newly formed body, the HL7 EHR Special Interest Group (SIG), has begun a project to harmonize a variety of EHR approaches and forge a functional model that will achieve broad industry acceptance.
Common Building Blocks for Bioinformatics
On a parallel track, standardization efforts are well under way in the fields of medical informatics and bioinformatics. The Clinical Data Interchange Standards Consortium (CDISC), for example, has been forging standards to describe all data involved in clinical trials. Recently, CDISC activists established an HL7 technical committee that is attempting to express CDISC standards in HL7 "language," with the aim of eventually creating an American National Standards Institute (ANSI) standard.
Furthermore, members of HL7 have initiated a new HL7 SIG to define all usages of genomic data in healthcare practice. Bioinformatics markup formats, such as MAGE-ML for gene expression and BSML for biological sequencing, are being utilized, similar to the way controlled medical vocabularies are referenced.
How do these bioinformatics-oriented efforts relate to the EHR standardization effort? The answer lies in a recent "methodology shift" that major standards-development organizations are adopting. HL7, for example, has produced a number of ANSI standards for healthcare messaging, but these standards have lacked a common language, and their development methodology has been inconsistent.
The new Version 3, however, represents a quantum leap with the emergence of a central Reference Information Model (RIM) that consists of common building blocks from which each committee can derive its specifications. These include core classes, such as acts, entities, and associations, all organized in a coherent framework called the Unified Service Action Model.
The result: a common language on which all HL7 standards may be based. Different health standards-development organizations will eventually agree on a common Health RIM, and this agreement will boost the usability of all specifications derived from such an RIM, enabling the desired goal of semantic interoperability.
This would represent one more step toward reaching the final frontier for healthcare informatics — a resource that could one day save time, money, and perhaps lives by avoiding medical errors with a virtual medical chart that would give any healthcare provider in any location the ability to instantly review a comprehensive, up-to-date patient history.
Amnon Shabo is a research staff member at the IBM Research Lab in Haifa, Israel. He may be reached at firstname.lastname@example.org.