As was pointed out in the pages of this publication not too long ago (From the Editor, April 2007), many healthcare technology vendors claim their systems help facilitate interoperability. As a matter of fact, while at the HIMSS Annual Conference in February, I visited the Vendor Directory kiosk and conducted a keyword search on the term, which produced a list of more than 90 companies—about 10 percent of the exhibitors—asserting their systems provided interoperability.
As was pointed out in the pages of this publication not too long ago (From the Editor, April 2007), many healthcare technology vendors claim their systems help facilitate interoperability. As a matter of fact, while at the HIMSS Annual Conference in February, I visited the Vendor Directory kiosk and conducted a keyword search on the term, which produced a list of more than 90 companies—about 10 percent of the exhibitors—asserting their systems provided interoperability.
It is obvious those invested in health information technology recognize that interoperability represents the next significant advance in the industry. What’s less clear, however, is how each of these players defines the term.
Definitions
Currently, three definitions are typically used to describe the overall concept of interoperability, and each represents a progressively advanced level of functionality:
Interface engines. Many have used the terms “interface” and “interoperability” interchangeably—albeit erroneously. In reality, an interface engine routes information from one system to another, but stops short of enabling systems to use the data that has been transferred. A common example is an interface between a laboratory system and a billing system. When lab work is performed, the former sends a message to the latter so a claim or bill can be generated. These systems represent an important step forward by ridding organizations of the need to build and maintain many point-to-point interfaces between the discrete systems.
Interoperability. Systems that are truly capable of interoperability enable data to be transferred or shared in a meaningful way—typically in conjunction with other data. In other words, the receiving system can consume and assemble information from the transmitting system. For instance, technology that exhibits true interoperability can create a composite list of all lab results associated with a specific patient, even if the labs reporting the results rely on disparate systems.
Semantic interoperability. Representing the most sophisticated stage of development, semantic interoperability allows two systems not only to share information, it also enables the receiving system to understand and make use of the incoming data while maintaining the original “meaning” of that data. Labs may report results for the same tests using different nomenclatures, for instance. Semantic interoperability will coalesce the values, empowering providers to make optimal use of the information. In the context of service-oriented architectures, true semantic interoperability also enables systems and services to make use of software services provided by other systems.
Optimizing the Value of Shared Data
Clearly, the concept of semantic interoperability offers the greatest promise for the healthcare industry. It will enable providers to analyze trends, put discrete sets of data and services into a larger context and, ultimately, provide improved diagnostic and therapeutic services.
When measuring diabetic quality markers, for instance, one laboratory system may denote levels as glycohemoglobin within the chemistry panel, another may represent it as an independent test called “HgbA1c” and a third may describe it as using an appropriate LOINC code, for example 4548-4. Semantic interoperability allows the receiving system to recognize each of the three terms as representing the same value (i.e., glucose levels). As a result, providers can use functionality to make use of the common meaning—generating a trending analysis of glucose readings over time—regardless of the fact that the data originated in very different formats.
Experts Disagree On Approach
Few providers will argue the benefits that semantic interoperability can offer. Progress towards achieving this level of functionality is often impeded, however, by experts who can’t agree on how to achieve it.
On one end of the spectrum, many say that semantic interoperability requires perfect and complete harmonization of standards between technology systems. They recognize the great variations among technologies and believe these must be fully resolved to ensure reliable access to, and accuracy of, shared data
At the other extreme, some maintain that semantic interoperability can be achieved simply by mapping codes from one system to another. This approach is somewhat misleading and represents an oversimplification of the challenge.
The Middle Ground
In reality, the answer lies somewhere between these two extremes. Those subscribing to the former approach fall prey to “paralysis by analysis,” taking no steps towards semantic interoperability until all conditions are perfect. While some abilities to translate information from one system to another have already been developed, healthcare is nevertheless a long way from achieving the ideal. Therefore, healthcare leaders who advocate sitting back and waiting for the creation of flawless standards do a grave disservice to the industry and, ultimately, to patients.
Experts supporting the latter point of view, on the other hand, demonstrate a significant level of naiveté, since some code systems cannot be translated or mapped directly to another. A system like SNOMED, for example, cannot transfer data into a system like ICD-9-CM without losing specificity.
To be more pragmatic one must recognize that, within specific domains, a set of baseline codes to which all systems are mapped may be available. Even in cases where only part of the codes can be mapped to a baseline, this limited subset may have considerable value from a clinical and information technology perspective. These baseline codes can then form the common language for all variations of different representations of results—allowing them to be mapped to a baseline without requiring that the contributing system change the way it represents the inherent value of the results.
While this potential is not yet available within or between all types of code systems, we nonetheless have sufficient tools and technologies to begin the voyage towards full semantic interoperability today—delivering real and considerable value to providers and patients alike. As an industry, we need to ensure that practicing clinicians on the front line have access to these practical interoperability tools. An ever-failing healthcare system cannot afford to wait for “stars to align.” Similarly, an oversimplified approach will not adequately address an increasing demand for patient safety, quality improvement and a change in healthcare economics.
I believe that we can attain many of these lofty goals today by: implementing technologies that deliver on the levels of semantic interoperability already achievable; and, building the foundation to make use of the continuous improvement in the levels of terminology standardization that we as an industry deliver in the future.
Ilan Freedman is vice president of marketing for dbMotion. Contact him at [email protected].
