As health care institutions continue to implement electronic medical records (EMR) across the country, finding out how to unite disparate information systems across a community is becoming increasingly more important in achieving the Obama administration's vision of health information technology in America. Vanderbilt University Medical Center CIO, Dr. William Stead, and Informatics Corporation of America CEO, Gary Zegiestowsky, spoke with DOTmed about interoperability and the universal health record.

Based on two national surveys supported by The Office of the National Coordinator for Health Information Technology (ONC), hospital adoption of full, application-rich EMR is at about 2 percent and ambulatory center adoption is at about 4 percent. Slightly more substantial are physicians' offices. About 17 percent of physicians have implemented some basic form of EMR, and approximately 8 percent of hospitals have a basic EMR that at the very least allows doctors to order medications and view laboratory and radiology reports electronically.

For the few institutions that have implemented EMR, the next step is leveraging that technology to form larger networks that allow health care workers to pull patient information and review data about broader disease populations. This interoperability not only accelerates patient care by enabling detailed medical information to follow patients across community hospitals and clinics and can reduce redundant procedures, it also allows regional health care organizations the ability to better gauge, manage and treat specific disease states that are common within a community.

In 2000, a few Vanderbilt University Medical Center informatics programs began working on two complementary software applications dubbed StarChart and StarPanel, which aggregated and organized medical data, as well as improved communication and clinical decision-making within a single interface. Dr. Stead, Vanderbilt's chief information architect, has worked in biomedical informatics at Vanderbilt for well over a decade.

"The simple idea was to assemble information from any source and to use computational algorithms to turn it into something that can be used," says Dr. Stead. "It has no boundaries and it's analogous to what Google has done. Google answers questions by crawling over any number of sources of information -- each of which are used for a single purpose but none having the original purpose of answering your question."

In 2005 the technology behind StarChart and StarPanel was licensed to Informatics Corporation of America (ICA) and rebranded as CareAlign, which has become a leading web-based software as a service. CareAlign provides community information networks -- such as health information exchanges (HIEs)-- and regional health information organizations (RHIOs) -- the integration and "matching" of structured and unstructured data unique to a patient to form a complete, longitudinal and universal health record.

Dr. Stead, also a senior adviser to ICA, urged the National Library of Medicine, the advisory body of the United States Department of Health and Human Services and the agent for managing current Health IT standards efforts, to implement a new definition of interoperability, which the agency did in May of this year.

"The definition we need to think about is 'information that can be assembled and interpreted in the light of current knowledge and reinterpreted as knowledge advances,'" states Dr. Stead. "That definition stands in pretty stark contrast to the more simplistic view that we would get a single correct answer and that the answer would be explicit and understandable by any system that you move it to. Very few of the things we use to describe biological systems fit that degree of explicit single definition."

Dr. Stead uses the example of testing a patient's potassium level. Lab procedures provide signals that are correlated to reach a clinical decision. As the medical profession advances, the way clinical decisions are made is going to change. EMR solutions that function only as transactional and reporting applications are not going to be useful in the future if the data comprising those records are static and cannot be dynamically reinterpreted in the future. This trend in EHR is termed "data liquidity."

"Making the data mean something outside the application is the one thing we can do to ensure that the data is useful over time," says Dr. Stead. "We know that technology is going to change dramatically if we change the care process to achieve the goals of health reform. Any technical approach that locks us in concrete is not going to survive these changes."

On the commercial front, Gary Zegiestowsky had a lot to say about the challenges that ICA and CareAlign have faced while integrating multiple proprietary EMR solutions within community networks. In less than six months CareAlign can be deployed to aggregate data from any source and in any form. As successful as ICA has become in this field, it is still a difficult feat to unite data from several vendors across a complex network of hospitals and clinics. Cooperation is key and proprietary competitiveness must be kept in check for it to work.

For example, MidSouth eHealth Alliance is a network of 16 primary care sites and 14 emergency rooms in 16 hospitals located in Memphis, Tenn. ICA has aggregated more than two years of data, including almost 4 million patient encounters.

The proprietary systems that CareAlign has partnered with in this network include Allscripts, Ulrich and NextGen EMR solutions, Cerner and McKesson hospital information systems, and Meditech's MAGIC laboratory, nursing documentation, pharmacy and radiology information system. Most vendors are doing what they can to accommodate data integration, but competition and proprietary wariness are still an issue.

"Since we are working more community-wide versus enterprise it was an easier bubble to burst, because it wasn't so competitive," says Zegiestowsky. "On the hospital side it's a little more challenging because now you're moving in on vendors' territory, but if you're not providing the client in the hospital or physician practices everything that they need, either you're going to cooperate or you're not going to stay in business. It's the nature of the game at this point. You're not going to see homogeneous, closed systems."

Zegiestowsky says that the first barrier is often a vendors' inability to feed outbound data. ICA works with vendors to free-up their data in whatever form or language it takes. CareAlign also provides interface templates that fill in the blanks when legacy EMR applications don't provide needed functionality. CareAlign can build nursing documentation templates or system-wide messaging and alerts that build intelligence into the workflow.

On a grander scale, ICA's population management tools allow health care networks the ability to set up specific disease dashboards to help communities manage trouble areas, like ventilator-associated pneumonia and diabetes. These dashboards can drive clinical effectiveness and may be under the American Recovery and Reinvestment Act's umbrella of meaningful use, which will be officially defined by the ONC later this year.

Dr. Stead co-chaired a study for the National Academies in January and released a report that signaled a warning: If the medical community stays the current course toward haphazard implementation of health IT without the proper knowledge and design for a united, dynamic system, "not only will it not lead to the fixes in health care that we want, it actually could make it worse," says Stead. "That's because of the focus on transaction process and the competition between suites of applications, each of which are designed to work very well with its own parts but not with anything else. It just won't work."

Zegiestowsky and Dr. Stead assert that interoperability is as critical as the functionality of an EMR, and that broad changes are going to have to be made on several fronts to make sure the universal health record reaches its full potential to contribute to a transformation of American health care.