Recently, I interviewed the CIO of a large Midwestern healthcare system about its plans to become an accountable care organization. The healthcare system was installing a well-known electronic health record that will allow its providers to access patient data across inpatient, outpatient and post-acute care settings. The biggest obstacle the CIO saw to health information exchange was the lack of national standards that would enable the system's EHR to communicate with the EHRs of private practices and other providers outside of the enterprise.

The Obama Administration is trying to address this problem. In the near term, the government is testing its Direct Project, an open-source exchange standard that can help providers "push" care summaries and other data to one another. At the same time, the state health information exchanges receiving federal funds are required to figure out how to get providers to exchange Continuity of Care Documents (CCDs), which contain the key information on diagnoses, medications, and allergies that providers need to treat a patient.

But in the long run, that won't be enough. To create richer data interchanges, a Presidential commission recently recommended the development of a "universal exchange language" for healthcare data and a standard infrastructure for locating patient records while protecting privacy and security. Extrapolating from the experience of other industries, the President's Council of Advisors on Science and Technology (PCAST) also recommended that all healthcare information be broken down into "meta-tagged data elements" that can be easily transported.

It sounds good, except that it's not so simple. To start with, HIMSS pointed out in a comment on the PCAST proposal, inaccurate patient identification could undercut the goals cited by the Presidential advisors. The algorithms used to identify patients uniquely are not always accurate, HIMSS noted, and should be improved. Of course, unique patient identifiers could solve that problem; but PCAST ruled those out, probably because they're politically controversial.

Meta-tagged data elements could probably overcome the problem of incompatible databases. But then there's the problem of semantic interoperability. Myriad terms and codes exist for many medical terms, and there are upward of 70 different medical terminologies in use. Transforming each of these terms into a unique data element would be a tall order, although there are companies dedicated to mapping terms to a common vocabulary.

In short, a standardized method of exchanging all healthcare information remains a Holy Grail. But even if that doesn't emerge in the near future, there still are ways of exchanging less robust data sets that will allow providers to communicate and coordinate care better than they do today. - Ken