What are the possibilities for RDF (Resource Description Framework) as a Universal Healthcare Exchange Language? It’s an issue to be explored next week at a SemTechBiz workshop in San Francisco.
The healthcare sector is rife with medical vocabularies and localized terminologies. In fact, says David Booth, Senior Software Architect, KnowMED, one of the leaders of the upcoming event, “some people have characterized the problem as not being one of a lack of vocabularies but of too many vocabularies.” To some extent that can’t be helped, because specific languages have grown up with various medical specialties and healthcare subdomains. What can be helped, though, is to create semantic connections among these vocabularies, to avoid the disconnects that can harm patients, researchers, and others.
The issue for a universal healthcare exchange language, he says, has been clearly articulated back in 2010, by the President’s Council of Advisors on Science and Technology (PCAST). “That was identifying a visionary need that has not been met yet,” says Booth.
In the meantime, the price is paid by patients, who can’t get as good medical care because doctors don’t have a full and complete picture of their health histories. Most people use multiple healthcare providers and institutions, and their health information winds up in different forms and vocabularies, so it’s hard to connect up, he says. Researchers have the same problem connecting the data dots – the very thing they need to do to have important new insights, he notes. Booth relates a real-world example: A patient was receiving treatment for a type of cancer and also taking a different medicine for a completely unrelated disease. At some point, doctors linked up the data to discover that the medicine not being used to treat cancer was effectively helping to treat it in the patient.
“This resulted in the first new medication for that type of cancer in 30 years,” Booth says. They didn’t make their data links in that case with the help of RDF, but those kinds of discoveries could be made more efficiently and effectively using a standard language for data interchange. And, he says, “the consumer will benefit by getting more effective and more cost-effective treatment both directly and because of the new research insights that can happen.”
Moving in this direction is well-aligned with the push to electronic health records. “It’s the natural extension of those changes to make those changes more effective, to make electronic health records more useful, and make meaningful use more effective,” says Booth. “One key thing is that this is not about replacing current [providers’ and institutions’] IT infrastructures, and it’s not about changing existing EHRs. It’s about exchange, about how do we exchange healthcare information so that it can be assembled and meaningfully integrated much more effectively.”
Once the mappings from standard vocabularies to RDF have been accomplished, and once vendors of healthcare applications support that in their products, “it’s quite easy to expose data as RDF, very easy, for example, to transform relational data into RDF. So it’s not a burdensome element,” he says.
The healthcare field tends to skew conservative and to go slow on new technologies, he acknowledges, but on the other hand, RDF is mature now and Linked Data principals are being increasingly accepted as important principles for making information more amenable to machine processing. “This technology has been quite well-proven already now in uses both outside healthcare and in some cases inside healthcare, and quite a bit in life sciences,” Booth says. All that may be needed to move the needle, so to speak, in this direction is a demo by important parties in the sector, like the Veterans Administration and Department of Defense, whose healthcare systems still don’t talk to each other effectively, he says, or a mandate for it as part of the government’s meaningful use definitions or other healthcare guidelines.
You can apply to attend the workshop here.