Health care organizations like other businesses are getting ready for the deluge. Nearly three years ago, technology research firm Gartner listed “dealing with data” as the second biggest IT trend for businesses, pointing out that data would grow 800 percent over five years. What’s more, 80 percent of that data would be unstructured.

What does that mean? The difference between structured and unstructured data is that structured data usually exists in a fixed field or database. In health care, this might include patient details typed into an electronic health record. Unstructured data is, in essence, everything else. For hospitals, it usually refers to text documents or “images,” in this case scanned patient records or the many faxes an institution might receive — for referrals, orders or prescriptions. But it can also include everything from electrocardiograms and digitized pathology slides to imported datasets from CT scanners, MRIs, X-ray and other equipment. “Unstructured data is much more difficult,” says Lydia Washington, senior director of HIM practice excellence with AHIMA.

That’s a problem, because the amount of unstructured data hospitals have to contend with is mind-boggling. Geisinger, an integrated health system in Pennsylvania that runs several hospitals, had more than 2 billion distinct unstructured documents sloshing around its repository as of 2010, according to Joe Stewart, general manager of the health care business unit at OpenText, the company that provides back-end IT infrastructure for Geisinger. That’s the equivalent of hundreds of terabytes of data (one terabyte is 1 trillion bytes, or 1,000 gigabytes). By now, that number’s inching closer to 3 billion, he says.

Enterprising software
How to deal with billions of documents? One answer could be enterprise information management, or EIM, an umbrella term for technologies that help businesses handle both structured and unstructured data, and can help them archive old data, share it with people across departments and eliminate redundancies. So far, medicine is mostly not on board. Eighty-five percent of health care organizations have weak or nonexistent EIM, according to Washington, citing figures from another Gartner study.

“We find health care, like with many information-related things, is a little more behind the curve than some other industries,” says Washington, who gave a talk on EIM at the HIMSS 2013 show in March. “We have not treated information as an asset in the past.”

This can create conflicts and confusion, as the problem with the data glut is not just how to store information or share it, but to ensure it’s accurate and useful. For instance, if patient information gathered in the admissions department doesn’t match up with similar data acquired by the health information management department. “It’s happened for years,” Washington explains. “There’s a realization now, as we move into analytics, that that’s going to be critical for new payment models or care models. You’ve got to have good data, and data has to go across departments and can’t be siloed anymore. EIM is an attempt to bring down those silos, those islands of data.”

For the past six or seven years, some vendors, including big ones like IBM and Microsoft, have started to provide EIM products, and they’re moving into health care. OpenText says one of its customers is EMP, a staffing group that provides emergency room doctors to hospitals on long-term contracts. OpenText helped develop the technology that allows EMP’s physician portal apps to be developed on mobile phones, letting doctors check claims denial rates, satisfaction scores and other things that track their performance.

But for hospitals looking to set up EIM, they should start soon. Washington says it takes at least 12 to 18 months to get one program off the ground. “It’s not really a system, it’s more of a plan or a strategy,” she says. “It doesn’t happen overnight.”