Every advance in imaging technology—the proliferation of slices per CT scan, the mounting use of MRI and CT to diagnose heart disease, and the increasing prevalence of digital mammography—also advances one IT challenge in particular: storing and protecting image information in the event of a disaster. “Imaging is growing exponentially,” Corey Zeigler notes. Zeigler is health IT program manager for
Fort Drum Regional Health Planning Organization (FDRHPO), Watertown, New York, a payor-funded regional health information organization (RHIO).
He adds, “As you bring in these new technologies—3D cardiac imaging and digital mammograms, which you want to keep forever—you’re seeing this curve. It’s like a half pipe for skateboarders. How do we manage all that?”
Zeigler oversees the effort to link five regional hospitals and 40 freestanding physician practices, including an imaging center, via their respective electronic medical records. “We’re a federated model. There are a lot of islands of information, and we set up pointers,” he explains. “We do cache some information, but we’re not responsible for storing it. The entity that’s required to store it is still the individual or hospital.”
The RHIO’s larger hospitals have their own redundant data centers, protecting their information in the event of a disaster or system downtime, but the physician practices and smaller hospitals still store information on local servers, Zeigler reports. “We’re figuring out how the folks who are running the data centers can share their space. We’re trying to figure out how to move these images so people can access them when they need to,” he says.
Familiar ChallengesThe six entities within the RHIO that perform imaging (the five hospitals and the imaging center) each have their own PACS—which, Zeigler says, presents an immediate problem in terms of storage management. “You want to find the balance between what you have to have right now versus what images you can have in near-term storage, and the management is proprietary to the PACS software, so you’re bound to whatever the software will support. Fortunately, all six are on the same PACS, but its global viewer doesn’t scale well; when you start to get big, it slows down,” he says.
Two of FDRHPO’s hospitals are now performing digital mammography, according to Zeigler, and in response to the explosion of image information that needs to be stored, one hospital has adopted a tiered storage architecture. Studies less than 180 days old are available on spinning disk before being archived on content-addressable storage. “You really just need a dictionary to tell you where the data are,” Zeigler notes. “Your shortcut resides on your fast drive, so you can quickly find out where the data are.”
This method of storage and archiving, however, creates fresh challenges when it comes to disaster recovery and business continuity: repopulating one data center from its mirror and restoring images to their appropriate short- or long-term archives. “Retrieval takes longer,” Zeigler says.
To address this issue, FDRHPO just finished a fiber project that put 811 miles of 100MB fiber in a loop throughout northern New York. “It spans a long way, and it gives us a closed, secure network that’s health-care specific,” Zeigler says. “We have a couple of big players that have data centers that are redundant that are hooked into that fiber, and we’re hoping the smaller hospitals can maybe buy some space on it. We’re trying to pool our resources here, sharing existing storage space over this closed network.”
Sharing ResourcesPooling resources between RHIO members is also not without its hurdles, Zeigler notes. “The trouble is hospitals aren’t in the IT business,” he says. “You have to be very careful and extra vigilant when it comes to service-level agreements. The divorce paperwork—how to separate the data, if you need to—is challenging. Most hospitals have never written a contract like this, and that’s really what it takes to be successful.”
The economies of scale resulting from sharing these resources, however, might be necessary if smaller hospitals are to adopt the kind of storage-management and disaster recovery strategies pioneered by their larger peers. “It’s cheaper to be on the dedicated network,” Zeigler says, noting that the network’s bandwidth can be bumped to a gigabyte as necessary. He adds, “If the hospital can choose between a $100,000 SAN and a $150,000 SAN that has three times the capacity, it just makes more sense for everyone to share.”
The larger hospitals will only charge their smaller counterparts what is necessary to break even, Zeigler says. “Technology should not be a differentiator between facilities,” he observes. “We’re trying to give them a common footprint to compete in so they can focus on their core business, which is providing great health care and customer service. When you ask people what they’re looking for in a hospital, they don’t say anything about a tier 1 data center.”
Preventing any downtime in data availability, stands to benefit small and large hospitals alike—both financially and in terms of providing the best possible care, Zeigler notes. “An eight- to 24-hour downtime is an exponential cost. As we move to the electronic record and there is no paper system, the light bulb is going on in people’s minds. They’re realizing we can’t go down,” he says.
Innovative solutions for data protection will be increasingly critical, Zeigler concludes, as medical information becomes entirely digital and as facilities face increasing pressure to make the most of narrow profit margins. “Every facility wants to take great care of its patients, but they also want to survive in a tough market,” he says. “We’re trying to support the little hospitals as they try to get into enterprise reliability and business continuity—without the capital actually to do it for themselves.”
Cat Vasko is editor of ImagingBiz.com and associate editor of Radiology Business Journal.