Controlling Their Own Destiny

The volume of medical imaging data isn’t just growing at a steady rate – it’s exploding. In just a short period of time, advances in modalities such as CT, MR, and full-field digital mammography are creating significantly larger data sets than previously encountered, and there’s no sign of slowing, either. With an aging baby boomer population, radiology faces a harsh demand for diagnostic imaging exams. How can a facility prepare for this wave of data that will come crashing down on them in the coming years?

For some institutions, such as The Children’s Hospital of Philadelphia (CHOP), it’s never too early to be thinking about the future. Chris Tomlinson, MBA, CHOP’s administrative director of radiology, is taking a proactive approach by implementing a concept that completely diverges from the traditional model of data management – an independent enterprise imaging archive.

Common Obstacles

CHOP’s decision to stray from the traditional method of PACS and enterprise storage was driven not only from a stark awareness of the impending flood of information, but from the frustrations the institution faced trying to manage the different departments across the enterprise. Having to deal with images from the other “-ologies” such as pathology and cardiology, not just radiology, often turns into a compatibility and access nightmare – not to mention a significant ongoing expense maintaining each department’s disparate PACS viewer, archive, and storage solution.

Once it was time to replace the PACS, all of the data would need to be migrated to the new system, taking a tremendous amount of time and money. With image sets growing exponentially, Tomlinson recognized the urgent need to find a better way to store images and move them around. He also made it his mission to break away from the traditional PACS model, which didn’t allow CHOP to control any of its data because of the proprietary data formats created by most PACS.

PACS has been traditionally viewed as a unified solution. But Tomlinson questions why it can’t be taken apart and viewed as separate components. “We have these old architectures and infrastructures for PACS that don’t really make sense and were built for film-based radiology departments,” he says. “If you could blow it up and do it again, how would you do it?”

Tomlinson attributes his way of thinking to not having been brought up in a film-based department. His background is in IT and consulting. “I didn’t come up through film to witness the old ways MRIs or CTs were done, or even the old reading rooms. I don’t have all that baggage, so it helps me to see it a bit clearer,” he says.

Once he identified the areas that needed to change, Tomlinson’s next step was to find out how he could save CHOP from expensive migrations and unify storage across the hospital’s departments. Part of “getting out of the proprietary game,” as Tomlinson puts it, involved the implementation of a vendor-neutral archive (VNA), an intelligent IT infrastructure that would put control of data into the institution’s hands by separating out the storage and proprietary archiving from the clinical application.

Under CHOP’s new medical image management solution, St. Paul, Minn.-based Acuo Technologies LLC is supplying the vendor-neutral archive, which sits on top of a grid medical archive solution (GMAS), supplied by Armonk, N.Y.-based IBM Corp.

How It Works

A complete vendor-neutral archive solution involves several layers. At the top is the clinical application layer, which involves clinical and enterprise viewing and contains the PACS viewer, the 3-D viewer, etc. Underneath that is the interoperability layer, or archive layer, and thirdly comes the storage management layer. The VNA takes the data obtained from the modalities and “virtualizes” it down to one of the layers in the storage device and distributes it to the viewers that require the images. This approach takes the viewer out of the “ingesting” or initial storage step. It allows the PACS to be a viewer and not an archive.

Tomlinson describes it as a traffic cop, directing where the studies go. When someone in the institution needs to call the study to a viewing application, the VNA knows where the information is stored within the storage application and brings it back to the user. The intelligent storage device makes and stores two copies of each study – one housed onsite at CHOP and the other 23 miles away at a back-up data center in Norristown, Pa.

“So what we look at as the vendor-neutral archive is the idea of taking all of the archiving/data management functions that reside in a PACS, removing them from the PACS, letting the PACS continue to acquire images and move them around in the department, but anything beyond a very short-term cache should immediately be put into a centralized archive at the enterprise level,” says John Koller, founder and president of KAI Consulting, in Larkspur, Colo.

And this can be done for any department that acquires images, not just radiology. Tomlinson says the other “-ologies” have the same problems but that they’re in varying degrees of awareness.

“Enterprise imaging isn’t just radiology. If people are worried about who owns PACS – IT or radiology – they’re missing the boat. It’s an enterprise solution. It’s for any area in a hospital that acquires images, stores images, and needs to distribute images to the enterprise,” Tomlinson says.

And many of these clinical departments are producing fixed content objects that aren’t DICOM. Endoscopy, for example, produces digital images. These, too, will need to be stored electronically, so a section of the vendor-neutral archive will have the ability to store and recognize non-DICOM objects, as well. Pathology is another area where the associated files are very large, adding to the tsunami of data. Koller, who is on the DICOM working group for digital pathology, sees this area as experiencing volumes of data that will dwarf cardiology and radiology combined.

Dealing with Data

According to David Stalder, executive vice president and co-founder of Abrio Healthcare Solutions Inc., in Brentwood, Tenn., there are two key drivers that could spur a hospital to move to an enterprise imaging archive model. The first is electronic medical record (EMR) adoption. To implement EMR effectively, there needs to be an enterprise imaging archive that can consolidate all the different departmental systems.

For example, radiology, pathology, ophthalmology, pediatric cardiology, and adult cardiology are all departments that may have their own independent systems and may also use systems from different vendors, with their data locked up in silos.

“The challenge is bringing all imaging data into an EMR where it can be managed as a clinical information data warehouse,” Stalder says. The enterprise clinical image repository can be used by the EMR that physicians can access, providing them with a more comprehensive view of a patient’s studies.

“The vendor-neutral concept is a starting point in many large-scale institutions that are in the race to provide patients and their healthcare providers with one holistic patient record, whether they are in-hospital or remote,” says Paul Markham, vice president of global strategy and marketing for Milwaukee-based TeraMedica Inc., a company that has been working with large-scale institutions such as the Mayo Clinic, Johns Hopkins, and M. D. Anderson to provide them with patient-centric solutions to move to EMR.

Another major motivator stems from the painful and expensive process of data migration. Many of Abrio’s clients are on their second or third generation of PACS. “They’re sick and tired of basically having to start over every three to five years with a new PACS vendor or even the same PACS vendor who comes out with a new platform, forcing a migration of all their imaging data,” he says. “And they don’t want to have to do that anymore.”

Koller agrees, saying, “With every change from one PACS vendor to another, we keep moving larger and larger volumes of data, which increases the time, cost, and challenge of moving all that data with each subsequent PACS migration.”

At CHOP, Tomlinson’s business case states that by taking ownership of the data, hardware, and storage, along with never having to migrate again, his hospital has cost-avoided $3 million over five years versus a traditional PACS model. The payback increases even further when you factor in the cost avoidance of having to migrate PACS vendors every three to five years. With the explosion of image data that’s expected to occur, moving the data now as opposed to waiting for later easily pays for it now, according to Tom Rose, IBM’s national healthcare solutions manager for the U.S. “It is such an easy financial argument,” he says.

Compounding the issue of migrating data is the amount of data that PACS not only generates, but keeps around. Many hospitals are not sure what data to delete, so they store it all. However, storage media changes periodically, requiring the hospital to physically move their data from one media to another. As studies become larger, there is more to move.

Children’s hospitals such as CHOP are required to keep studies for a longer period of time, and the same goes for mammography and oncology studies. But for most normal findings, the business value of an imaging study might be near-zero after the interpretation has been posted, but the clinical value of the data is significant for years, and the legal value of the data requires that it be preserved for each state’s legally mandated retention period (at least five to seven years). Koller says CIOs and department heads need to start thinking about data retention/deletion strategies from a business perspective.

If the business value of a normal chest X-ray, for example, goes to near-zero after it’s read, does it make sense to continue paying for its migrating and storing on high-cost media, which involves power, technology, human interaction, and maintenance? However, most hospitals are keeping everything of low and high value on the same high-performance cache for longer than is cost-effective. In cases like this, Koller says, the infrastructure needs to be intelligent enough to move those studies with little business value to the lowest cost storage system.

This is where an information lifecycle management (ILM) policy will help a facility cost-effectively manage the flood of data. “You have it coming at you so fast and so much, it would be nice if you could get out in front of it and design in where it’s going to be kept so you can handle it with better cost performance,” Rose says. The ILM intelligently and automatically purges data a facility is no longer required by law or necessity to keep. This frees up storage space, allowing more head room for the critical studies and saving institutions from purchasing more storage or paying for maintenance costs after the technology goes out of warranty.

“It’s really important to understand that we do have to take advantage of a purge strategy in order to get the volume of data to a manageable size that still complies with legal regulations,” says Michael Gray, owner of Novato, Calif.-based Gray Consulting.

Gray, who states “PACS-neutral archive” is a more accurate term, encourages facilities that have already budgeted for more storage to use those dollars more intelligently and begin migrating data to a standalone neutral archive. “You will make that first baby step toward building out what will eventually become a full-fledged enterprise-neutral archive,” he says.

Moving Forward

It’s not surprising that Tomlinson, with a strong background in IT, saw VNA as an obvious solution. From an IT standpoint, it makes sense to have a centralized enterprise-wide VNA instead of multiple systems having independent storage. Part of an IT department’s responsibilities is to manage data. But when radiology started going digital, many clinicians handled storage and IT issues, even though it was out of their area of expertise.

Koller thinks too many doctors are getting involved with their day-to-day IT systems. “If you look at it, managing data through its lifecycle: protecting it from a security, integrity, and availability perspective – these are things that IT does in their world,” he says. “This is their sweet spot.”

Gray notes that the vast majority of clients in the past couple of years have been IT departments that are responsible for the infrastructure issues of a PACS and think of it more technically than clinically. “They’re the ones who get it right away,” he says. “They’re responsible for managing data for all the different departments, and they want to bring common sense into it by consolidating the storage.” CHOP’s solution was built with a high degree of collaboration and design input from information services.

Stalder says he runs into some Abrio clients who are implementing VNAs to focus their systems management and support to a centralized IT department perspective, while others have a hybrid approach in which the clinical application is managed by the department and the enterprise archive is supported by IT. “As our clients are moving toward an IT-centric support model,” he says, “it is clear that IT is getting more involved in the decision-making and support of any of the departmental PACS strategies.”

The next step in implementing a VNA is moving data from the existing proprietary archive into the VNA. For CHOP, this process has taken about nine months, but once the transition is completed, they will never have to move their data again. “Being able to control our own destiny is huge,” Tomlinson says.

Any future PACS purchases, for example, will not be as complex or expensive. All of the departments can choose their own user interface or viewer and it will be connected to the VNA. “The beauty is doctors don’t care how images are stored. All they care about is that you get them the images quickly in the viewer of their choice,” Tomlinson says. In the vendor-neutral archive set-up, the PACS or viewer becomes the tool for specialized viewing and workflow applications and does not control how the data is archived and managed. Koller says, “The PACS can be plugged and unplugged almost at will and with almost no trauma attached to it as far as moving the data.”

This is good news for clinical departments who have pre-existing preferences for PACS systems. Stalder says, “Rarely do we see a consensus across departments in terms of which application vendor they want to use. It’s very hard for IT departments anywhere to mandate decisions to any given department by saying, ‘We’ll make that clinical tool choice for you.’ “

The Forecast

Now that CHOP and other early adopters are in various stages of implementing this new model, it will only be a matter of time before the rest of radiology and the other “-ologies” catch on. “People have done components and different variations,” Tomlinson says. “We’re trying to put the whole thing together.”

How will the vendor-neutral archive change the PACS game? The experts have varying opinions, but the general consensus is that many integrated delivery networks will turn to this solution. Koller says, “Any hospital small or large that’s not looking at a VNA approach is short-sighted and leaving a lot of money on the table. And in today’s times, we cannot afford to not take
advantage of these long-term cost-saving methods.”

Rose believes hospitals will readily embrace the VNA. “They want to carve out their patient data and own it and not have it tied to another company,” he says. Stalder agrees that the “unbundling” of the archive, storage hardware, and other components that were historically sold as a turn-key solution from the PACS vendor generates significant savings and strategic flexibility, but cautions all of his clients to base their decisions on their unique vendor-sets, internal support structures, and stated organizational goals. “We have seen some very ugly situations in the marketplace when the process is not coordinated effectively,” he says.

Gray asserts that we can easily predict the way the trend will go. “Those of us who are close to this emerging market – PACS-neutral archive – recognize that its adoption will follow almost exactly the curve defined for radiology PACS. There are the early adopters, and there’s this chasm separating hundreds of users on the other side that are waiting to see if this actually works,” he says. “And they’re going to need ways to get started, primarily because it’s 2009 and it’s the financial crisis. Hospitals are having difficulty finding money to borrow to do anything.”

Koller acknowledges the shift is inevitable, but doubts we will see this happen quickly based on healthcare’s historical pattern of slow adoption. “It could be years before we get a significant enough adoption from all these hospitals to actually see a real industry-wide impact,” Koller says.

— Jane Kollmer is editor of rt image. Direct questions and comments to jkollmer@rt-image.com.