Many radiology administrators remember purchasing their first or second PACS: It was probably no small financial commitment, and justifying the expense might have been a painful process. Just as those older PACS began to collect a little dust—not to mention massive archives—the vendor-neutral archive (VNA) came creeping into the picture. Skeptical equipment purchasers naturally want to know: Why VNAs—and why now?
Jonathan Shoemaker, senior consultant at Ascendian Health Care Consulting, has heard these questions before; their answers stem, in part, from what radiology and cardiology PACS were not designed to do, he says. Software engineers created PACS to run in a structured environment, to address the workflow of radiology and cardiology departments, and to digitize analog processes.
“Outside of the radiology and cardiology ecosystem, it’s the Wild West—because not everyone adheres to DICOM,” Shoemaker says. “As health-care organizations see that imaging is occurring in a lot of other areas within the enterprise, they must address the other departments’ workflow and image capture. If you ask the radiology PACS to ingest these other images, especially if you’re wrapping everything in DICOM, you end up potentially imposing the radiology workflow onto these other departments—and it simply does not fit.”
Data ownership beyond the PACS allows for flexibility, in addition to what Shoemaker describes as building the data flow from the bottom up—essentially, creating a message from the object header, if needed, to support the clinical archiving system’s data flow. He says, “Take dermatology: If the department acquires an image, it may not produce an order in an information system (as radiology does in the RIS for a chest radiograph). Dermatology, however, will charge for the event, on the back end. If the radiology PACS requires an order, you are now forcing dermatology staff to log into a RIS and create an order—just to send the images through an encapsulation device into a PACS.”
ACO–ready Solution
As large integrated delivery networks acquire more hospitals and imaging centers every year, things get even more complicated, and the need for a neutral solution becomes more evident. Shoemaker has seen huge efforts from IT to deploy the hospital information system and RIS across a newly created enterprise.
“You also have the need for cost reduction because reimbursements are going down,” Shoemaker says. “It does not make sense to have all these technical silos with different storage devices. In these situations, IT professionals must manage the data and adhere to the data-protection policy of the organization by using image–life-cycle management (ILM). VNAs bring in single storage solutions for all imaging systems across the enterprise, not just radiology and cardiology.”
Shoemaker has helped organizations implement full VNAs, which address the archive needs of DICOM and non-DICOM data. “The full VNAs can drive the data flow that clinical systems require,” he explains. “You also have the functionality and flexibility for self-managed migrations. If a radiology PACS gets replaced for $800,000, the migration can sometimes cost almost as much as the PACS itself.”
The consolidation of the health-care marketplace will only make these migrations more common, particularly as accountable-care organizations (ACOs) must increasingly share data. “VNAs allow hospitals to become more agile, internally, with clinical-system replacement,” Shoemaker adds, “but VNAs also position hospitals to share their image data more easily through health information exchanges (HIEs).” In a more standardized approach, they can use Cross-enterprise Document Sharing (XDS) and XDS for Imaging, or XDS-I, IHE implementations.
Hospitals aren’t the only ones consolidating, as evidenced by Lexmark International’s recent acquisition of Acuo Technologies—a trend that Shoemaker also views as a natural progression. “I expect to see a few more content-management systems acquire VNAs,” he says, “especially as VNAs begin to move out of the radiology/cardiology space. There is a large need for interoperability between enterprise content-management systems and clinical systems to provide a complete record of structured and unstructured content to the EMR systems.”
To achieve these efficiencies—and, ultimately, to be a full VNA—a system must have an interoperable platform for clinical-systems archiving, and that archive must accommodate both DICOM and non-DICOM data. Some of the full VNAs also come with their own vendor-provided universal viewers.
All VNAs use an interface for third-party universal-viewer integration. “There are also hybrid VNAs that are (kind of) crossing into being full VNAs, but their roots are strong within the DICOM world,” Shoemaker says. “How they manage the non-DICOM support is really vendor specific.”
The Right Reasons
A VNA is not a small investment, but Shoemaker is convinced that in the right situation, it can prove to be the perfect solution. In his work as a consultant, Shoemaker addresses the specific needs of each business to determine whether a VNA makes sense.
He makes sure that all clients know that, by definition, VNAs are hardware agnostic—which allows IT personnel to purchase storage in bulk. Clinical systems can use a smaller online cache of whatever supported storage solution they need. “You also have the ability to communicate with HL7, DICOM, and Web services, which are coming into health care in force,” he says.
Some administrators are surprised to learn that they can use existing assets with a VNA. “Let’s say you have a storage device that has been retired from the strategic storage plan, but it hasn’t reached the end of its life,” Shoemaker explains. “You could repurpose that as a third or fourth tier for compressed images, or images beyond the typical life cycle for retention. A true VNA will bring in both the clinical rules and the technical rules for enhanced ILM capabilities. Storage devices can only determine when the object was created, updated, and accessed: They do not know what the object (such as a chest CT exam) is.”
In contrast, he says, “A VNA that is receiving order information (or admission, discharge, and transfer information), will know the object is a mammogram and that it has a different retention policy than a chest radiograph has—or that the patient is under 18 years old. Depending on the clinical information, a VNA can determine where to move those data, or whether to compress them or delete them.”
For those merely looking for an additional DICOM archiving system, Shoemaker concedes that a VNA might not be worth the money. A PACS only provides access to the data managed through it, and if it is a radiology PACS, the electronic medical record (EMR) is only seeing the radiology images. “The VNA provides access from all departments that are integrated with it,” Shoemaker says. “Therefore, a VNA integration, if it is fully implemented, will provide physicians a more complete read-only imaging record than a PACS could.”
The key is choosing the right vendor for the right reasons—a task that can be daunting, in and of itself. Not all VNAs are the same, and contacting a consultant is one way to cut through the confusion.
“Because a VNA touches the entire enterprise, not specifically radiology, the client needs to understand its needs for today and the next five years,” Shoemaker says. “The wrong vendor choice will have an impact on your entire organization. If you buy the wrong VNA, the paradigm of never having to migrate your data again goes out the window—because when you replace that VNA, you are going to have to do the migration again.” Ultimately, Shoemaker sees a certain inevitability to the progression toward VNAs, reasoning that there will be a continuing need for health-care systems to bring in external data and map them to their internal clinical systems to create a complete medical-imaging record for the EMR.
“They will need to ingest images into their clinical systems so physicians can view them side by side, but yet not archive them,” he says. “PACS are typically too structured for that type of data flow. Full VNAs, on the other hand, can provide accurate interoperability with different imaging systems, leveraging IHE profiles, HL7, DICOM, and Web services to deliver or ingest those data into clinical systems efficiently.”
Greg Thompson is a contributing writer for Radinformatics.com.