By David Yeager
Vol. 13 No. 3 P. 6
The explosion of digital medical images during the past decade and images’ important role in the medical record have elevated the archiving from a departmental concern to an enterprisewide issue. While some practices and institutions have chosen to maintain their own archive on site, an abundance of off-site options are also available, some using proprietary methods and, more recently, some that bill themselves as vendor neutral. With so many options, deciding which solution is the best fit for a given facility can be confusing.
Part of the problem, says R. L. “Skip” Kennedy, MSc, CIIP, technical director of imaging informatics for Kaiser Permanente medical centers in northern California, is that the need to build these archives has outpaced the industry’s archive strategy. That need for petabyte-scale DICOM storage drove the building of enterprise archives before their role was fully defined. Clearly defining the archive’s purpose goes a long way toward ensuring that the chosen solution meets an organization’s needs.
“One of the things we jumped into when we first started deploying vendor-neutral archives and enterprise archives was that we did so without clearly getting at what their functional requirements—really their mission statement—are,” says Kennedy, who will be cochairing a learning track at the Society for Imaging Informatics in Medicine’s annual meeting that will examine next-generation archival models. “If it really is to be the clinical repository for all imaging, whether it be cardiology, radiology, dermatology, the digital operating room, etc, that’s a different role than having a series of best-of-breed, front-end image management systems that will then rely on an image archive for deep storage [rather than] near-line clinical storage.”
Because clinical storage and long-term storage play fundamentally different roles, Kennedy says expecting an enterprise archive to serve as a clinical repository is not realistic. Advances in the performance of clinical systems have raised expectations for archives’ performance beyond what most solutions can deliver. Clinicians expect to see images “at YouTube speeds,” not in 30 seconds or more. Most archives can’t achieve those speeds without relying on proprietary protocols rather than DICOM protocols.
Rather than linking an EMR to it, which many healthcare facilities do, Kennedy suggests linking the EMR to the radiologists’ workstations. DICOM allows clinicians to share contacts and presentation states, and the workstations are much faster than most archives. This would remove the clinical burden and allow the archive to serve its primary function, which is storing vast amounts of medical data for a very long time.
How long data will be retained is a significant factor in determining a healthcare organization’s archiving strategy. Kennedy says most enterprise archives have never implemented data life cycle rules. Although storage is cheaper and more available than it’s ever been, it’s still not free, and the rate at which it’s being used is increasing exponentially. Considering this trend toward storing more rather than less, storage has the potential to take a big bite out of an IT budget.
“It’s the box that nobody wants to open because nobody wants to answer it, but forever is a really long time,” says Kennedy. “If we start stashing petabytes worth of data away, [it’s going to be astronomically expensive].”
Taking into account statutes that govern data retention requirements, healthcare organizations need to carefully consider storage costs. Some states require more record retention than others, but the general trend is toward more. As an example, Kennedy points to California law, which requires all pediatric files to be saved for 30 years past maturity. It’s possible, or even likely, that somebody is storing a neonatal film of a 19-year-old, 270-lb football player, even though that film most likely has little, if anything, to do with the patient’s current state of health.
A significant stumbling block is that legal requirements for record retention are built around paper documents, not digital ones. Currently, there isn’t even a legal concept of a lossy compressed version of a document. Kennedy believes that a professional society, such as the ACR, will have to publish detailed retention recommendations and allow them to be tested in court to achieve some clarity in this area.
Because of the uncertainty of record retention requirements and the growing reliance on storage, owning and maintaining an archive can be an expensive proposition. This has led many healthcare organizations to explore off-site storage options. While these cloud-based services offer some advantages, such as copious storage space and redundant, multisite archiving to minimize risks to the data, it’s not a simple calculation.
To begin with, many archiving companies are relatively new, while many hospitals have been around for more than 200 years. A healthcare organization needs to have a contingency plan in place in the event that its storage vendor goes out of business. This is extremely important for two reasons. One is that the hospital is legally responsible for maintaining those records, regardless of its archive partner’s status. Two is that migration costs can be daunting, and migrating an entire archive is difficult under the best circumstances. Kennedy says the industry is still grappling with how to mitigate risks related to vendors’ durability.
“And I don’t mean to say that in a way that makes it sound like I’m not in favor of cloud storage,” says Kennedy. “I actually think we’ll move to something like that, in slightly different models than we’ve seen so far.”
Kennedy believes some of the larger, older IT companies will begin acquiring some of the archive companies, which would provide more of a comfort zone to chief information officers looking at a sizable, long-term financial commitment. He also thinks imaging vendors will offer vendor-neutral services, either on their own, through partnerships, or through acquisitions. While he’s not sure what form those services may take, he believes a more diversified approach to archiving will become increasingly desirable.
“There is a strong need for long-term, enabled DICOM storage that may or may not be part of your PACS system,” says Kennedy. “There is also a place, I think, for a single enterprise repository for different departmental content. Rather than having to fight the battle of ‘I’m going to have to migrate the cardiology system, I’m going to have to migrate the dental system, I’m going to have to migrate the ophthalmology system,’ [if I have] a single place where I can keep all of that data in DICOM, I may move it, but I won’t be moving the full package.”
— David Yeager is a freelance writer and editor based in Royersford, Pennsylvania, and a frequent contributor to Radiology Today.