Building a Better PACS: Part 5 -- Vendor-neutral archives

2009 01 30 12 15 27 600 Better Pacs 70x70

There has been a lot of discussion about archive schema, and especially vendor-neutral archives (VNA), in the past few years. Most of this has centered on the role of a VNA in radiology PACS, yet establishing a VNA for PACS is merely the tip of the iceberg for most information technology departments.

2009 12 03 13 22 40 438 Cannavo Head Shot
PACS consultant Michael J. Cannavo.

The bigger fish in most hospitals is instead establishing a facility-wide central data repository (CDR) that allows multiple clinical systems to utilize a central archive that provides data storage in a vendor-neutral format using a single industry-standard database. This can be done onsite or offsite, with offsite archives used by those who want to avoid creating and supporting a central archive as well as not having to implement a disaster recovery plan. A disaster recovery plan is crucial to any successful archive implementation.

The jury is still out on what does and does not constitute a VNA. None of the standards committees such as DICOM, HL7, or IHE (Integrating the Healthcare Enterprise) offer a single, universally accepted definition, yet most definitions include several areas that overlap. IHE also offers several profiles that are beneficial to a VNA as well.

As is typical in the PACS arena, confusion abounds as well as to differences between VNA, DICOM archives, and enterprise archives. Almost every vendor claims they have a DICOM archive and for the most part they do. But DICOM archives and VNAs are as different as night and day. Without going into a lot of detail, the basic difference between a DICOM archive and a VNA is context management and the ability to process the data differently from the way it was stored.

This is comparable to only being able to look at a photograph, compared to being able to process the photograph differently on one's computer. Storing data in a DICOM Part X format is really all that's required to be considered a DICOM archive, and most vendors do that consistently. Unfortunately, each vendor supports (or does not support) all the various data elements available as defined within DICOM.

This "pick and choose" approach creates problems when data are migrated from one system to another and is what makes these archives more proprietary than open. This also requires the use of a third-party data migration service to help bring the data into a VNA in a format that can preserve the majority, if not all, of the image and database information.

Unfortunately, much of the data stored in a vendor's DICOM archive are also still stored in a proprietary manner. The data often do not transfer across doing a simple DICOM query/retrieve function (GSPS -- grayscale presentation state -- is one area that is often "lost"), which can be and often is an issue if not properly addressed. A VNA eliminates the need for concerns about anything of a proprietary nature.

Penny-wise, pound-foolish?

While VNAs have been getting a lot of attention in the PACS arena, buying a VNA that only benefits PACS is penny-wise and pound-foolish. Now don't get me wrong: VNAs for PACS definitely have their place, but why would you spend the money for a dedicated PACS VNA when a central archive, central repository, enterprise archive, or whatever you care to call it can bring a facility that much closer to the golden ring called ARRA (American Recovery and Reinvestment Act) dollars. Those dollars are out there for the taking -- why not take them?

The healthcare IT provisions of ARRA include at least $35 billion to "incentivize" full electronic health record (EHR) deployment at hospitals nationwide, allowing five to seven years to reach that goal. This may sound like a short time frame, but most hospitals larger than 300 beds already have several digital systems in place, PACS among them, so it is a question of tying them all together to meet the ARRA's definition of "meaningful use" as a justification for funding.

"Meaningful use" has three stages and is overseen by the U.S. Centers for Medicare and Medicaid Services (CMS). Stage 1, which begins in 2011, proposes 23 objectives/measures for eligible hospitals that must be met to be deemed a meaningful EHR user. These focus on electronically capturing health information in a coded format, using that information to track key clinical conditions, communicating that information for care coordination purposes, and initiating the reporting of clinical quality measures and public health information.

Stage 2 would expand upon the stage 1 criteria in the areas of disease management, clinical decision support, medication management, support for patient access to their own health information, transitions in care, quality measurement and research, and bidirectional communication with public health agencies. The people who oversee all this may consider applying the criteria more broadly to both the inpatient and outpatient hospital settings.

Stage 3 would focus on achieving improvements in quality, safety, and efficiency, focusing on decision support for national high-priority conditions, patient access to self-management tools, access to comprehensive patient data, and improving population health outcomes.

Hospitals that meet the meaningful use requirements with the establishment of an EHR by 2014 will benefit from three reimbursement years only, with a transition factor of 75%. Those who do not achieve meaningful use by 2015 will benefit from only two reimbursement years beginning with a transition factor of 50%, while those who fail to adopt EHR by 2015 get no incentive money. Confused yet? So is everyone else, so you are in good company.

The bottom line is that those wishing to achieve maximum payout need to fully adopt an EHR before 2013, and that includes PACS and nominal uses of VNAs as well.

A centralized approach

I've read more articles than I care to admit about the explosive growth of data volumes in healthcare, and everyone has their own approach to managing this. One thing is clear -- the individualized archive silos that have been part and parcel of nearly every clinical system in place today need to go by the wayside in favor of a centralized archive solution. This can be either one created in-house or external to the facility.

There are several reasons for this change. The first is support, both internal and external. Each individualized archive no doubt has its own specific database manager as well as associated hardware. It probably also has multiple databases to support, and multiple image file formats, all residing on hardware from multiple vendors. Unfortunately, this nightmare defines how most hospitals operate today.

Almost every clinical information system has some form of archive attached to it, be it a storage area network (SAN), network-attached storage (NAS), direct-attached storage (DAS), or other long-term storage silo. A midsized hospital can easily have five or six of these or more, one each for every department in the hospital.

By consolidating each of these individual silos into a single centralized archive, you eliminate the annual service cost of multiple archives (which can be significant), as well as internal support costs and related costs (significant as well). Most importantly, the end user now only has to query a single database to obtain the data, significantly speeding up information retrieval time.

Performing this migration to a single archive is not without cost, however. Because most of the image data in radiology, cardiology, and possibly even lab and pathology systems are stored in proprietary file formats, they need to be converted over to a standardized form that can be understood by all vendors.

In the case of PACS, the data should ideally be reconciled against the RIS to ensure that the studies in the RIS match those on the PACS, at least from a report status. The same can be said for cardiology as well. The cost of this varies based on volume but can be as little as 15¢ per study for a straight data migration to as much as 50¢ per study for lower volume (< 50,000) fully reconciled migrations. Once the data are converted into a standardized vendor-neutral format, all future data coming into the system will be stored in that same format with no additional costs incurred.

The one thing that isn't often talked about in archives is disaster recovery. Very few of the smaller IT solutions have disaster recovery plans in place. The best disaster recovery solutions offer offsite storage and retrieval with fairly rapid recovery plans (> 2 TB/day, starting with the most current studies and moving backward).

Recovery is the item almost never considered by most IT departments. Having the data on tape or disk and backed up is one thing; having either the internal or external resources postdisaster to reload and, if required, to validate the data is entirely another. That is one of the advantages of an offsite archive solution -- not only is everything typically redundantly stored, but the disaster recovery option can be put in place almost transparently by the offsite provider to the organization.

The cloud

There was a lot of discussion about cloud computing at the 2009 RSNA meeting. Clouds have been available in general computing for several years now but have only recently been embraced by healthcare.

In its simplest form, a cloud-based system allows multiple disparate systems to upload images and reports to a central data repository, or "cloud," and then resend them out to various clients from a single site. The cloud is responsible for consistently formatting everything in a singular fashion, so that it can be read by the Web-based viewing software on the receiver's PC. This is typically aimed at the primary care physicians as opposed to radiologists, but can be used by both.

The upside of the cloud is that a primary care physician only has to get information from a single source instead of querying multiple disparate systems. The downside is that the hospitals that own those multiple disparate PACS and RIS networks have to pay what amounts to a fairly significant sum for cloud usage.

How much? Prices are all over the board, but they range from less than $1.00 per study to well over $12 per study depending on size, volume, and so on. I'm sure you'll see these prices stabilize at around $1.50 to $2.00 a study and possibly less, especially with volumes coming down. This will also lead to the development of the iPHR (Internet-based personal health records) market being actively pursued by several large computer entities.

The development of clouds in healthcare is being done primarily to replace Web-based image distribution systems found in virtually every PACS. Prior to 2000, most clinical systems were largely isolated. They became more and more "online" from 2000 forward and will stay that way for at least the next three to five years, phasing out when most facilities create an EHR that requires cloud usage.

Cloud computing in healthcare started evolving in earnest last year and is expected to reach its peak around 2020. Many challenges still abound, though. Standards adoption and utilization, which is crucial in VNAs as well, remains the number one barrier.

For reasons that seem to evade logic, virtually every PACS vendor maintains certain aspects of their archives as proprietary. What's even more fascinating is these very same vendors have recently begun adding middleware from third-party providers to make their own archives more "vendor-neutral."

Faced with competition from third parties, they finally recognized that adopting standards-based archiving is basically a "do or die" scenario for them, even though many still make arguments for holding onto their proprietary schema. Until the industry becomes largely standardized and the use of anything proprietary eliminated, the expansion of cloud-based networks, the development of RHIOs (regional health information organizations), and the expansion of PHRs (personal health records) and other standards-based information sharing will be stymied.

Ways of insuring compliance with HIPAA is also crucial to the growth of this industry, as clouds can be a hackers' paradise.

Lastly, once healthcare facilities become more fully electronic and embraced by end users, these very same consumers will drive the adoption of standardized records and, with them, VNAs and clouds.

The future of healthcare is reliant on information sharing. That said, the only way that information can be shared is if it is stored, disseminated, and reviewed using industry standards such as those found in VNAs.

By Michael J. Cannavo contributing writer
February 8, 2010

Michael J. Cannavo is a leading PACS consultant and has authored nearly 300 articles on PACS technology in the past 16 years. He can be reached via e-mail at [email protected].

The comments and observations expressed herein do not necessarily reflect the opinions of, nor should they be construed as an endorsement or admonishment of any particular vendor, analyst, industry consultant, or consulting group. Rather, they should be taken as the personal observations of a guy who has, by his own account, been in this industry way too long.

Related Reading

The 2009 PACSman Awards: Red, black ... or both? December 3, 2009

Building a Better PACS: Part 4 -- Warranties and liability, September 15, 2009

Building a Better PACS: Part 3 -- Look before you leap, July 2, 2009

The DRA and PACS: One hospital's story, April 14, 2009

Building a Better PACS: Part 2 -- Standards, guarantees, and flat-rate pricing, February 19, 2009

Copyright © 2010

Page 1 of 775
Next Page