HIMSS TECHNOLOGY FOCUS:
The Vendor Neutral Archive
Are the Days of PACS Numbered?
by Jean Phillips, editor
The advantages of an integrated, patient-centric digital medical record across the continuum of care are varied and numerous. Because of the fragmented American healthcare system itself and the clinical IT infrastructure that supports it, a patient-centric record has been beyond the reach of most healthcare enterprises. By standardizing and integrating patient information at the archival level, a Vendor Neutral Archive (VNA) provides a relatively straightforward and cost-effective strategy for development of such a patient record. VNA-enabled patient files can be easily shared across medical specialties and add new imaging information to an Electronic Medical Record system. VNAs are also being used to store a range of non-imaging healthcare data. No wonder the technology was in the limelight at HIMSS this year.
The Need for a Consolidated Record
University hospitals. Emergency care clinics. Free standing radiology practices. Government supported healthcare. Hundreds of disparate systems and dozens of applications. The fragmented state of healthcare IT mirrors the fragmented state of the US healthcare system overall.
Indeed, it comes as no surprise that today’s push for a consolidated electronic patient record across a patchwork of disconnected providers—all enabled by equally disconnected IT applications—is fraught with pitfalls and challenges.
The benefits of a single comprehensive digital record with patient-centric clinical information that is distributed across the continuum are readily apparent. Availability of this information to all providers regardless of geographic and institutional boundaries means more informed clinical decision-making and treatment efficacy and efficiency. It means enhanced patient safety with elimination of unsafe drug interaction and cumulative unsafe radiation dose. It means prepping an ER in advance of a patient’s arrival when seconds are precious and less duplication of medical testing—as well as diminishing the cost and management of mountains of redundant healthcare data. Recent government mandates and financial benefits for achieving Meaningful Use have brought the issue of patient records to the forefront. But unfortunately, a single patient file isn’t quite as simple as it seems. Or is it?
Currently, most digital healthcare data is stored in discrete silos tethered to the specific clinical application that created it. John Doe’s radiology exams are, of course, located in PACS, while his EKGs are sequestered in a cardiology system. The exponential growth of specialized healthcare IT systems as the digital age progresses also makes this an exponentially growing problem.
Ironically, the difficulty in solving the problem is inexorably tied to the cause itself. Moving to a more comprehensive enterprise-wide IT system can be extremely costly and challenging because of the difficulty in aggregating proprietary data across multiple applications and then migrating it to more comprehensive applications. So the vicious cycle continues.
Hail the VNA
According to many industry experts, the light at the end of the IT tunnel is what has been termed a Vendor Neutral Archive (VNA). A VNA is a vendor-agnostic data repository serving multiple clinical IT applications both within and across medical specialties. The specific definition of the term varies significantly, as does acceptance of the actual term itself. Some prefer PACS Neutral or Vendor Independent Archive, while others suggest Vendor Neutral Architecture. The thought is that this better describes the implied open approach to information sharing and data ownership as well as the significant paradigm shift inherent in the VNA strategy. As a general definition, most would agree that such a repository is standards-based and accepts, distributes and enables display of information from numerous other clinical systems—whatever the originating department or IT vendor. A VNA is patient-centric, easily displaying comprehensive data for a single patient. Systems typically integrated with the VNA include HIS, PACS, RIS, and EMR, as well as some specialty departmental systems. The goal is to enable the availability of relevant data and the interoperability of clinical IT systems through a standards-based data format, eliminating today’s hodge podge of costly, clumsy IT interfaces and home grown workarounds. Key to that format is a consistent metadata format—which provides a means of identifying the data as relevant to a particular patient, wherever the information enters the IT system
In a perfect VNA-enabled world, treatment location becomes irrelevant. With a consistent patient identifier, the result is a comprehensive treatment record, easily retrievable by any specialist and IT application that needs it.
Today’s healthcare data difficulties are rooted in the history of healthcare and healthcare IT itself. Catering to individual specialists and niches, current IT systems were simply not built to facilitate interoperability and information exchange. Because most perform specific functions within a particular department, there simply was no impetus to enable cross-application communications.
Further, even within a particular system, such as a RIS or PACS, data is often incompatible across vendors. Many believe vendors have purposefully perpetuated this, creating proprietary data formats that keep their customers in an ongoing relationship with them. Clearly, vendors have had every opportunity and incentive to do so, and the result is today’s vendor/hospital love/hate relationships—and the islands of isolated data rather than more logical and useful consistent patient-centric organization.
The Problems of Proprietary PACS
Because diagnostic images and reports are among the most frequently shared information across the continuum of care, the isolated PACS archive—often held hostage in proprietary vendor formats—poses a particular problem. The longstanding DICOM standard fails to fully specify the metadata tags, or fields, used to identify and annotate data, enabling PACS vendors to retain vestiges of their proprietary systems. Additionally DICOM permits the addition of private tags, which will not read on other vendors’ systems, and allows relational data fields to be left empty. All this becomes extremely problematic for information sharing.
While communication of DICOM information across disparate PACS solutions is difficult enough, the promise of an EMR enabled with DICOM images and related information is even more complex. This can require custom interface engines to link unrelated systems and possibly data transformation as well as even duplication of entire archives. Interfaces are costly to purchase, implement and maintain and create security risks and scalability issues. Data duplication astronomically multiplies the storage and maintenance costs.
The entire situation is even more difficult due to the nature of the DICOM beast. Imaging files are enormous and becoming larger—and therefore are complex, costly, cumbersome and challenging to share.
PACS Beyond Radiology and DICOM
Compounding the problem, today PACS is also often called upon to store a growing variety of data—both DICOM and non-DICOM—from within radiology and from other medical specialties.
DICOM data stored may include cardiology, neurology, and an ever-growing list of the ologies. The departments that generate this data often have their own identification numbers, requisition systems and even accession numbers that all must be reconciled and managed within the PACS just to create an integrated record in this single application.
Radiologists are increasingly adding non-DICOM data objects—such as scanned prescriptions and reports—to the PACS archive. Additionally, the data contributed by other departments may not be in DICOM format. Without an accepted metadata management standard to help systems identify these objects, sharing them beyond PACS also is problematic. Many experts note that storage of this data is well beyond the core competencies of these radiology-centric systems and places stress on their infrastructure.
Whether or not non-DICOM information is managed in a PACS, it is of course managed somewhere. If archived in a separate IT application with even more department-specific formatting, the difficulties of introducing it into any patient-centric system are even greater.
A Better Archive
Representing a significant paradigm shift in archiving, the VNA follows methods and data formats that objectify healthcare information independent of the vendor application that created it. Developed by the widely recognized Integrating the Healthcare Enterprise (IHE) initiative, these standards include DICOM, XDS and XDS1, HL7 and others. Using this common data formatting, the VNA provides one comprehensive archive as the single point of information access for all the healthcare IT applications connected to it.
Naturally, existing non-standard data will have to be transitioned to this formatting as part of the VNA migration process. However, medical sites weary of the ongoing cycles of data conversion that accompany a move to a new IT application will be pleased to know that this will be the last data migration they will ever have to perform. And the process can be handled efficiently and cost-effectively by many VNA vendors.
Because the VNA contains diagnostic images, image-enabling of an EMR and other IT systems becomes far less challenging than with a siloed PACS archive. Providing this patient-centric storage and information sharing will put a facility on the road to Meaningful Use.
Versatile Standardized DICOM and Non-DICOM Storage
As a comprehensive system, in addition to DICOM images, a VNA stores non-imaging DICOM data. This includes DICOM Structured Reports, containing CAD and measurement data, and DICOM Presentation States with user manipulations such as shutters and overlays.
The VNA also stores and manages non-DICOM information from across the healthcare enterprise. This typically involves formatting such as JPEG, PDF, waveforms, various video and audio file types, and MS Word—in short whatever modern medicine sends its way. In particular, these versatile storage capabilities make the VNA an appropriate repository for data from specialties as diverse as the lab—growing in importance—and ophthalmology to speech pathology and dentistry.
Whether for DICOM or non-DICOM objects, once the object header is formatted with consistent patient identifiers as defined by the IHE, a simple search in the appropriate fields will gather the information necessary to create a comprehensive patient record. To facilitate this, the VNA includes a master image manager database, which captures information from the object header such as patient name, patient ID, accession numbers and more. Information is recalled against the database using a standard storage query and retrieve protocol.
If IT systems storing to the VNA produce slightly non-standard data, many VNAs stand ready to help. Through DICOM tag morphing, advanced systems can neutralize non-standard DICOM data, reconcile and manage patient identifiers from multiple sites and resolve inherent metadata discrepancies.
It is of note that some VNAs will also allow the altered DICOM header data to retain initial tagging so that it can still interact meaningfully with the system that created it. This allows departmental IT systems to retain some autonomy. Whether and how dynamic data neutralization is handled may vary significantly with the VNA vendor and has important implications for the true neutrality of the archive.
Standardizing Unstructured Data
VNAs typically ingest non-DICOM objects using APIs and FTP. Known as unstructured data, non-DICOM objects must be turned into structured objects with headers providing identifiers similar to the DICOM standard. A VNA commonly uses techniques such as DICOM Wrapping or DICOM Encapsulation or relies on the XDS/XDS-1 standard to structure this information for the image manager database. The choice between encapsulation and conversion is largely determined by whether the data needs to be returned to its native format for use in its original system. Currently, XDS/XDS1 is generally a more costly strategy to implement but is more sophisticated and is rapidly gaining traction.
In addition to storing and retrieving data, an intelligent VNA also can route and pre-fetch studies and manage certain functions based on user-determined rules.
Enabling the EMR
A key goal of the VNA is to provide access to complete, accurate and timely patient data at the point-of-care, which most often takes place through the Electronic Medical Record (EMR) system. While IT has struggled with a broad spectrum of clumsy and costly ways to enrich the EMR with images and related data, the VNA simplifies the task. A simple link to the standardized archive will bring up all these images, imaging-related data and information in the VNA for a particular patient. Embed a lightweight DICOM viewer, and the result is an easy and elegant hospital-wide image viewing solution delivering a rich clinical picture of the patient.
Naturally, this same link and viewer can bring VNA information to any physician portal connected to the archive with a simple click.
Liberating Your Healthcare Data
In addition to aggregating data for a more global view of the patient record, the VNA provides other significant benefits.
True Data Ownership
The VNA brings about a philosophical change in data ownership. With a VNA, IT applications are no longer tied to physically addressed storage and communication with other systems through proprietary integration. A VNA breaks the connection between the application and the archive, freeing data. It provides a virtualized approach to information management with storage fully controlled by hospital IT—not the IT vendor.
Any application using the archive requires only a single point of integration to a standardized service for access.
Streamlining PACS Migration
As a result, a site also has the ability to elegantly switch to any new standards-based PACS or other clinical system without costly data migrations or significant data reconciliation and cleansing. The VNA ends the era where a medical facility was tied to a particular vendor simply because their data was held captive in a proprietary format. Typically a new PACS can be connected to the VNA with a simple plug-in with little interruption of department functions.
Analytics and Trending
With this access to open standards data across multiple applications, medical sites can also treat their data like the business intelligence that it truly is and run cross-departmental and enterprise-wide analytics—examining, for example, financial trends, staff performance and patient referral patterns.
With its greater management sophistication, a VNA is the logical choice to automate data lifecycle management, including retention and purges as well as file duplication according to both legal requirements and institutional policies. A PACS solution may simply flag certain studies for specific handling, while a VNA can save time and costs by directly driving the lifecycle management process itself. Based on complex rules, for example, it can age off older studies to a secondary archive or purge them altogether, as well as duplicate information for disaster protection.
Market research suggests that today’s 300 bed hospital with 30 TB of imaging data will require 3 petabytes of data to support similar patient services by the year 2020. Instituting a VNA, of course, is a long term decision, and vendors are keenly aware that they must support exponentially growing volumes of data. By definition, a VNA is scalable. The best VNAs are built to scale transparently, on the fly and should not require added administrator management time or add operational complexity to the system. According to healthcare IT expert Herman Oosterwjik, “Even if data should surpass a ‘semi-artificial limit,’ such as 1 million images, a VNA will be able to preserve information and that data should not have to be migrated or ported.” (Although, he notes that conceivably a VNA database might have to be ported—which is a simple process in which all information remains intact—if it became too large for the original media.)
In today’s age of shrinking budgets, a major advantage of the VNA is that centralized storage cuts costs. Quite simply, enterprise-wide storage is less expensive than department-based archiving, due, in part, to an economy of scale in purchasing. Naturally, a VNA also eliminates costs for expensive data replication in multiple siloed archives. It also consolidates IT staffing and maintenance, paring down the departmental hours required.