VNAs: Useful, but not a silver bullet
In his latest AuntMinnie post, Michael J. Cannavo brings up the topic of Vendor-Neutral Archives (VNAs), aka PACS-Neutral Archives (PNAs). I agree with Mr. Cannavo’s insightful points about the necessity of information sharing in future healthcare and the described benefits of VNAs. However, our long experience from deploying cross-enterprise archiving and sharing shows both good and bad sides of VNA adoption. Therefore, I would like to add to the discussion by pointing out potential drawbacks of a VNA approach. Whether the benefits outweigh the drawbacks is an open question, but the following is important to consider before a decision is made:
- Archiving vs. sharing: A VNA is an archiving solution for images in DICOM format. However, the real need is often to share the full patient history across multiple enterprises, which includes reports and clinical documents. For every scenario, the need must be carefully analyzed to determine if archiving or sharing is the way to go. If the latter of these is deemed appropriate, do not turn to VNA, but look for solutions based on the IHE profiles for sharing, XDS and XDS-I.
- Data consistency: If the data is stored in a central VNA, there must be a mechanism for making sure that alterations of, for instance, patient demographics are propagated into the VNA. In Sectra’s experience, the consistency issue is complex and error-prone. From this perspective, distributed storage with a clearly defined “original” is a safer and probably less costly solution.
- The data explosion: Increasing data volumes is actually a strong argument against centralized storage. Most would agree that network congestion is the primary concern with respect to performance of clinical IT systems. The best way to minimize network traffic is to keep data close to the source, which is also where it will most frequently be accessed. Imaging data is sent to other locations only upon request. This is also the scenario envisioned in IHE XDS-I. Many Sectra multi-site customers are approaching data volumes where for which centralized storage is not a plausible option, since not even Gbit bandwidth would be sufficient from archiving alone.
Finally, I think the term vendor-neutral can be misleading. The vendor delivering the VNA should not be ignored. If it is not one of your PACS vendors, then you have another vendor to deal with. Bringing in an additional system, such as a VNA, means that there are more interfaces to deal with. There is nothing magical about VNA interoperability – a thorough effort is required as always when connecting IT systems.
There are undoubtedly numerous scenarios for which a VNA/PNA is the perfect choice. But as with many new popular concepts, the ideas about what it can be used for sometimes lose touch with the buyer’s real needs.
