September 2010

Moving Toward Interoperability
By David Yeager
Radiology Today
Vol. 11 No. 9 P. 8

Patients already shuttle among multiple doctors, and healthcare reform means an estimated 35 million people will swell insurance roles over the next four years. The need for fast, efficient image transfers will only increase. The requirement for interoperability between PACS is moving from luxury to necessity. While technological issues are among the obstacles to widespread interoperability, technology isn’t the biggest challenge to systems that more seamlessly share data becoming standard practice.

“We have examples of some of the infrastructure that’s necessary in a local PACS system,” says Kevin McEnery, MD, an associate professor and the deputy division head for informatics at the University of Texas M. D. Anderson Cancer Center. “Most institutions have the ability to send images from a modality to another modality to a workstation. And just as if you do a query/retrieve for images of your local PACS to a local workstation, one could envision that to be extended to the ability to exchange images between institutions.”

A major barrier to better image sharing is the lack of ground rules to foster PACS interoperability. Although DICOM has been around for a while, vendors have traditionally inserted proprietary tags that prevent systems from effectively reading other manufacturers’ images. And while vendors have taken small steps toward image neutrality, there is no consensus about the best way to proceed.

To allow collaboration, Koller says PACS vendors will either have to build filters that read their competitors’ tags or institutions will need to find intermediaries that adjust the data to a standard format. Middleware vendors are beginning offer a neutral online space for imaging data, but whether this type of cloud computing will become widespread remains to be seen. One limiting factor is that it requires institutions to opt in to such a solution, which can be problematic if a patient is seen at multiple facilities, especially if the facilities aren’t familiar with each other. It takes time to develop working relationships among institutions. McEnery believes building such trust between organizations will be a significant challenge. Alleviating those trust concerns will go a long way toward spreading interoperability.

“There is always the option of throwing away all the proprietary stuff and just using the standard stuff, which will give you the image, but it may not give you all the manipulation, annotation, and other stuff that was done by the previous system. It may, it may not. But the core image that was originally acquired should be there in an unaltered form, available to the second system,” says John S. Koller, president of KAI Consulting. “But if the purpose of this exchange is collaboration, there’s a lot of value lost when you can’t access what the previous physician has done.”

“If you can have a federated system where you have a central trust broker, and you create an information exchange based on that trust broker intermediary,” McEnery says, “I suspect that a lot of institutions will go that way.”

Digital image transfer is faster and more efficient than the traditional methods of transporting films or CDs. McEnery believes patient demand will be a key driver of interoperability, since patients control who can or can’t view the images. He believes patient permission for a hospital to access an intermediary may eventually become part of a standard intake form.

Proprietary Tags
While there is growing demand for image neutrality among PACS customers, proprietary tags are still prevalent. Some larger PACS vendors are building neutral archives, but the trend is far from widespread in the industry. A trusted intermediary can help with this problem as well.

“There are still vendors who do proprietary things in a standard format,” says Koller. “Until that totally goes away and everything is truly standards based on how you store and manage and move that data, you will have some intermediary or some tool that normalizes that data.”

Due to the size and complexity of the U.S. healthcare system, many people in the field believe it’s likely that true interoperability will require a mandated solution along the lines of an Integrating the Healthcare Enterprise (IHE) profile. One attractive feature of IHE is that it provides an impartial space for bridging differences between institutions and standardizing data from disparate vendors. The IHE XDS-I profile, which provides guidelines for image transfers, is already routinely used in Canada and Europe and is starting to make inroads in the United States. Koller believes that trend will continue.

The United States may have recently missed an opportunity to promote image interoperability. There is no provision for it in Health and Human Services’ EMR meaningful use criteria, which McEnery says would have sped up the process. He believes radiology’s example of using DICOM to move terabytes worth of data between disparate systems may have helped guide the planning for other types of data interoperability. But even if the drive for meaningful EMR use eclipses PACS interoperability in the short run, he’s convinced that the barriers between systems and institutions will eventually dissolve.

“Clinicians will use systems to look at images, and where the images originate from will be irrelevant,” says McEnery. “I think that in the long term you’ll have the ability to have knowledge of images in the different systems. And so when a radiologist needs to view an image and needs comparison studies, those studies may not necessarily come from their local system, they may be retrieved from other institutions.”

— David Yeager is a freelance writer based in Royersford, Pa.