Enterprise Workflow: Imaging’s Next Frontier?
Three hospitals using two separate PACS are a recipe for headaches, and Steven C. Horii, MD, is determined to find relief. As clinical director in the Medical Informatics Group at Philadelphia’s University of Pennsylvania Medical Center, Horii wants his enterprise imaging platform to boost crucial connectivity, while maintaining most of the capabilities of the “full blown” workstation. The transformation is ongoing, but Horii is successfully leveraging his department’s already strong enterprise workflow to boost organizational growth, all while driving quality and efficiency. The University of Pennsylvania Hospital, Presbyterian Hospital, and Pennsylvania Hospital have all benefited from the new enterprise imaging capabilities—with patients and professionals alike appreciating better communication. With patients often going back and forth between the facilities, the process of moving images around must be smooth. “With two different PACS, we could move the images, but it was not trivial,” Horii explains. “The only way we could do it was to use the electronic medical record (EMR) system that our hospital web applications group had put together. I give a lot of credit to the hospital IT developers, along with the folks from radiology IT. They built a lot of what we could not reasonably buy." The EMR setup worked, but not perfectly. One problem was that different departments had their own databases and information systems. “Even though there is a medical record number that is uniform across a given facility, it is not uniform between hospitals,” Horii notes. “We are just now converting to an enterprise identifier that will be uniform across all three facilities—including our outpatient facilities.” It’s a challenge, but the main objective is to get everyone on the same PACS, to facilitate the same capabilities wherever radiologists in the health system may find themselves. Cross-coverage will improve, in addition to the efficiency of processes and the familiarity of viewing tools. “We have had to scale up to do that,” Horii says. “We are putting all these images into a storage system, and that must be increased in size to meet our needs.” The viewing tools are Web-based, requiring that radiologists load the active controls; the same is true of the enterprise viewer, which can be utilized anywhere with Internet access. Those outside of the enterprise, either at home or at a different facility, can access images via a virtual private network. As the imaging platform evolves, Horii favors a “thin client” system akin to a Web viewer that is easy to install, but not as memory-intensive. “The Web viewer works fine, but just does not have as many options for how you display things,” he explains. “It’s designed as a single screen system, whereas our principal PACS workstation software will run on a single monitor, and up to four.”Clinical Case in Point Radiology departments looking to scale up their enterprise imaging solutions can boost patient care dramatically with the right additions. Horii has seen the benefits firsthand, with real-world results hanging in the balance. In the case of a 70-year-old man who had an MRI study at a non-network facility, an outside radiologist first indicated a tumor in the left kidney. “We made those original MRI images available to our urologist, and the urologist wanted one of our radiologists to look at it,” Horii says. “We had a policy to generate an order number, so the image got entered into our PACS, which allowed viewing throughout the enterprise.” The urologist and radiologist from the University of Pennsylvania saw the image and agreed that the man had a tumor in his left kidney. The urologist planned to remove the small tumor and preserve the rest of the kidney. The urologist prepped the man in the operating room, but decided to summon another radiologist prior to the actual surgery. “I got called and went up to the OR to perform an ultrasound, because the urologist wanted a better idea of exactly where the tumor was, and how close it was to other structures,” Horii explains. “I scanned the kidney and found the tumor, but I also found a second mass in another part of the kidney that was not described in the MRI, and I was concerned.” Since Horii had the ability to look at images in the OR, they called up the original MRI for an immediate comparison. The second mass was not evident on the MRI, and the scan had been done within a month, so both clinicians agreed it should be visible. “I took another look at the ultrasound, and I said, ‘I think this thing is really there,’ and the surgeon said, ‘If that’s real, I’m going to take the whole kidney out,’” Horii recalls. As the urologist pondered his two options, Horii explained that the small size of the second mass was likely the reason it failed to show up on the MRI. In the end, the convenient comparison on the enterprise system helped to make the final decision. In this case, it was the lack of clarity on the MRI that actually sealed the deal. “If the urologist looked at the MRI and decided, based on the MRI, that the secondary mass looked more like a cyst, then I probably would have been convinced that it was nothing and advised to leave it alone,” Horii explains. “Because I did not see it on the MRI, I knew it was not a cyst—and that helped me. A cyst would be obvious on the MRI. The fact that the original radiologist did not see it made it more likely that it was something to be concerned about.” Ultimately, the urologist took out the whole kidney. Pathology results, accessed through the upgraded enterprise system, showed that the second mass was indeed cancerous. At minimum, the patient was saved from another operation; at worst, without intervention the cancer could have spread. Horii created a teaching file from the case that illustrated the benefits of interoperative imaging. Through the enterprise system, he linked up pathology reports, ultrasound results, and the MRI. “People went back and looked closely at the original MRI study and where I found this secondary mass,” Horii says. “They concluded that there was something there, but it was so small they could understand how our radiologist and the outside radiologist were reluctant to call it. It was too small to definitively call as a second tumor. Having those images in the OR meant that we could conveniently, and quickly, compare them.” The Light Box Dilemma Back in the 1970s and the 1980s, skeptics questioned how digital images could possibly match the wide physical distribution of light boxes. Horii recalls one speaker challenging attendees to count light boxes in their facility. “I got to the third floor of 18 and I hit 1,000,” Horii says with a chuckle. “The reasoning was that if we had to replace every light box with a workstation, it would not be affordable. Figuring out how to solve that problem has been the modern dilemma.” When the Internet came along, reproducing those light boxes became easier. The evolution only continued with new viewing stations, laptop computers, smart phones, and electronic tablets. Even today, however, Horii maintains that some radiologists still tend to view enterprise imaging as “dumbed down.” The attitude stems from the historical challenges of pushing images through a Web interface (thin client) and displaying them on a single-screen PC. “You often don’t have all the capability that you would have at a workstation,” Horii says. “But you can see images, and it is often a misconception that they are not as good. The complaint I hear from other specialties is that it can be cumbersome.” The inherent differences between an enterprise imaging platform and a full-blown workstation are still sizable, but the conveniences and workflow improvements of the enterprise system are well worth the effort, Horii says. “We do a lot of cardiac imaging and spine imaging where 3D is useful, and we have a thin client 3D viewing application,” he notes. “You can run it over the Internet, and through a small machine. It runs well on a PC connected from the outside, so you can even spin those spine images around at home.” Greg Thompson is a contributing writer for Executive IT Insight.