Meaningful use stage 2, the linchpin to the next round of federal financial incentives for health IT adoption, involves radiologists more heavily than stage 1. Computerized physician order entry of at least 30% of radiology procedures is required of physicians eligible for meaningful use, meaning radiologists need the capacity to securely receive electronic orders for diagnostic tests.
Furthermore, stage 2 requires diagnostic test results, including radiology reports -- the text accompanying the scan or X-ray the radiologist writes evaluating the image content -- must be available to patients within 24 hours. Whether that includes the image itself is up for debate. Sources within CMS, the government agency overseeing the ONC, said it's left to the discretion of the provider to determine which types of information should be included.
Far more complicated is the stage 2 optional "menu measure" for sharing radiology images and the text reports that go with them. The measure does not require that EHRs store radiology images, but it does mandate that primary care physicians (PCPs) or attending physicians in hospitals be able to access at least 10% of their patient’s studies (the image plus accompanying text) via links in the EHR.
The idea, said Joseph Marion principal of Healthcare Integration Strategies LLC, a consultancy based in Waukesha, Wis., is to give patients their data but also to prevent unnecessary, redundant tests by providing several patient's physicians with access to past tests. Imaging studies can be big-ticket diagnostic tests, and payers are hoping this particular facet of meaningful use will unlock costs savings by unlocking a patient's past test results that are currently locked in picture archiving and communications systems (PACS).
Imaging software vendors initially fought the idea of sharing radiology images, Marion said. They argued that letting patients have access to images would require them to download and use complicated proprietary viewers.
"That's baloney," he said, because the viewers are needed only for diagnostic purposes, and patients don't need to, for example, zoom in 6,000% on their MRI scan. "There's no reason your average consumer would have to have one, because they're not going to look at it form a diagnostic perspective. As long as there's linkage from the EHR, they can launch a simple [Web] viewer and display the images."
But opening up these systems contrasts with the way radiology reports customarily are delivered from the radiologist, who interprets a scan or X-ray, to the patient's PCP, who in turn interprets the report's technical medical terminology for the patient. Stage 2 breaks that pattern, leading some radiologists to experiment with sharing reports directly with patients.
The idea made some attendees of last year's Radiological Society of North America conference bristle. Some attendees worried delivering those results, especially when they are confirming cases of diseases, would be the wrong thing to do, ethically. Others worried taking results straight to the patient might foul business relations with referring PCPs. Despite the change in culture, radiologists who want to avoid CMS penalties will have to turn to IT in order to open up their data silos.
The movement toward more open sharing of test results between physicians and patients would be happening regardless of meaningful use's specifics, according to Dan Morton, the director of medical informatics at Penn Medicine, a health system employing nearly 2,600 physicians across three hospitals and outpatient facilities in greater Philadelphia and southern New Jersey.
"I don't think it will be meaningful use that drives it so much; I think it will be consumer demand," Morton said. The menu measure is optional in stage 2 and could technically be ignored for now by radiology practices, but "I don't think that would position you well against your competitors," he added.
In terms of IT support, radiology image-sharing poses a host of technical issues for Penn Medicine. For starters, most systems don't support the same standards, making it difficult for radiologists to access a health system's PACS and EHR-like radiology information systems (RIS).
Moreover, the variety of system vendors and locations across health systems of Penn Medicine's scale make PACS integrations a complex IT project that requires making sense of a knot of application programming interfaces and connectors designed to ensure radiologists can access the patient images they need to evaluate. PACS integration was needed, Morton said, because the radiology department decided to base its workflow on RIS technology, which is used to organize and share diagnostic and imaging tests. The plan called for the creation of one master medical imaging archive and clinical viewer solution (as opposed to a more complex diagnosis viewer) that could be used to call patient studies from the many PACS systems scattered throughout its facilities.
"One of our team members came up affectionately with the term 'spackleware'" to describe Penn Medicine's PACS integration, Morton said. "Every application has holes in it, and sometimes you need to bridge applications with a certain amount of spackle just to tie things together."
One example, he said, was two instances of a GE PACS running at two hospitals, which have since been merged into one. When they were separate, code was needed to enable radiologists to be routed to the correct PACS when calling an image from the RIS, setting the context for viewing and calling the image on screen.
Morton's colleague and Associate Chief Information Officer of Operations Tom Riesenberg said imaging operations across radiology, cardiology, ophthalmology, emergency department, OBGYN, dermatology and other departments use at least 12 different PACS systems. They're all stable systems and working well. IT's job is keeping the business running smoothly, keeping the software up to date, and ensuring closer integration with Penn Medicine's Epic Systems Corp. EHR system, which also pipes image data and test results into a Web-accessible patient portal.
The Epic integration covers Penn Medicine's meaningful use needs. Long term, however, the health system is evaluating a vendor-neutral archive (VNA), into which it will merge all medical image data from existing long term PACS managed archives. Each department PACS will continue to provide specific departmental functions, but its standards-based relationship to the VNA will also make patient studies accessible to clinicians throughout Penn Medicine, as well as the patients. The eventual VNA implementation represents a strategic infrastructure investment to keep medical images and documents in one place, accessible with one EHR interface and viewing tool. This will eliminate the "spackleware" situation, satisfy regulatory needs and set the stage for coming big data projects.
"We have a lot of fragmented archives out there, all around the place," Riesenberg said. "That's creating certain complications for us economically speaking, but it's also creating complications in terms of clinical review of these studies across these many different PACS systems makes it very, very challenging, requiring unique EHR interfaces for different PACS viewers, each with a different graphical user interface, unique provider training, and other technical complications along those lines."
VNAs could be the answer to enabling meaningful use attestation for healthcare CIOs faced with PACS systems that don't necessarily integrate well with EHRs and other data systems. They make sharing images with primary care or attending physicians easier, Riesenberg said. Another payoff is that VNAs also give a radiology department more vendor-independence when choosing future PACS systems, which isn't always the case today "when the PACS vendor owns the archive," he said.
Penn Medicine will likely migrate out of its current PACS-centric long term archives into a VNA one PACS archive at a time. The project will be driven by business reasons, such as avoiding the high cost associated with replacing PACS archive storage systems when they run out of space or become too expensive to maintain (approaching end of life).
"At each one of these business events, you seize the opportunity to migrate [a PACS archive] to the VNA," Riesenberg said. "It might take us four or five years to migrate them all. That's the vision; it's usually not driven by image-viewing interoperability or a lower SAN cost or something like that. It tends to be more investment-driven. You've already bought the [VNA] software, all you have to do is add some storage, maybe some servers and redundancy, and add the next PACS archive in there."