The world is changing rapidly, and with it, the U.S. healthcare system—and with the healthcare system, so, too, the world of radiology and imaging informatics. Among the most bracing trends, emerging out of a very broad range of different sources, that are affecting the practice of radiology and the use of imaging informatics:
A massive and unprecedented shift from volume to value in the policy and reimbursement area is taking place now. This shift had already been underway for several years because of the implementation of the mandatory value-based purchasing program and the avoidable readmissions reduction program for physicians and hospitals, under provisions of the Affordable Care Act (ACA); but that shift has been accelerated by the passage of the MACRA (Medicare Access and CHIP Reauthorization Act of 2015) by Congress in April 2015, with the MACRA final rule being published on Oct. 14, and set to be implemented beginning Jan. 1, 2017, and with all Medicare-participating physicians then being pushed either into advanced payment models (APMs) or into the MIPS (Merit-based Incentive Payment System) program. And those changes only compound the policy changes facing radiologists, as referring specialists will soon be required to make use of appropriateness criteria when ordering diagnostic studies.
> Partly driven by policy and payment shifts, a massive consolidation of the healthcare provider sector is underway, with physicians, including radiologists, in the smallest practices, joining either larger radiology groups or large multispecialty groups, or choosing direct employment by hospitals, and with imaging centers also consolidating.
> A combination of advances in information and communication technology is making it far easier for radiologists to work remotely, and spurring a shift of many radiologists into virtual groups or working as individuals for remote-reading services.
> Advances in information technology and in interoperability are changing imaging informatics, with PACS (picture archiving and communications systems) systems increasingly being folded into VNAs (vendor-neutral archives), which are incorporating not only radiological images, but also images from such medical specialties as cardiology, pathology, dermatology, and gastroenterology.
> Changes in technology and business are accelerating an already vigorous pace of consolidation among imaging informatics vendors, resulting in a far smaller number of increasingly larger solutions providers, and, with electronic health record (EHR) products increasing in sophistication, a plummeting level of interest in standalone RIS (radiology information system) solutions, except at the smallest, non-networked radiology practices.
> At the very leading edge of technology, the Armonk, N.Y.-based IBM announced in June the creation of the new Watson Health Medical Imaging collaborative, a global partnership that is bringing together 16 vendor, health system, and academic partners to improve cognitive imaging for radiologists and referring providers in many specialties. That collaborative’s leaders are looking to leverage big data to improve clinical decision support for radiologists and for referring physicians. Other initiatives are aiming at bringing machine learning-based tools into radiological practice in earnest. And the implications of all such initiatives are major, in terms of how radiologists will work in the future.
All of these changes in U.S. healthcare are fundamentally changing the landscape of radiology practice and of imaging informatics, at a time of both innovation and uncertainty for radiologists, always historically among the most technology-embracing of medical specialists.
What Do Industry Leaders Think Will Happen Next?
Rasu Shrestha, M.D., the chief innovation officer at the 20-plus-hospital UPMC health system in Pittsburgh, and a practicing radiologist, says that the release of the MACRA final rule will be excellent for radiology and for radiologists, particularly as it replaces some of the quality-focused measures under the meaningful use program under the HITECH (Health Information Technology for Economic and Clinical Health) Act with a new set of measures that he believes will more accurately measure radiologists’ and other physicians’ contributions to the quality of care delivery. “MACRA really is about moving to a continuum of scoring versus an all-or-nothing system, as Farzad [Farzad Mostashari, M.D., the former National Coordinator for Health IT] said [recently], and I think that’s spot-on. Meaningful use has been about looking at quantity—the numerator-denominator game we’ve been playing. Now we’re actually looking at quality, so there’s a movement and a validation, and a push, for us to actually get to value. I think this is actually awesome for radiology and radiologists,” Shrestha adds. “This basically substantiates a lot of the discussion that radiology has been leading in the last couple of years, moving from volume-based to value-based imaging.” In fact, he says, “I think radiology has also been the bellwether for healthcare on this.”
That doesn’t mean that the shift into value will be easy for radiologists; quite contrary, says Ezequiel Silva, M.D., a practicing interventional radiologist at the 71-radiologist South Texas Radiology Group in San Antonio, and, since April, the chairman of the Commission on Economics of the American College of Radiology (ACR). “The challenge we’re seeing with new payment models and paradigms is that the shift towards them is happening very quickly and evolving very quickly,” says Silva, who has been in clinical practice since 2001. “Most of us who have been practicing for 10 or more years have gotten used to the fee-for-service system, and we’ve been fully comfortable with those kinds of mechanisms. So now with the ACA and in particular with MACRA, we’re seeing more of a focus not just on quantity of CPT codes, but on quality and value within that.”
Silva concedes that, “On one level, you want to say, well, we should always have been focused on quality. But most of the mechanisms around quality were based on giving bonuses, and then later on, based on minor little penalties around quality. But now, we’re seeing a wholesale shift towards value. And from an informatics perspective, we’re seeing more and more of a drive towards using information technology. And we’re even seeing the use of IT being a piece of new payment mechanisms. The challenge that practices are facing is that they don’t really have a playbook yet for doing what’s needed,” he adds. “And the regulations are evolving so quickly that it’s really hard to move forward and do what is being asked.”
Can Radiologists Play a Role in the New Population Health-Driven Healthcare System?
Indeed, Eliot Siegel, M.D., the chief of imaging for the VA Maryland Healthcare System, and vice chair of research information systems for the University of Maryland Department of Diagnostic Radiology and Nuclear Medicine, believes that radiologists can play a significant role in the emerging universe of population health management. “I really believe that finding some carotid plaques way earlier before you develop carotid stenosis then have to intervene with a stent, being able to find patients who are at-risk of coronary artery disease by looking at calcium scoring way earlier, that’s a way to save money,” Siegel says. “If we have a certain amount of money, we have patient populations to track, I believe if we use detection early on and end up intervening, imaging would actually result in saving some dollars. I think in 10, 15, 20 years, maybe five years from now, the emphasis is going to be on spending a lot of money to keep healthy people healthy for a longer period of time.”
Of course, that’s partly where the advance of information technology, and the adoption of new tools on the part of radiologists, especially clinical decision support (CDS) tools, will be so important. Asked what he thinks CIOs need to know about the current landscape around radiology practice, Siegel says, “I’d like to work with CIOs to figure out how we can extract information from the radiology images and reports. I’d like CIOs to know that radiology is becoming increasingly structured, and so there will be data that they can start to make machine intelligible and in their algorithms, when a patient gets submitted to a hospital with a stroke, you can use algorithms. There is data in radiology that they can discover; it’s not just dark matter, it doesn’t exist to answer just one question. There is a wealth of information, and that data, pixel data, should be thought of a great source of information in EHRs. And yet,” he says, “In the same way that EHRs currently don’t support the use of genomic information, we have to struggle to figure out how to do that, the same is true for radiology: we’ve got a lot of interesting, complex information, and it’ll be great to work with them to figure out how we make that information discoverable, actionable.”
At UPMC, Peering Into a New World of Machine Learning-Driven Medicine
At UPMC, Shrestha and his colleagues are already on the case. “As chair of the RSNA Scientific Program Committee for Informatics, I have insights into all the scientific papers and presentations that are submitted,” Shrestha says, referring to the Oakbrook Terrace-based Radiological Society of North America. “And one of the most important trends that’s emerging this year is specifically around machine learning; and the context here is leveraging pattern recognition.”
Explaining what the pioneers in that area are doing, Shrestha says, “Machine learning is all about looking at relationships that happen between Fact A and Fact B or Data Point A and Data Point B, and looking at patterns, and drawing conclusions. And what better place to really push with this notion, than in radiology? And that’s because you have so many different types of data elements in radiology. And you have capabilities to leverage machine learning to look at relationships between data and text—and imaging data as well. So progress is being made in leveraging machine learning in terms of correlating text-based data with meta data, and also in terms of looking at patterns around data itself. So we’re not just looking at images one image at a time—for example, prostate detection or lung cancer nodule detection, in CAD, or breast imaging detection—we’ve had that for the last ten years or so. But now with machine learning, we can look at countless images at once.”
The implications of these emerging capabilities are startling, Shrestha says. “What does it really mean for us to be able to mine through these immense amounts of data sets, and have machines learn patterns, so that it will then aid us in diagnosis and tell us, is this lesion increasing or decreasing? Is this a lesion we should consider, as an element in diagnosis? These are sort of the early stages of, say, Watson coming into radiology. I don’t think it means the radiologist will be displaced anytime soon,” he reflects. “But for sure, there is a role for machine learning and artificial intelligence in order to separate the signals and the noise; and also in terms of patient-related data and context around the specific pathology of the image; and last but not least, helping in the diagnostic process, based on not just one or two, but millions of similar images and studies.”
Changes in Technology are Changing the Business of Radiology
In that regard, how are changes in the business and technological landscapes impacting radiology practice? Joe Marion, principal at the Waukesha, Wis.-based Healthcare Integration Strategies, LLC consulting firm, says that a combination of technological, business, and payment trends is going to seriously upend radiological practice going forward. Not only are radiologists banding together to provide remote-read services all across the U.S.; those newer arrangements are changing the whole concept of proximity, as immediate geographical proximity is no longer a limiting factor for where and how radiologists practice. “I’m working with a group in New Hampshire,” Marion says, “where the radiology group serves five facilities in the area; but those radiologists recently hooked up with a group in Connecticut, and will be doing readings through that group in that state as well. Furthermore, people are realizing that they need backup capabilities, and want to own their own data.” So the whole concept of location is morphing now when it comes to the provision of radiological services nationwide, Marion says.
“And,” he adds, “you’ve got hospitals signing up with other entities for larger integrated delivery networks. And the question is how you manage radiology services across all those facilities. So load-leveling, and managing services. We’ve got a major integrated health system here in southeast Wisconsin that recently created their own doc-in-the-box in the Brookfield area [west of Milwaukee], and all the main studies, except for interventional studies that have to be done on site, are done out of that doc-in-the-box in Brookfield.” All that activity is also being impacted by a move into the cloud on the part of radiology practices and multispecialty group practices, hospitals, and integrated health systems.
Meanwhile, he says, the growing universalization of VNA and other broader architectures will in turn change who delivers services and how those services are delivered, Marion says. “So, for example, a hospital is concerned about getting reports economically and smoothly. So if a vendor could provide such services, why not? I know that a major PACS vendor looked into this years ago, and it came down to a marketing issue—whether they might be perceived as competing with their customer—and that sort of put the kibosh on a lot of it. But the concept is there.”
Changing Radiologists’ Incentives
Meanwhile, the opportunities and challenges of federal and private-insurer reimbursement systems will both be significant in terms of radiologists in practice. “When you look at the quality buckets under the MACRA law, the practices that have the informatics tools in place to allow for reporting to qualified clinical data registries, will find themselves in a position to perform more successfully on quality,” says Silva. “And that involves two elements: first, you have to report satisfactorily enough to retain your payments. But second, you want to engage in benchmarking, to improve processes and outcomes. And it’s difficult to do that from a traditional claims-based standpoint. But if you’re improving radiation dose or predictive value around lung cancer screening or CT radiography, etc., then it becomes, how do you show your quality externally, to your hospital systems, to your patients, to larger payers? Because we’re talking about alternative payment models.”
UPMC’s Shrestha believes that this decisive shift towards quality will actually empower radiologists, many of whom have been stressing out over the increasing commoditization of radiology services in recent years. In many ways, until the automation and digitization of processes, images, and data, have only served to speed up processes for radiologists, not change them—or radiologists’ place in care delivery. “And what that has propagated is the reality of how radiology has become commoditized in the last few years,” he says. “The radiologist or radiology service who provides the quickest reads at a good price—that could be a guy in India or the Philippines, or wherever. But the next phase is about a focus on value—on these clinical quality metrics. We’re tying payment to outcomes, and using evidence-based medicine and tying that to quality of care. And in that context, we need more collaboration, not less, collaborating with the ordering physician, collaborating with the surgeon who’s going to operate on the patient the next day. So the radiologist becomes not just a service provider, but a consultant. That’s the evolution that is allowed by technology and payment changes.”
A Call for Technical Standards to Overcome Remaining Obstacles
Yet for all the talk of the future, healthcare IT leaders say that numerous obstacles remain in the present, and those are becoming more urgent as the new policy, regulatory and reimbursement changes sweep the industry. The simple fact of having to continue in many cases to rely on the use of CDs to transport diagnostic images rankles, says Kent Hoyos, vice president of information technology and CIO at the 437-bed Pomona Valley Hospital Medical Center, a freestanding community hospital located in the Los Angeles suburb of Pomona. “The sharing of images across images, via health information exchange or some kind of DropBox lite—it remains challenging to move images around, in many cases.”
Indeed, Hoyos says, “We were in a revenue cycle meeting, and we ended up talking about the process of care delivery around mammograms. We ended up having to change in our revenue cycle system the length of time it was acceptable for a bill to be paid, in cases in which a mammogram had been done outside the hospital, because the process” of acquiring the images and moving the billing process along around outside mammogram procedures “is still clumsy.” The reality, he says, is that “We need to have [technical] standards. That’s an easy answer but hard to actually do. We need standards that are transferable and that are collaborative in the truest sense. When you can have a diagnostic imaging procedure down the street at the imaging center, but we’re still relying on moving CDs around, that’s terrible.” What’s more, Hoyos says, he believes that health IT vendors are dragging their heels on achieving true, vendor-neutral interoperability, because “They’re protecting their revenue stream however they can, just as we are. You wouldn’t get an argument from anybody about what’s right for the patient. But it’s all about money, about protecting the revenue streams around our populations.” In his view, it will take federal government pressure in order to finally break the remaining logjams around the interoperability needed to birth a truly new era in imaging informatics across U.S. healthcare.
An “All-For-One Kind of Thing”
Certainly, healthcare IT leaders will need to be front and center in order to help radiologists—and everyone—optimally navigate the next several years in U.S. healthcare. “What we’re seeing a lot of at the physician level,” says Silva, “is that a lot of what MACRA drives is physician-driven payment metrics. Initially, what we’ll see physicians say” to CIOs and other healthcare IT leaders, “is, help us to score better on this metric by giving us information systems. And the IT people already have a lot on their plates. But by the next three years and certainly five years, you’ll see the need for physicians and hospitals to partner. So that’s an important message to communicate: that this is very quickly becoming a one-for-all kind of thing” for clinicians and everyone else in the patient care enterprise.