Minnesota’s Bold Experiment: Radiologist as RBM
In June 2008, the Government Accountability Office (GAO) released a long-awaited study on imaging utilization. If CMS follows GAO’s recommendations on effective methods for curbing overutilization, private payors are likely to follow. That’s why the radiology community was disturbed to hear GAO’s statement that overutilization is a far-reaching issue that can be best addressed with stringent prior-authorization programs. “Several plans attributed substantial drops in annual spending increases on imaging services to the use of prior authorization,” the report states. “In contrast, CMS employs an array of retrospective payment safeguard activities that occur in the post-delivery phase of monitoring services and are focused on identifying medical claims that do not meet certain billing criteria. The private plans’ experience suggests that front-end management of these services could add to CMS’s prudent purchaser efforts.”¹ There is no question that inappropriate utilization of advanced imaging services is occurring, but is it as widespread as the GAO report suggests? Who is responsible for the problem, and what’s the best way to manage it? That question of management is what three Minnesota payors—HealthPartners, Blue Cross Blue Shield (BCBS), and Medica—were asking themselves in 2004. Although they all considered some form of prior-authorization program administered by a third-party radiology benefit management (RBM) company, in the end, they chose a different route: offering medical groups the option to run physician-focused decision-support solutions. A New Age Liz Quam, director of the Center for Diagnostic Imaging (CDI), Minneapolis, says, “Health plans all over the nation were realizing that they needed to do something about utilization, but they didn’t know how. RBMs had become the most politically correct way for health plans to address utilization, not because it’s the best idea, but because it’s the easiest to implement without getting grief from anybody.” Patrick Courneya, MD, a family practice physician in Roseville, Minn, and medical director of HealthPartners, was involved in the conversation about how best to prevent overutilization. “We started looking at diagnostic imaging as an issue in 2004,” he recalls. “We had some promising discussions about decision support, but they didn’t go very far, so we began to move ahead with prior-notification programs. Multiple payors in the market were doing it at the same time.” He stresses that the program initially adopted by HealthPartners was a prior-notification (not prior-authorization) system, so decision-making power would ultimately rest in the hands of providers, not a third-party RBM. “We were never planning to deny any orders,” he says. “We simply required that providers contact our utilization-management vendor. If the provider and that company disagreed, it would be duly noted, but the claim wouldn’t be denied.” Jim Tierney, CEO of Suburban Radiologic Consultants, Minneapolis, points out that this system still represents a significant hurdle for referring physicians. “The prior-authorization programs adopted by the three payors required any physician referring a patient for an MRI, CT, or PET scan to go through a Q and A in order to generate an authorization number,” he says. Tim Signorelli, CEO of Consulting Radiologists, Minneapolis, adds that Medica’s original plan also created inconveniences for imaging practices, and there was no way to know whether BCBS and HealthPartners would address the issue when they went live with similar programs. “[Medica] said that if a patient came to an imaging center without an authorization number, the center should not do the study, and if it did, the study it wouldn’t get paid,” he recalls. “That puts the imaging provider in a very difficult position.” Quam says that the tide began to turn when her team asked for permission to test a new decision-support tool. “It was very well received by everybody, and because of it, some of the other large imaging providers in the state looked into getting the same thing,” she notes. “It isn’t because it’s a better tool. It’s because it feels better if it’s the radiologist saying, ‘Are you sure you want this CT with contrast?’ than it does if you’re in Minnesota calling Houston or Connecticut to hear the same thing. It’s human nature, and it’s also efficiency.” Around the same time, Courneya and his colleagues at HealthPartners were paying close attention to Harvard Pilgrim’s health plan in Boston, which had allowed both Massachusetts General Hospital and Partners HealthCare to try decision-support systems in lieu of working with an RBM. “They’d had some success, so we wanted to hold out the option of doing that in our marketplace,” Courneya says. “We decided that we’d let any medical group that could meet our criteria try out a decision-support system. It needed to be based on appropriate guidelines and needed to include decision support at the point of care, and there needed to be a scoring system so that providers would know which were the most and least appropriate studies.” He adds, “We were operating on the belief that, given the right information in the exam room, providers would make the right decisions.” RBM Versus Point of Care In generating its report, GAO looked at the efforts of 17 payors to reduce utilization; 16 of them use RBMs. What the report fails to assess is whether the efficacy of these programs stacks up against the claims made by RBMs.For his part, he’s not sure what to expect from utilization rates in the months and years to come. “We suspect utilization will level off and may even begin to trend back up again,” he says. “We’re comfortable with that, as long as we believe the technology is being used with respect to best-practice recommendations. There’s also the legitimate argument that as we understand these technologies better, we’ll find meaningful uses for them that we don’t have now.” When it comes to our changing understanding of imaging technology, decision support may actually benefit radiologists. Courneya notes that uses for imaging technology change so rapidly these days that it can be difficult for referring physicians to keep up with best uses, but with clinical decision support, new information can be rapidly disseminated. He offers the example of the American Cancer Society’s recent revisions to its breast-cancer screening guidelines. “We were able to make sure the new criteria were implemented in the decision-support process so that women who needed the test were getting it more consistently, and women who didn’t were having a good conversation with their primary care physicians,” he says. Signorelli offers a more ambivalent perspective, saying, “The best I can say, for now, is that we’ve taken lemons and made lemonade.”
"Whenever you add a middle entity between the clinician and the patient, you risk some erosion of the quality of care." —Liz Quam, Director, Center for Diagnostic Imaging“One of the RBMs has been in Washington a lot in the last year, trying to convince Congress that the solution to all of its woes is to hire an RBM for CMS,” Quam says. “Nobody’s getting into the details of whether the RBM method is actually better, but for those of us who understand how the whole system works, we know it would be, unequivocally, a mess.” Maybe that’s why, in the recently passed Medicare Improvements for Patients and Providers Act of 2008, Congress calls for a GAO comparison of RBM-based and point-of-care utilization-management strategies. Quam, however, does not need to see a report to be able to compare the two approaches. CDI has centers in nine states, and in some of those locations, the practice does work with RBMs. “In Minnesota, we’re closer to the referring physicians, and I would argue that that’s better patient care,” Quam says. “Whenever you add a middle entity between the clinician and the patient, you risk some erosion of quality of care. We have more leeway to deal with situations that don’t fit the criteria, and we can do it immediately. With an RBM, it could take days, and some of our patients don’t have that much time.” For the same reason, Quam suspects that the point-of-care model may be a cost saver, compared with contracting with an RBM. “What kind of money are we saving when ordering a scan takes the referring physician half an hour?” she asks. “It’s penny wise, pound foolish if you’re costing three time as much somewhere else in the system to save money here.” Consulting Radiologists opted for an especially expeditious system, combining its decision-support module with its electronic scheduling so that providers can check appropriateness while ordering online. “It takes referring physicians to an algorithm that meets ACR guidelines, and then the payors get a report that verifies that all exams meet the guidelines,” Signorelli says. He adds, though, that this isn’t too different from what his staff was already doing. “We have the name Consulting Radiologists because consulting is a big part of what our radiologists do,” he says. “I’ve often seen them recommend a lower-end study when it was more appropriate.” Tierney sees advantages in terms of referring-physician service. “The decision-support tool allows us to conduct the same Q and A that the RBM would and get an OK for the exam by going through an algorithm,” he says. “Referring physicians still have to do what they’ve always done, but rather than calling a third party, they call us. It eliminates a huge headache for the physicians ordering the exams and maintains their loyalty to our imaging centers.” Tierney is quick to note, however, that providing this service comes with a significant downside. “It’s a good thing that inappropriate and unnecessary exams are no longer ordered,” he says. “The bad thing is that we’re incurring extra overhead to accomplish this, and that needs to be compensated for; we’ve taken over the grunt work.” Meanwhile, the health plans have succeeded in lowering utilization considerably, resulting in significant cost savings. “In 2007, we saw about a 7,000-scan reduction among our membership, based on the projected trend,” Courneya says. “That’s a saving of about $6.6 million.” While the RBMs continue to serve as clearinghouses for the volumes and other data submitted by the imaging providers, they do so at a reduced price, meaning further savings for the insurers. “We pay them to manage that information, but it’s less than it would cost for a full-support approach,” Courneya notes. “We’re continuing to figure out community approaches to providing the capability as broadly as possible and as cost effectively as possible. The rate they charge is based on membership.” Reduced Utilization There’s no doubt that the measures undertaken by the Minnesota payors have had the intended impact. The depth of that impact, however, has taken its toll on the state’s imaging community. “I would say, anecdotally, that most of the local freestanding outpatient centers seem to be reporting anywhere from a 5% to a 15% drop in MRI, PET, and CT,” Tierney says. “The health plans’ costs have decreased. As for us, when our volume’s down 5% to 15%, our revenue’s down 5% to 15%, but our expense base stays the same.” The Minnesota Department of Health collects imaging referral data, but the statistics for the relevant period have not yet been published. Signorelli confirms that there’s been a decrease in utilization, but doesn’t assign it a figure. “In the first quarter of last year, we saw lower utilization and a much slower quarter than what we had experienced previous to that,” he says. “We think a lot of that was due to the implementation of these programs. The people who were the hardest hit were the small practices, because they didn’t have the infrastructure or the wherewithal to absorb this additional complexity.”
Medicare Part B spending on imaging by setting, 2000 and 2006¹:Quam says that her referral sources have been extremely vocal about their feelings on the subject. “They’re so much happier calling us than calling the RBM,” she says, “but it has cost us a lot, and not just in terms of the staff needed to run the system. As the vendor continues to upgrade the tool, we have to install the upgrade. We had to give extra training to the intake people on our end, and it takes up radiologists’ time. Our medical director devotes hours to it weekly.” A theme emerges of practices suffering at the hands of payors trying to cut costs. Was overutilization that much of a problem in the first place? The numbers seem to suggest that up to 15% of scans ordered prior to implementation of the decision-support system were unnecessary, but numbers don’t tell the whole story. “If something’s deemed inappropriate, it bounces to one of our radiologists to talk to the referring clinician and figure out what the problem is,” Quam says. “So far, what we’re hearing from radiologists is they’re not getting any more consults than they did before the tool was put in place. What they are hearing more of is that the decision-support tool is inadequate, so it gets bumped to them, and their advice to the referring physician is just to say ‘other.’”
"We’re coming into a second year, and what we’re hoping is that as we look at year-over-year numbers, our volume levels off. You wouldn’t think it would go down any further, and if it doesn’t we’ll probably be OK." —Jim Tierney, CEO Suburban RadiologySignorelli has also experienced some glitches in the system. “In the chiropractic community, which tends not to be large in terms of practice concentration, what we began to see is some of the chiropractors saying, ‘I’m just going to refer over to a neurologist and have him or her order the study.’ Where the study is done can change, and that doesn’t mean the study wasn’t ordered,” he says. The impetus for the increased focus on appropriateness of imaging services was the rapid escalation in the volume of studies performed, leading insurers like Medica to claim that between 10% and 20% of scans are unnecessary.² Is overutilization the only reason for this growth? “Were there inappropriate patterns of utilization previously? Absolutely,” Signorelli says. “I think there’s no doubt that there is some inappropriate utilization. Some is predicated on self-referral; some is defensive medicine. How big is the problem? That’s for the researchers to tell us.” A 2008 report from the Minnesota Medical Association attempts to do just that. It states, “For much of the overall volume growth in health care, technological change is often cited as the reason. This fact may be particularly relevant in the technology-dependent field of imaging. Technological changes include both treatment substitution (replacing old technology with new) and treatment expansion (treating more people due to new technologies).”³ The report goes on to recommend both clinical decision support and mandated accreditation as two potential solutions to the problem. In the meantime, Tierney has his fingers crossed that utilization won’t slide any further than it already has. “MRI and CT are usually the bread and butter of a successful imaging center,” he says. “We’re coming into a second year, and what we’re hoping is that as we look at year-over-year numbers, our volume levels off. You wouldn’t think it would go down any further, and if it doesn’t, we’ll probably be okay.” 2009 and Beyond Despite her mixed experiences, Quam remains optimistic about the future of decision-support tools, which she sees as a necessary evil. “We’ll stick with it, and we’re very involved with helping our vendor refine some of the decision-support trees,” she says. “I’m getting a lot of calls recently from radiology groups asking about decision-support tools. I tell them it’s a good idea, but it is a work in progress.” Courneya concurs that the system is far from perfect, and speaks of a multitude of improvements he hopes to see. “Through our community collaboration with the Institute for Clinical Systems Improvement, we are exploring a consistent source of content for decision support (so that the groups in our region have one source they can use to populate the programs), and we’re looking at more robust evaluation of the impact of the studies on clinical decisions and outcomes,” he says.