Will Decision Support Deflect Preauthorization?
Decision support might be able to remove the target taped to radiology’s back, according to a tandem presentation that was made by two of the specialty’s respected leaders on May 3, 2010, at the annual meeting of the American Roentgen Ray Society in San Diego, California. Pat A. Basu, MD, MBA, explained how that target came to be there in the first place; he presented “Overutilization: Background, Trends, and Causes.” Basu, an attending radiologist at Stanford University Medical Center and at the VA Palo Alto Health Care System in California, is Stanford University’s course director of health finance, policy, and economics. Now serving as a White House fellow and special assistant to the president, he is the first radiologist to be appointed to that position. By understanding the reasons for radiology’s current struggles, he says, both individual radiologists and health-care organizations can begin to find ways to identify and correct inappropriate utilization of imaging. James H. Thrall, MD, provided a real-world example of how decision support could help decrease radiology’s vulnerability to external control (or outright attack) and declining reimbursement, presenting “Opportunities for Reducing Over-utilization.” Thrall is radiologist-in-chief at Massachusetts General Hospital (MGH) in Boston and Juan M. Taveras professor of radiology at Harvard Medical School. Even if the overutilization of imaging is a problem created largely outside radiology by self-referring nonradiologists, patients’ underinformed demands, and referrers’ errors and malpractice worries, it has become radiology’s problem because radiologists bear the brunt of measures meant to curb utilization, Thrall notes. In clarifying the need for radiology practices to move beyond simply accepting overutilization as part of keeping referring physicians happy (or as part of maintaining current revenue levels), he says, “It’s not a sound business principle to try to build a practice or a business based on things people don’t need—on providing services that are unnecessary.” Growth Versus Overutilization It is important, Basu says, to avoid confusing growth with overutilization in imaging, as this mistake has done much to damage the reputation of radiology in the eyes of lawmakers, regulatory agencies, insurers, and the public. Growth in procedural volumes can often represent completely appropriate increases in the utilization of imaging services in response to technological advances and to improving medical evidence of the best applications for imaging modalities and procedures. Many of the major growth spurts that imaging volumes underwent in previous years were wrongly characterized by policymakers as the simple addition of yet another costly, high-tech, and possibly unnecessary study to the diagnostic protocol for a given disorder, with the economic effects of that addition then multiplied, across the nation, by the number of patients initially suspected of having that condition. Basu point out that this oversimplified and erroneous model has harmed radiology both by exaggerating the cost of imaging and by overlooking its value (in clinical and monetary terms). In many—or even most—cases, the imaging-utilization statistics could have been showing growth in the use of a procedure or modality for sound reasons. Often, one type of imaging is substituted for another as its superiority becomes more obvious to referring physicians. In such instances, figures demonstrating growth in that modality must be weighed against any declines seen in the utilization of other modalities, if any realistic conclusions are to be drawn. This is relatively obvious for analyses of the utilization of one modality versus another (but not often brought to broad attention, even so). This type of comparison has been far more difficult to make when one procedure largely replaces another within the same modality, partly due to lack of data and partly because this kind of work hasn’t often been undertaken. Radiologists in the field might see a strong trend toward one less-expensive but diagnostically equivalent study replacing a more costly one within a modality, for example, but this kind of positive change is unlikely to be tabulated anywhere—and still less likely to reach a level where it could gain the attention of policymakers. The value of imaging in replacing diagnostic procedures outside radiology is even more pronounced, but just as often overlooked. How often, today, does anyone undergo exploratory surgery? It is still necessary, under some circumstances, but it is certainly far more rare than it once was—because the same exploration can be conducted using imaging. Many of imaging’s steep gains in procedural volume happened because an imaging study could replace a diagnostic procedure that was less useful, but the most beneficial studies also replaced invasive procedures that were more likely to cause postprocedural complications, that were far more uncomfortable and unnerving for the patient, and that usually cost a great deal more than an imaging-based diagnostic progression. For example, an MRI study might be perceived as expensive by payors, but it certainly never approaches the cost of the exploratory surgery, pathology work, and associated expenses that it can (and so often does) replace. Any fair attempt by policymakers to assess the cost of growth in the utilization of imaging procedure must be certain to subtract the cost of all of the invasive diagnostic procedures that this additional imaging replaced (and the extra risk and pain that invasive procedures add to the patient’s burden should not go unnoticed, either). Utilization Realities For these and similar reasons, Basu says, the frequently repeated claim that as much as a third of imaging is inappropriate is probably untrue. He adds, however, that there is reliable evidence (both direct and indirect) that overutilization is more than a figment of payors’ imaginations. It is real, but it has many causes; several of the reasons for overutilization have their roots outside radiology (and even outside medicine), but there are also factors driving overutilization within the specialty. Overutilization is hard to quantify, but Basu explains that retrospective review of referring physicians’ imaging orders is one direct measure that can be used to verify the extent of the problem. Analysis of the indications for imaging in individual cases will make overutilization more obvious, for example—often, in the form of duplicated exams. Indirect clues to the presence of imaging overutilization are seen in the wide variance of procedural volumes for imaging, both from region to region and from one ordering-physician specialty to another. In some places, the average annual cost of imaging per Medicare enrollee is twice the national average, for example. Because the variance in imaging utilization among locales and specialties is large, Basu says, it is improbable that all of the lower-use areas and specialties represent underutilization (especially in a litigious society): The presence of overutilization is more likely to be the reason that the variance exists. A more localized indirect indicator of overutilization is the abrupt increase in the use of a modality that is sometimes seen when a physician has just acquired equipment in that modality for the first time. Radiologists’ Actions Radiologists face multiple external pressures that can drive overutilization, but other influences favoring it come from within many practices and departments. Some radiologists might feel that reviewing the appropriateness of a requested study is not their role, and might thus be unwilling to take this step. Others could be willing, but not experienced enough to feel confident that they are able to perform such a review. Unfortunately, some radiology practices also have a long tradition of self-referral that takes the form of recommending additional studies (which might not always be appropriate). Less questionable motives, such as adherence to a clinical protocol or to a policy or standard previously established by the practice, can lead to recommending more imaging than necessary. In some cases, radiologists’ recommendations and/or referring physicians’ orders can be inappropriate because the imaging guidelines of specialty societies need to be updated to reflect evidence from recent studies. In addition, Basu says, some radiologists remain unaware of the harmful long-term consequences that imaging overutilization could bring to patients, payors, providers, their practices, and themselves. Others are not yet convinced that these effects are bad enough to warrant changes in their day-to-day activities. These radiologists are unwilling to assume a gatekeeper’s role in determining whether an imaging order is appropriate, and they fail to see themselves as the expert consultants that such a role requires. The research that would be expected to help more radiologists assume the expert consultant’s position with confidence has not been supported by policymakers to anything approaching the necessary degree. Evidence of effectiveness is needed, and the effort to tie imaging to patient outcomes requires more attention than it now receives. Lawmakers, Basu says, also have failed to enact measures that would substantially reduce the overutilization attributable to financially motivated self-referral outside radiology. Patients’ Demands The public’s demand for imaging studies remains strong, while its understanding of what those studies can do generally remains poor. Consumers can feel cheated of the best possible health care if they are denied the most advanced imaging exams, even in cases where those exams are clearly inappropriate. Because some health plans impose inconsequential (or no) deductibles or copayments for outpatient imaging—under rules left over from the era when insurers were working to discourage unnecessary inpatient admissions for diagnostic testing—patients can be subject to financial detachment from the consequences of their imaging demands. In the view of patients who have not been helped by their physicians to understand whether the results of an imaging study would have any effect on their treatment, more is always better. When such patients demand additional exams, too many physicians give in and order these inappropriate studies—and too many radiologists give in and perform them. “Radiologists have the best clinical, operational, and economic understanding of when imaging is and is not useful. To serve their patients and physician colleagues better, it is imperative that radiologists play a lead role in optimizing appropriate imaging and curtailing inappropriate imaging. Specifically, I encourage radiologists to hold meetings or seminars with referring-physician colleagues, payors, and policymakers to find solutions to these issues,” Basu says. Taking Action Speaking after Basu, Thrall observes that even though imaging became the subject of fee-reduction measures due to growth—with actual overutilization being far less responsible than simple growth for attracting regulatory attention—imaging is, nonetheless, a handy target. The probability of further cuts in imaging reimbursement, therefore, is high. Thrall defines genuine overutilization as the performance of imaging procedures that are not expected to improve patient outcomes (in light of the individual circumstances that apply to a case). The existence of overutilization is unquestionable, Thrall says, but there are no empirical data to indicate how prevalent it really is. The figures often cited for excess imaging could be too high, but might also be too low, and overutilization rates for different care settings (and among radiology practices) also vary a great deal. Thrall reports that there appear to be three factors that are the main contributors to imaging overutilization. These are self-referral, defensive medicine (including acquiescence to patients’ inappropriate preferences), and referring physicians’ poor understanding of appropriate imaging. This lack of knowledge is probably a factor in 10% to 15% of imaging referrals, Thrall says, with self-referral accounting for 8% to 12% of problem orders. In the recent past, high levels of reimbursement for imaging studies tempted too many physicians into self-referral for imaging; they might not otherwise have bothered to offer imaging services. In particular, Thrall says, the Stark law—which was intended to combat self-referral—has, instead, encouraged and protected it through the inclusion of the in-office ancillary-services exemption. This ineffective measure might have been replaced by more useful deterrents through the enactment of health-care reform, but that opportunity was lost during the creation of the new law. The influence of defensive medicine on overutilization is the most variable of the three main drivers of inappropriate imaging, Thrall notes, partly because the states that have enacted tort-reform laws are less likely to see patients being encouraged by malpractice attorneys to sue their physicians. The percentage of imaging orders issued for fear of medical liability could, therefore, be as low as 5% in less litigation-prone areas, but as high as 25% in others. These three overutilization-promoting problems often overlap in individual cases, so their probable percentages cannot simply be added together to yield an estimate of total overutilization. All three factors could even be involved in persuading a physician to order one inappropriate imaging exam. Educating Patients While Thrall acknowledges the difficulty that researchers have encountered in attempting to quantify the multiplying effect on imaging overutilization that patient requests can have, he also says that this is one area where individual referring physicians can have a considerable influence on improving the appropriateness of imaging orders. Naturally, educating themselves about the best uses of today’s imaging studies is the first step that both primary-care physicians and specialists should take to reduce their contributions to overutilization. That should not be where their increased level of knowledge ends, though. By educating their patients, referring physicians can help them understand what they should expect from imaging; that, in turn, should reduce the demand for unwarranted procedures. For example, Thrall says, prior medical encounters have led many patients to expect imaging exams to be ordered for them whenever they report a new health problem to their physicians. Often, these patients then respond to the physician’s decision not to order any imaging studies by contending that their health-care needs are not being adequately met. An investment of the physician’s time in letting such a patient know why imaging is not called for, in his or her current case, will not only reduce inappropriate utilization during that encounter, but will improve that patient’s satisfaction with the care provided (and could reduce the likelihood that he or she will demand unnecessary imaging studies during future medical visits). Thrall adds that there is an overlapping effect between patients’ demands for imaging exams and physicians’ liability concerns: A patient whose demand for a medical service has not been heeded can become a patient whose poor (but unrelated) subsequent outcome then leads to a malpractice suit alleging that the necessary medical care was not provided. For many physicians, this prospect can be threatening enough to prompt the ordering of imaging procedures that the physician knows are unlikely to be helpful. Decision Support There is, Thrall says, a fourth reason for the overutilization of imaging studies, but unlike the other three factors, it is not a motive: It is the absence of the kinds of decision-support or utilization-management systems that can reduce overutilization due to any of its other causes. This lack of access to useful systems means that there are still many unexploited, promising opportunities to address overutilization. An ideal utilization-management system for imaging would have eight primary characteristics, Thrall says. First, it would be effective in helping to control the inappropriate utilization of imaging. Second, it would be transparent, working by applying criteria that are made known to referrers and to imaging providers before they try to issue or carry out an order for imaging. Third, the ideal system would yield answers that are reproducible. Whenever identical exams are ordered for any two patients who share the same characteristics, medical history, past studies, and suspected clinical problems, the utilization-management system should produce the same results. Fourth, it would be efficient. This means that it creates no added costs that referrers or imaging providers are expected to bear; it also means, Thrall says, that its use can be adopted without disrupting workflow in the radiology department, imaging center, or referrer’s office. Fifth, it would enhance safety for patients, alerting users to patients’ allergies or kidney problems and indicating that a proposed exam duplicates one that has already been performed. As a result, the risks of reactions to contrast media and of avoidable additional radiation would be reduced. Sixth, it would be flexible, Thrall says, giving users a choice of alternative procedures to consider and accommodating unusual circumstances described by the ordering physician. Seventh, it would be educational, allowing users to learn more about appropriate imaging each time they ordered a study (and, therefore, helping them improve their ordering patterns over time). Eighth, it would be developmental. In a field that is (when all goes well) constantly evolving new technologies and new applications for established technologies, a utilization-management system must always reflect the state of the medical evidence for the benefits of imaging. Imaging providers and their organizations should be able to adjust the system, so that it can adapt as radiology’s knowledge grows.
Table. Comparison of Two Methods Used to Manage Imaging UtilizationA side-by-side comparison of the characteristics of computerized provider order entry (CPOE) augmented by decision support with the characteristics of the preauthorization-based approach taken by radiology benefit managers (RBMs) makes it clear that decision support is far more likely to be able to reach all eight of these ideals than an RBM model, based on its typical business structure and methods, ever could (see table). Experience at MGH In mid-2003, MGH implemented its radiology order entry (ROE) system, following preliminary research and pilot programs undertaken in the preceding years. Some of that groundwork had been laid through observation of the effects that displaying information about quality and utilization during the ordering process had on the imaging-ordering patterns of primary-care physicians. Other background work that informed the ROE project (some of it performed using the very large databases of major payors) included analysis of variation in radiologists’ recommendations for additional imaging for the same clinical indication, in addition to evaluation of the reasons for unnecessary repetition of the same (or similar) exams for a patient. MGH’s ROE is a utilization-management system that uses CPOE to provide decision support at the point of care. It uses a modified version of the ACR® appropriateness criteria (which cover more than 160 topics and 800 variants). The ordering physician uses a simple, quickly learned interface to indicate the patient’s situation, primarily by clicking the relevant checkboxes. Pull-down menus are used to choose the imaging modality to be used and the body part to be examined. ROE makes it easy for users to alert the technologist and radiologist to any special consideration that might apply to the patient (such as the need to avoid contrast media or the presence of allergies, pregnancy, or implanted medical devices); this step serves as an extra safeguard. On every ROE screen, there are hyperlinks to further information on relevant topics, allowing ordering physicians to find out more, from authoritative sources, without leaving the CPOE system.
Figure 1. After all patient information has been entered and a procedure has been ordered, the user of the decision-support system is provided with a results screen that assigns a score to the probable diagnostic value of the ordered exam, as well as to possible alternatives. In this case, the MRI study that was ordered is scored as likely to be clinically useful, and it is also considered more likely to help than CT or radiography exams would be.Boxes are also checked to indicate the specific requested exam, as well as the signs and symptoms that are the reason for the physician’s imaging order. The decision-support system then assigns a number that predicts the utility of the ordered exam, on a scale of 1 to 9; a traffic-light color is also assigned (Figure 1). Yellow, for example, is used to indicate an exam of intermediate probable utility, with a score of 4 to 6.
Figure 2. Trends in appropriateness of MRI orders following implementation of computerized provider order entry with decision support (based on modified ACR® appropriateness criteria).
Figure 3.The impact of decision support on actual MRI, CT, and nuclear-cardiology 2006 procedures performed per 1,000 health-plan enrollees. The system was fully implemented in the second quarter of 2005.There is no prohibitive element involved in using ROE. Even if the system indicates that an exam is of low predicted utility, the physician always has the options of proceeding with the order, modifying it, or canceling it entirely. This makes the impact of ROE on overutilization (Figures 2 and 3) even more remarkable, in comparison with RBM methods, since those often rely on denying coverage for exams to produce the decreases in utilization with which they have been credited. Protecting Good Growth Imaging’s growth is decelerating, despite ongoing demand for imaging services and the aging of the population (which increases the need for medical services of all types). The slowing of growth, therefore, could be the response that policymakers predicted—and sought—when they imposed drastic payment decreases for imaging services. Thrall says that multiple factors underlie growth in imaging. Undoubtedly, there was (and still is) bad growth, defined as profit-seeking activity on the part of imaging providers who do not care whether the services that they provide confer any medical benefit. Good growth, in contrast, is a response by physicians to the growing clinical utility of imaging in the diagnosis and treatment of a constantly expanding list of disorders. Because growth and overutilization have been mistaken for each other, that good growth has been placed at risk, both through direct payment cuts and through the implementation of burdensome preauthorization requirements. In taking aim at self-referring physicians—using decreases in per-exam reimbursement—payors have, Thrall says, wounded the ethical providers of imaging services as well. By adopting sound decision-support systems, he hopes, radiology will be able to make the difference between growth and utilization clear. That distinction would then help the specialty protect its beneficial utilization growth from getting caught in the crossfire of the battle against inappropriate imaging. Kris Kyes is technical editor of Radiology Business Journal.
Table. Comparison of Two Methods Used to Manage Imaging Utilization
Figure 1. After all patient information has been entered and a procedure has been ordered, the user of the decision-support system is provided with a results screen that assigns a score to the probable diagnostic value of the ordered exam, as well as to possible alternatives. In this case, the MRI study that was ordered is scored as likely to be clinically useful, and it is also considered more likely to help than CT or radiography exams would be.
Figure 2. Trends in appropriateness of MRI orders following implementation of computerized provider order entry with decision support (based on modified ACR® appropriateness criteria).
Figure 3.The impact of decision support on actual MRI, CT, and nuclear-cardiology 2006 procedures performed per 1,000 health-plan enrollees. The system was fully implemented in the second quarter of 2005.