Introduction and summary
For more than two decades, experts have warned that Medicare relies far too heavily on recommendations submitted by the American Medical Association’s Relative Value Scale Update Committee (RUC) when setting payment amounts for different physicians’ services.1 Moreover, because Medicare’s relative values are also used by Medicaid and private insurers, the distortions in payment caused by the existing RUC process ripple through the entire health care system.2
The American Medical Association (AMA) resisted changes to the RUC well before the Trump administration, but the current administration has sent a clear message that the AMA need not worry about any pushback or additional oversight.3 Instead, during last year’s physician fee schedule rule-making process, Medicare signaled that it plans to defer to the RUC even more than in past years.4 Since then, the powerful AMA lobby has been successful in keeping even the most limited reforms from becoming law.5
It is clear that the RUC process is broken and that the AMA has no interest in even minor changes. And given the Trump administration’s extreme deference to the RUC, and the unlikelihood that there is sufficient congressional support to overcome that position, reform will have to wait. But the importance of reform is growing, and it is essential that policymakers and stakeholders outline what changes must be made to Medicare’s current system.
First, improving the accuracy of Medicare payment rates is critical to ensuring appropriate payments as the program transitions from a fee-for-service system toward one that links payment to quality and outcomes. All of these value-based payment reforms are built on top of the existing fee-for-service architecture.6
Second, proposals to guarantee universal health care coverage, including the Center for American Progress’ Medicare Extra for All, must include appropriate payments for doctors in order to finance affordable universal coverage.7 Medicare’s current payment constraints are insufficient to support an effective financing system; it is critical to undo Medicare’s current bias in favor of procedure-based, specialty care and provide additional resources for critical primary care, mental health, and other services that Medicare currently underfunds.
In this report, CAP outlines a new approach to setting physician payment amounts based on empirical data that also reflects today’s understanding of the importance of primary care and other nonprocedural services.
The current system for valuing physician services
The Relative Value Scale Update Committee’s design and function create myriad interconnected process and conflict-of-interest problems. There are also substantive problems with how the RUC values nonprocedural work by physicians. The Government Accountability Office (GAO) has concluded that “weaknesses in the RUC’s relative value recommendation process … present challenges for ensuring accurate Medicare payment rates for physicians’ services.”8 Moreover, the RUC “is made up of practitioners who have a financial stake in” Medicare’s final payment decisions; it operates largely out of view; and its members’ subjective assessments comparing the relative ease or difficulty of various tasks are confidential.9
How the RUC makes value recommendations
About 20 percent of the nation’s health expenditures are for physicians’ services.10 Payments for physicians’ services under traditional Medicare total about $90 billion per year.11 Medicare pays a separate amount for each service that corresponds to one of more than 7,000 codes under the physician fee schedule.12 Each payment amount is adjusted to reflect the relative resources needed to perform that particular service compared with other physician services. Federal law requires the Centers for Medicare and Medicaid Services (CMS) to review all relative values at least every five years and, on an ongoing basis, identify services that are likely to be misvalued.13
The relative value of a service includes three components: 1) the value of physicians’ work, which accounts for just more than 50 percent of the total value; 2) the practice expenses, which are about 45 percent of the total; and 3) the cost of professional liability insurance, which is the small remaining amount.14 The physician work component takes into account both the time that the physician needs to provide the service and the intensity—including the cognitive effort and judgment, technical skill, and psychological stress—associated with performing the service.15 While there is a chance for error in both of these assessments, it is especially challenging to estimate the intensity of services.16
Medicare’s periodic adjustments to the relative values of physician services are intended to account for changes in medical practice, coding, and new information about each of the three components. Any changes must generally be budget-neutral. This ongoing process begins by both CMS and the RUC identifying potentially misvalued services for review.17 In addition, another AMA panel—the Current Procedural Terminology (CPT) Editorial Panel—identifies new and revised codes for RUC review.18
After the RUC has a list of services to review, it consults with members of the physician specialty societies that perform the services; once these groups decide to proceed with re-evaluating a code, they develop surveys of society members to assess the time and intensity of a specific service, as well as a recommendation for the total work relative value.19 These surveys also include questions that compare the service relative to other selected services.
The relevant specialty society then reviews the survey results and prepares recommendations for the RUC about the relative value of the specific service. Before review by the entire RUC, one of its members will pre-review the recommendation and provide feedback to the specialty society.20
The full RUC then considers the recommendation at one of its thrice-yearly meetings; the specialty societies present their work relative value recommendations to the entire panel.21 At that time, RUC members discuss and then vote on the recommendation, and the full 31 member RUC can either approve it or refer it back for further work. CMS officials attend these meetings, and while they may comment or ask questions during the deliberations, they generally do not comment on the merits of a recommended value.22
The RUC then forwards its recommendations to CMS. CMS then reviews the recommendations by looking at the survey results and other supporting data and comparing the recommendations with the values of other physician services.23 The agency publishes its preliminary relative value decisions as part of its annual proposed physician fee schedule rule, and it later finalizes the values through this rule-making cycle.24
The GAO has noted that CMS “does not have its own data sources to validate RUC recommendations because such data sources do not exist,” and “[t]he RUC is currently the only source of comprehensive information available regarding the physician work.”25 For this reason, it’s not surprising that in most cases, CMS accepts the RUC’s recommended work relative values.26
The RUC process cannot be fixed
The core features of the RUC—its evaluation process and specialty society membership—are also the source of its harms. Columbia University professor Miriam Laugesen’s Fixing Medical Prices offers the most comprehensive review of the evidentiary problems and conflicts of interest inherent in the RUC, detailing the methodological problems with the evidence collected by surveys, as well as the ingrained conflicts of interest present in this work.27
The RUC’s heavy reliance on surveys developed by specialty societies is just one example. As Laugesen notes, “the method of surveys of physicians is not sufficiently strong given the use to which this information is put.”28 The sample sizes of the survey are extraordinarily small, the surveys have extremely low response rates, and the results can have wide ranges.29 Together, according to the GAO, these “suggest shortcomings with the data” and raise significant concerns that the results may not accurately account for the time and effort of most doctors performing the service.30
Given these shortcomings, it is not surprising that when researchers have compared the survey results with operating room logs, they have found that the time estimates in the surveys are much higher than those recorded in the logs.31 One study “found that RUC survey data overestimated the procedural time by an average of 31 minutes across 60 procedures when compared with objective data from operative logs.”32
Further evidence that these subjective surveys are unreliable comes from reviews of specific work relative values “for many high-volume, high-cost services [that] haven’t been adjusted downward to account for automation, experience, personnel substitution, and other productivity improvements that have substantially reduced the amount of physician time and work involved.”33 The largest discrepancies occur with imaging and test interpretations, outpatient department and ambulatory surgery center procedures, and office-based procedures.34
But there is a more fundamental problem with relying on a doctor to assess her work when she knows that her answers can increase her income. And when physicians fill out these surveys, they are also subject to unconscious biases, such as anchoring effects and recall bias.35 The former describes the psychological finding that people are “highly suggestible to or influenced by baseline numbers that anchor their responses.”36 The latter refers to the fact that recollections of events and their duration can be inaccurate—most likely overestimating duration.37
The RUC has pushed back against this criticism, citing its subsequent review of the survey results as an adequate check on reliability. The committee relies on magnitude estimation to assess if the survey results are consistent with relative values of related services, and it often recommends a work value that is lower than what is suggested by the survey data.38 Yet even with these qualitative modifications, the resulting work values are still likely to be inaccurate because the underlying survey data are fundamentally flawed.
But the RUC’s flaws are not limited to the use of biased surveys or other methodological problems. The very structure of the RUC creates a conflict of interest; membership of the RUC includes physicians who serve Medicare beneficiaries and whose incomes vary based on the relative value of the services they provide.39 More than two-thirds are appointed by national medical specialty societies that advocate on behalf of their members, and consultants hired by these societies also attend the RUC meetings.40 The RUC has taken steps to minimize conflicts of interest: RUC members may not participate in the review of or vote on proposed relative values for services with which they have or a family member has a direct financial relationship. And more generally, the RUC tries to present itself as an expert panel, not an interested group of advocates, in order to push back against claims of self-dealing.
Yet conflicts of interest remain. The RUC remains a small, insular group with “interlocking and deeply embedded relationships between specialty societies and committee members.”41 Furthermore, Laugesen’s interviews with RUC members revealed “deals and coalitions that contradict the [claim that the RUC is] an expert panel” and not a group of advocates.42 More fundamentally, even if RUC members never consciously worked together to influence various votes, conflicts of interest remain as long as there is even an appearance that a person may be at risk of acting in a biased way because of personal interests.43
Voting is also secret, making it difficult to “determine whether the RUC’s recommendations are biased in favor of certain specialty societies.”44 Given the RUC’s role in allocating hundreds of billions of taxpayer dollars, it should be easy for the public to attend its meetings, and the proceedings should be transparent. Their meetings are ostensibly public, but their location and attendance requirements belie this claim. For example, at the RUC’s January 2016 meeting in Miami, which an author of this report attended, organizers claimed there was limited room available for the public, even though the meeting was held in an enormous ballroom with plenty of empty spaces. All attendees were also required to sign sweeping confidentiality agreements.
The RUC undervalues important provider services
Today, Medicare continues to pay significantly less for cognitive services than for procedural services, which reflects the RUC’s tendency to value codes primarily on the basis of the physical, mechanical skill involved in an interaction with a patient. Using such a narrow lens to evaluate the work involved in physician services, the RUC systematically undervalues cognitive services that include “critical thinking involved in data gathering and analysis, planning, management, decision making, and exercising judgment in ambiguous or uncertain situations.”45
While many procedural services may have been more complex and intense than cognitive services in the past, there is growing evidence that this dichotomy is outdated and increasingly inaccurate. As the Medicare Payment Advisory Commission (MedPAC) has noted, technological innovations and the increasing role of midlevel practitioners may mean that the relative values of certain procedures are overstated.46 At the same time, many cognitive services are increasingly difficult and labor-intensive, for both primary care physicians and specialists whose practices are not procedure-intensive.47 For example, care coordination for a high-risk patient can involve far more judgement, analysis, and decision-making than a physician’s preparation for a standard in-office procedure such as taping ascites or a spinal tap. And a meeting between an oncologist and her patient to discuss end-of-life wishes can require as much skill, expertise, and judgment as an endotracheal intubation in an emergency room.
More fundamentally, the RUC’s ongoing bias in favor of procedural services is at odds with the critical role that primary care plays in the health system; the nation’s health is directly linked to the strength of its primary care delivery system and workforce. As Harvard professor Dr. Asaf Bitton has explained, primary care is not only the front line of the health care system, but it also provides coordination, continuity of care, and other patient-focused care that is complex and extraordinarily valuable.48 Properly valuing primary care is, in fact, “about aligning a payment system toward our end outcomes and goals.”49
Rewarding procedural services to the detriment of cognitive services creates problematic financial incentives throughout the health care system. First, it creates a disincentive to spend time on cognitive care or select specialties whose focus is on these activities.50 One analysis concluded “that Medicare reimburses physicians 3 to 5 times more for common procedural care than for cognitive care.”51 In that study, the authors demonstrated “that 2 common specialty procedures [cataract extraction and screening colonoscopy] can generate more revenue in 1 to 2 hours of total time than a primary care physician receives for an entire day’s work.”52
Second, this practice also creates incentives for physicians to structure their practices to maximize other sources of income. For example, Medicare payments for physician-administered drugs increase revenue for medical oncologists whose cognitive services are undervalued. But these types of nonservice-related payments can, in turn, raise costs for the entire Medicare program by creating their own problematic incentives.
These incentives are at odds with the actual needs of the U.S. health system—and will diverge even more as the nation’s elderly population continues to grow. One study estimates that the number of primary care doctors will need to increase by at least one-third in order to meet the needs of patients with multiple chronic conditions.53
Every year, an estimated 26 percent of all Medicare beneficiaries experience some type of mental health disorder.54 And for those with severe mental illness or substance use disorders, Medicare spends five times more than on similar beneficiaries without these diseases.55 These beneficiaries need coordinated physical and mental health services to improve their health outcomes. But the current Medicare payment rules do not properly value these services, leading to fewer mental health professionals in the Medicare program. One study found that just 55 percent of psychiatrists accepted Medicare reimbursement, compared with 86 percent of physicians in other specialties.56
A new framework for evaluating codes
Policymakers must replace the Relative Value Scale Update Committee’s one-size-fits-all approach to assigning value to physician services. First, payment for nonprocedural services such as primary care and mental health services should not be set based on an arbitrary comparison to procedure-based services. Second, payment for procedural services should reflect empirical time data.
Set payment amounts to adequately reflect the value of cognitive services
The current payment system creates winners and losers each time codes are reassessed because all changes must be budget-neutral. Consistent with CAP’s Medicare Extra for All proposal, the first step to correcting the bias toward procedural codes is to increase average rates for primary care and other cognitive services by 20 percent relative to certain rates for specialty care on a budget-neutral basis. When deciding which nonprimary care services to apply this increase to, policymakers should prioritize services with existing provider shortages, such as services for mental health and substance use disorders.
This adjustment would correct Medicare’s substantial bias in favor of specialty care at the expense of primary care. After this initial adjustment, Medicare should periodically update this group of services on a budget-neutral basis, as well as engage with physicians, patients, and other health care practitioners to evaluate the adequacy of the current payment amounts and the impact of these changes on the number of physicians treating Medicare patients. Medicare should also consult with public health experts to assess if additional changes might improve access to care and community wellness.
Former CMS Administrator Gail Wilensky has noted that “the persistent focus on inputs and the costs of a particular input, as opposed to what you get from that input, put you in an undesirable position in terms of trying to acknowledge that there frequently are, can, and should be different ways of getting to a health outcome.”57 By reframing the process to place a greater focus on the value of the services provided to patients instead of the physicians’ time and effort, Medicare can encourage the use of services that improve patient outcomes and keep patients healthy.
Use empirical data to measure the time spent on services
The transition to using empirical data in place of surveys to evaluate physician services is a substantial undertaking. In order to be ready to implement these changes under a future administration, researchers and policy experts should continue to assess available empirical data, such as existing time-motion studies for specific physician activities and the expansion of current data collection and analysis efforts.58
In 2011, the Medicare Payment Advisory Commission sponsored a study to assess the feasibility of using empirical data to evaluate codes. During the Obama administration, the Centers for Medicare and Medicaid Services also announced a pilot project to address potentially misvalued services. The project developed empirical measures of physician service times and considered the implications of these new measures for physician work values. As part of this effort, researchers from the Urban Institute developed “empirical time estimates based on data from several health systems with multispecialty group practices.”59 They collected two types of data for 60 different services: administrative data from electronic health records and direct observation data from staff. Researchers also solicited clinical experts to review the data.60
These researchers encountered a number of challenges throughout the pilot, including difficulties in recruiting health systems to participate and difficulties in engaging staff. Particular challenges included working within union rules, obtaining institutional review board approvals, receiving patient and physician consent, scheduling complications, and training clinical staff. Nevertheless, the researchers believe that a larger study to collect empirical time data is “feasible,” and they offer a number of recommendations for future data collection efforts, including specific steps to encourage health system participation.61
One key is ensuring adequate resources for participating systems. According to the Urban Institute report, “Many of the practices that declined expressed interest in participating in a study like this but faced logistical or organizational barriers.”62 Funding these projects will be more difficult without the federal government’s participation. For example, if the Trump administration were willing, the Center for Medicare and Medicaid Innovation could provide grants in support of this work.
In the absence of federal support for this work, researchers will need to make the case to health care funders and other payers that these efforts are critical to building a higher-quality, lower-cost system that more accurately pays for physician services.
Any such project should also establish recommendations for evaluating the intensity of physician work. Empirical data can replace surveys in determining the time that a physician needs to provide a service, but there remains some need for qualitative, expert assessments about the resources and intensity required for specific services. Moreover, there may be a need to further modify payment in particular situations. If, for example, there are multiple procedures to treat the same condition with very different time and intensity findings, a shorter, less intensive approach may merit higher payment if it is safer and more effective.
Instead of relying on the RUC’s small group of interested physicians who review anecdotal surveys, policy experts should assess other ways in which experienced physicians could participate in these evaluations. For example, Urban Institute researchers engaged physician experts to validate their empirical findings. These review panels should include physicians, patients, and other health care practitioners, a majority of which should be noninterested physicians whose salaries and other income are wholly unrelated to Medicare payment levels. These individuals should review empirical data, consider updates to nonprocedural codes as outlined above, and recommend additional changes to payment amounts. Finally, Medicare will continue to propose and finalize updates to various codes in annual rule-making, while giving the American Medical Association and physician specialty societies the opportunity to comment on any proposed changes.
Conclusion
Rethinking how the health care system determines payment amounts for thousands of codes may appear to be a technical debate among a small number of policy experts. But it has far broader consequences because, as Laugesen noted, “what we pay for medical services as well as how we pay for them speaks to what we value: what kind of medical care we want, how we want to allocate resources, and how to fairly compensate physicians doing very different kinds of work.”63
A transition to a more thoughtful approach grounded in the needs of patients that is also data-driven and transparent will not be simple. Yet investing in these efforts today will allow for a successful move to value-based payment in the short term and to a health care system that meets the growing demands for primary care, mental health services, and care coordination in the future.
About the authors
Maura Calsyn is the managing director of Health Policy at the Center for American Progress.
Madeline Twomey is the special assistant for Health Policy at the Center.
This publication was made possible in part by a grant from the Peter G. Peterson Foundation. The statements made and the view expressed are solely the responsibility of the Center for American Progress.