C h a n g e s i n H e a l t h Ca r e F i n a n c i n g & O r g a n i z a t i o n Medicare’s Value-Based, Physician Payment Modifier: Improving the Quality and Efficiency of Medical Care July 2012 Changes in Health Care Financing and Organization is a national program of the Robert Wood Johnson Foundation administered by AcademyHealth. Medicare’s Value-Based, Physician Payment Modifier: Improving the Quality and Efficiency of Medical Care page 2 Table of Contents I. Issues Related to the Functioning of Medicare’s Value-Based Modifier for Physician Payment (VBPM) . . . . . . . . . . . . . . . . . . . . . . 4 A. Issues Related to the Functioning of Medicare’s VBPM in Need of Clarification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 B. Aligning Measures Used for Medicare’s VBPM with Other Public and Private Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 C. Gaining the Support of Physicians and Physician Groups for the Measures Used for Medicare’s VBPM . . . . . . . . . . . . . . . . . . . . . . . . . 7 II. Issues Related to the Selection of Quality and Cost Measures for Medicare’s Value-Based Modifier for Physician Payment . . . . . 8 A. Selecting Quality Measures for Medicare’s VBPM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 B. Selecting Cost Measures for Medicare’s VBPM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 III. Methodological Issues Related to Medicare’s Value-Based Modifier for Physician Payment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 A. Composite Measures of Quality and Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 B. Risk Adjustment of Clinical Outcomes and Health Care Costs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 C. Performance Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 IV. Phasing in Medicare’s Value-Based Modifier for Physician Payment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 V. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 VI. Endnotes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Medicare’s Value-Based, Physician Payment Modifier: Improving the Quality and Efficiency of Medical Care page 3 Medicare’s Value-Based Physician CMS’s development and implementation of Genesis of this report. Payment Modifier: Improving Medicare’s VBPM will be an iterative pro- The Affordable Care Act of the Quality and Efficiency of cess. In 2013, the ACA requires HHS to 2010 mandated the federal Medical Care publish in the Federal Register the measures government’s development of a The Affordable Care Act of 2010 (Public of resource use and quality and the analytic mechanism to allow Medicare Law 111-148 and Public Law 111-152) measures that CMS will use to determine to make differential payments (ACA) is a once-in-a-generation law Medicare’s payment modifier. Beginning to fee-for-service physicians intended to transform the U.S. health care on January 1, 2015, the law requires CMS based on the relative quality system by expanding health insurance cov- to apply Medicare’s VBPM to payments and costs of care they provide. erage, making care more patient-centered, of selected fee-for-service physicians and The Centers for Medicare and promoting the adoption and use of elec- physician groups. In 2017, CMS is directed Medicaid Services must phase tronic health records, and offering financial to apply the VBPM to payments to all (or in Medicare’s budget-neutral, and other incentives to health care provid- nearly all) physicians paid under Medicare’s value-based physician payment ers to improve the quality and value of the fee schedule. modifier between January 1, health care they deliver.1,2 2015, and January 1, 2017. “It is well established now that one The 2010 law charged the U.S. Department can in fact improve the quality of To discuss the develop- health care and reduce the costs at of Health and Human Services (HHS) ment and implementation with undertaking several bold initiatives the same time.” of Medicare’s new physi- aimed at recasting Medicare from a pas- cian payment modifier, Health Affairs Editor-in-Chief Susan sive payer to an active purchaser of higher the Robert Wood Johnson Dentzer quality, more efficient health care. Given Foundation, under its Changes that Medicare covers 47 million elderly and in Health Care Financing and CMS faces several challenges in the design disabled beneficiaries, engaging in value- Organization (HCFO) initia- and implementation of Medicare’s VBPM. based purchasing has enormous potential tive, hosted a meeting in One is likely to be a lack of agreement not only for ensuring that Medicare pro- Washington, D.C., on March among key stakeholders on priorities, mea- vides beneficiaries with high-quality care 29, 2012. Meredith Rosenthal, surement, and other design elements. CMS and remains solvent but also for catalyzing Ph.D., of the Harvard School of is collaborating with stakeholders inside changes in the entire health care system. Public Health moderated the and outside government, reaching out to meeting. physician groups and specialty societies, One of the value-based purchasing initiatives authorized by the ACA pertains to services holding public listening sessions, using the Economists, researchers, ana- Medicare Physician Fee Schedule rule-mak- delivered by physicians receiving fee-for- lysts, and federal policymak- ing process to develop equitable perfor- service payments from Medicare. Section ers as well as representatives mance measures, and relying on meaning- 3007 of the law mandates the Secretary of of physician groups and the ful and actionable feedback reports.3 HSS to develop a mechanism based “upon insurance industry engaged in the quality of care furnished as compared to a moderated discussion of the On March 27, 2012, the Robert Wood cost,” that provides for differential payment functioning and implementa- Johnson Foundation’s Changes in to physicians and physician groups receiving tion of the modifier and associ- Health Care Financing and Organization compensation under the Medicare Physician ated methodological issues. (HCFO) program convened a meeting in Fee Schedule. The ACA requires the applica- This report highlights key Washington, D.C. for a moderated discus- tion of Medicare’s value-based payment modi- points of the discussion and is sion on the development and implementa- fier (VBPM) to be “budget neutral” when intended to faithfully capture tion of Medicare’s VBPM. Participants applied to fee-for-service physicians and, the essence of the discussion included representatives of physician as appropriate, to promote “systems-based without endorsing any one posi- groups, representatives of large commercial care.” Moreover, the Centers for Medicare tion. Given that the discussion health insurers, health care researchers, and Medicaid Services (CMS) must take into was “off the record,” comments health economists, and federal policymak- account the special circumstances of physi- are not attributed to specific ers. The meeting topics included (1) the cians or groups of physicians in rural areas individuals. functioning of Medicare’s VBPM in the and other underserved communities. Medicare’s Value-Based, Physician Payment Modifier: Improving the Quality and Efficiency of Medical Care page 4 U.S. health care system; (2) the selection VBPM with measures used by other HHS to use standardized extracts of Medicare of costs and quality measures for the value and CMS programs as well as measures used Parts A, B, and D claims data to evaluate modifier; (3) methodological issues related in quality improvement and pay-for-perfor- the performance of providers of services to the value modifier, and (4) factors to mance initiatives undertaken by other public and suppliers.4 If providing support for consider in phasing in the modifier in 2015 and private entities. quality improvement does not come and beyond. from CMS, would it be possible to iden- I. Issues Related to the Functioning tify some type of support either to foster At the outset of the meeting, the modera- of Medicare’s Value-Based Modifier learning collaboratives at the regional tor Meredith Rosenthal, Ph.D., from the for Physician Payment level or to engage other qualified entities Harvard School of Public Health, empha- in using a hybrid of Medicare data and A. Issues Related to the Functioning sized that the discussion was to focus on data from private payers? of Medicare’s VBPM in Need of pragmatic issues related to the design and Clarification implementation of Medicare’s payment At the March 2012 meeting, physicians and • Budget neutrality. How does CMS modifier for fee-for-service physicians. other stakeholders expressed considerable interpret “budget neutral” with respect Meeting ground rules specified that the dis- uncertainty about the ultimate goals of to the effect of Medicare’s VBPM? Is cussion would not address stakeholder agen- Medicare’s VBPM and how CMS will inter- the modifier supposed to be budget- das and that problems were not to be raised pret the ACA’s provisions related to the neutral with respect to Medicare Part without proposing solutions. Moreover, in VBPM. Such uncertainty, they suggested, B (Medical Insurance) expenditures the interest of time, the discussion was lim- underscores the need for CMS to make only, or is it supposed to be budget- ited to issues related to payment rather than clarifications related to the following: neutral across Medicare Parts A to public reporting of the information that (Hospital Insurance) and B? What about will be used in Medicare’s VBPM. expenditures under Medicare Part D • Goals. Do HHS and CMS intend to (Prescription Drug Insurance)? implement the VBPM for purposes of This report highlights some of the observa- both improving health care quality and tions, concerns, and suggestions expressed reducing costs, or is their primary goal • Systems-based care. How does CMS by participants at the meeting with respect to save money for the Medicare pro- interpret “systems-based care,” in the to the design and implementation of gram? Is one goal to have Medicare’s context of the ACA’s mandate for Medicare’s VBPM for physicians. Noting VBPM serve as a core for the evolution Medicare’s VBPM to be applied, as that the ACA requires CMS to phase in of performance-based payment across appropriate, in the service of promoting Medicare’s VBPM for some physicians the country and to contribute to trans- such care? beginning in 2015 and for virtually all phy- formational change in U.S. health care? sicians in 2017, attendees remarked that Will Medicare’s VBPM for physicians • Physician group. The ACA stipulates time is growing short and that near-term align with the goals of the HHS National that Medicare’s VBPM is to be applied actions will lay the foundation for an itera- Strategy for Quality Improvement devel- to all physicians and physician groups tive process of improvement to the VBPM oped under Section 3011 of the ACA? serving Medicare beneficiaries on a fee- in the coming years. Will Medicare’s VBPM be applied in a for-service basis. Currently, CMS pro- way that reduces disparities in health and vides feedback data to physician groups Meeting participants emphasized that clear, health care or at least does not exacer- under Medicare’s Physician Feedback actionable, and timely guidance from CMS bate them? Program, but Medicare does not con- is essential for all physicians who strive to tract with physician groups. Will CMS be comply with the application of Medicare’s able to use Medicare’s VBPM program • Support for quality improvement VBPM. The challenge in effectively reach- among physicians. Do HHS and CMS to measure and pay physician groups ing a national audience of physicians and intend to rely on Medicare’s VBPM rather than individual physicians? If so, clearly explaining the components of program to measure and compensate what criteria will CMS use to determine Medicare’s VBPM will require a variety of physicians on the basis of quality and what constitutes a physician group and communication approaches. efficiency or, beyond that, to provide what the group—as opposed to the indi- support for physicians’ quality improve- vidual physician—is accountable for? Moreover, meeting participants underscored ment initiatives and the transformation the importance of ongoing efforts at the of the U.S. health care system? Section federal level aimed at aligning HHS and 10322 of the ACA requires the Secretary CMS health quality and efficiency measures. of HHS to establish a process that per- They recommended aligning Medicare’s mits qualified public and private entities Medicare’s Value-Based, Physician Payment Modifier: Improving the Quality and Efficiency of Medical Care page 5 B. Aligning Measures Used for cial health insurers, regional initiatives to tiative under Medicare that allows Medicare’s VBPM with Other Public improve the quality and value of health physicians to earn a bonus on their and Private Measures care, health care quality organizations, and total Medicare Part B charges paid Several public and private health care health care accrediting/certifying entities). under the physician fee schedule programs use health care quality and cost if they report to CMS on a mini- measures. Representatives of physician • Federal programs. Examples of federal mum of three quality measures.8 organizations and others at the March 2012 programs whose measures used in quality For 2012, physicians and other meeting underscored the need for ongoing improvement and pay-for-performance caregivers in hospitals or physician efforts to align the VBPM’s health quality initiatives might lend themselves to align- practices may report 208 quality and efficiency measures with measures used ment include several Medicare programs measures and 22 measures groups by other public and private payers. authorized under the ACA, the Health in the PQRS.9 By reporting on a Information Technology for Economic minimum of three measures on a The standardization and alignment of per- and Clinical Health Act (HITECH) Act, specified group of patients, physi- formance measures and methodology (and and other federal statutes. cians may earn a bonus payment reporting formats) for public and private of 0.5 percent on all Medicare bills quality improvement and pay-for-perfor- m Medicare programs authorized by the for 2012. mance initiatives would send consistent Affordable Care Act of 2010. ACA pro- signals to health care providers and move visions with measurement and reporting n Section 3003 (Physician Feedback the health care system in the desired direc- functions include the following:5, 6 Program). Medicare’s Physician tion of improved quality of care and great- Feedback Program, first estab- er value. It would also ensure that busy n Section 3001 (Hospital Value- lished in 2008 and extended and physicians and other health care providers Based Purchasing Program). enhanced by Section 3003 of are not overwhelmed with confusing, con- Medicare is the largest single the ACA, involves CMS’s provi- tradictory information and requirements. payer for hospital payments, and sion of confidential Quality and hospital payments account for the Resource Use Reports (QRUR) to In discussing the challenges of alignment, largest share of Medicare spend- physicians and physician groups participants raised several issues, includ- ing. Medicare’s hospital-oriented: to permit them to compare their ing whether new measures should align value-based purchasing program, performance with that of similar with established measures, whether the established in Section 3001 of physicians practicing in the same alignment of measures should differ from the ACA, links a percentage of specialty. the alignment of targets, and whether an hospital payments to hospitals’ overly prescriptive framework would limit performance on quality measures n Section 3022 (Shared Savings actionable measures. Even with a common related to common and high-cost Program for Accountable Care set of measures, targets for the Medicare conditions such as cardiac, surgi- Organizations). Section 3022 of population may not be suited to the com- cal, and pneumonia care. Starting the ACA authorizes Medicare’s mercial, privately-insured population. in fiscal year 2013 (which begins new Shared Savings Program for on October 1, 2012), CMS will Accountable Care Organizations 1. Existing Measures Used by Public and offer incentives to acute care hos- (ACO). ACOs are groups of health Private Entities pitals based on either how well care providers who volunteer to Physicians and other stakeholders at the hospital performs on certain take responsibility for providing the March 2012 meeting recommended quality measures or how much the the full spectrum of care provided ongoing efforts to align the health qual- hospital’s performance improves to at least 5,000 Medicare benefi- ity and efficiency measures used by HHS compared to its performance dur- ciaries and for meeting specified and CMS for Medicare’s VBPM with the ing a baseline period.7 quality benchmarks (including measures used by other HHS and CMS a mix of process, outcome, and programs, including those authorized by n Section 3002 (Improvements patient experience measures. They the ACA. In addition, they recommended to the Physician Quality and share in savings if medical expen- that CMS consider aligning measures used Reporting System). Medicare’s ditures per capita for Medicare by Medicare with at least some of the Physician Quality and Reporting Parts A and B fall below a certain measures used in quality improvement and System (PQRS), established under benchmark.10 pay-for-performance initiatives undertaken another name in 2007 and broad- A variety of provider-led enti- by other public and private entities (e.g., ened under Section 3002 of the ties may opt to become Medicare physician organizations, large commer- ACA, is a voluntary reporting ini- ACOs, including integrated deliv- Medicare’s Value-Based, Physician Payment Modifier: Improving the Quality and Efficiency of Medical Care page 6 ery systems, physician-hospital • Regional collaboratives for health portfolio of endorsed performance organizations, hospitals with a improvement. CMS might consider measures that may be used to measure primary care physician network, aligning Medicare’s VBPM measures and quantify health care processes, out- and large, multispecialty group with measures used by regional health comes, patient perceptions, and organi- practices. improvement initiatives, such as those zational structures and/or systems asso- located in the far West, the Midwest, ciated with the ability to provide high- n Section 3023 (Bundled Payment the Northeast, and Louisiana.13 quality care.15 CMS also might consider Pilot). By January 13, 2012, Section Regional collaboratives provide action- aligning its measures with those used by 3023 of the ACA requires the able information on the cost and quality accrediting entities such as the National Secretary of HHS to develop and of health care services, the health of the Committee for Quality Assurance subsequently evaluate a national population, and/or the extent to which (NCQA), which operates accreditation, savings program to encourage a community has adopted state-of-the- certification, recognition, and evaluation hospitals, doctors, and post-acute art methods of delivery, payment, and programs for a broad range of health care providers to improve patient health promotion. To foster regional care entities.16 care and achieve savings for the systems-level change, regional collab- Medicare program through bun- oratives aggregate data across payers • Large commercial health insurers. dled payment models.11 Under the and common performance measures. Many large commercial insurers such program, Medicare will offer a as BlueCross BlueShield, UnitedHealth bundled payment for acute, inpa- • Physicians and physician organiza- Group, and WellPoint serve Medicare tient hospital services, physician tions. Measures of health care perfor- beneficiaries on a fee-for-service basis services, outpatient hospital ser- mance best succeed when physicians and have instituted quality improvement vices, and post-acute care services and others involved in health care sup- and pay-for-performance initiatives. For for an episode of care that begins port such measures and, in turn, report four or five years, UnitedHealth has three days before a hospitalization them to the public continuous improve- been operating a transparency program and extends for 30 days following ment efforts.14 Many physician organi- that covers 245,000 physicians across 20 discharge. zations and state medical societies have specialties. It contracts with physicians undertaken initiatives to define quality in many geographic areas, some areas m Initiatives authorized by the Health and implement quality improvement more densely populated than others, and Information Technology for initiatives. Given the vital importance deals with solo physicians as the unit of Economic and Clinical Health Act of physicians’ acceptance of perfor- measurement. UnitedHealth Group has (HITECH) Act. The HITECH Act, mance measures adopted by Medicare built programs analogous to Medicare’s enacted as part of the American and to ensure that Medicare’s measures VBPM and is experienced in address- Recovery and Reinvestment Act of improve quality of care, CMS may want ing the challenges that accompany the 2009, included $22 billion to acceler- to consider aligning Medicare’s VBPM development and implementation of ate physicians’ and other health care with the evidence-based measures of such programs. Because commercial providers’ adoption and meaningful quality backed by a consensus of physi- insurers serve children, some of their use of health information technol- cian organizations. quality measures would differ from ogy—that is, the use of electronic those selected by CMS for the Medicare health records to achieve significant population, but other measures could be • Health care quality organizations closely aligned. improvements in the quality of care and accrediting/certifying entities. as specified in CMS regulations (e.g., It may be useful to establish collabora- entering data essential to creating an tions that align Medicare’s VBPM with electronic health record, using clinical measures used by the National Quality decision support tools, incorporat- Forum (NQF), a public/ private part- ing clinical laboratory results into nership created in 1999 that relies on electronic health records, and using a consensus process among a diverse electronic health records to support group of stakeholders to endorse patient transitions between care set- national standards for measuring and tings or personnel).11 publicly reporting on health care qual- ity measures. The NQF maintains a Medicare’s Value-Based, Physician Payment Modifier: Improving the Quality and Efficiency of Medical Care page 7 2. Framework for Aligning Measures • Section 3014 (Federal Pre- reward performance and fearing that Established Under the Affordable Care Rulemaking Process for Quality competitors might “reverse engineer” Act Measurement). Section 3014 of the all-payer statistics. Congress included several provisions in ACA requires the Secretary of HHS the ACA aimed at fostering the establish- to establish a federal pre–rule-making • Distributed data models for deal- ment of a framework for aligning federal process for the selection of health care ing with payer confidentiality issues measures used in health care quality and quality and efficiency measures in HHS have recently become available. In efficiency measurement: and CMS programs.21 2011, with support from the Robert Wood Johnson Foundation, the Data • Section 3011 (National Strategy for In 2011, pursuant to Section 3014 of the Aggregation Pilot of the America’s Quality Improvement). Section 3011 ACA, the NQF launched the public/ Health Insurance Plans Foundation of the ACA requires the Secretary of private Measure Applications Partnership undertook pilot studies in Colorado and HHS, through a transparent collab- (MAP) to advise HHS on the selection of Florida to demonstrate the feasibility orative process, to develop a national performance measures for public reporting of using a distributed data approach strategic plan to improve the delivery of and performance-based payment programs. to aggregate data on individual physi- health care services, patient health out- MAP is to be guided by the goals and cians from several health plans and to comes, and population health by 2011.17 priorities of HHS’s National Strategy for report aggregated data to physicians on HHS released the first strategic plan— Quality Improvement in Health Care.22 the quality of care they provide to their the National Strategy for Quality The NQF’s diverse membership—con- respective patients.24 Improvement in Health Care—in sumer organizations, public and private March 2011. The plan is to be updated health care purchasers, health care provid- annually and submitted to Congress no ers, and others—makes the organization C. Gaining the Support of Physicians later than January 1 of each year.18 well positioned to encourage the alignment and Physician Groups for the of measures across a variety of public and Measures Used for Medicare’s • Section 3012 (Federal Interagency private quality improvement efforts. VBPM Working Group on Health Care It is essential that physicians accept Quality). As a means to ensure align- In 2011, CMS submitted a list of health the measures of health care quality and ment and coordination across federal care performance measures under con- cost used by CMS in Medicare quality efforts and with the private sector, sideration for use in 2012 to MAP for its improvement and pay-for-performance Section 3012 of the ACA requires review and comment. In a subsequent initiatives such as Medicare’s VBPM. the Secretary of HHS to establish an report, MAP outlined a coordination Physicians have indicated that they need a Interagency Working Group on Health strategy for HHS on federal clinician common set of trusted, timely, and action- Care Quality for the purpose of pro- performance measurement,23 in which able measures—rather than several reports viding a platform for collaboration, MAP evaluated measures published for with varying results—in order to improve cooperation, and consultation among Medicare’s VBPM in the July 1, 2011, their performance. Many physician orga- 23 federal agencies with major responsi- Federal Register against the ideal characteris- nizations and state medical societies have bility for health care quality and quality tics and criteria for such a measure. undertaken initiatives to define quality and improvement regarding quality initia- implement quality improvement initia- tives.19 The working group began meet- An issue related to measure alignment is the tives. Actively involving these groups in ing in March 2011. utility of aggregating data from all payers. discussions on the selection of measures of performance improvement could foster • Section 3013 (Quality Measure • Aggregating data at the regional or stronger alliances and consistent messages. Development). Section 3013 of the national level by using a hybrid of ACA requires HHS, the Agency for Medicare data and private-payer data 1. Engaging Physicians in the Healthcare Research and Quality, and could yield more reliable measures of Selection and Application of Measures CMS to identify (1) quality measures health care providers’ performance. for the VBPM that need improvement, updating, or Large commercial health insurers, how- Physicians paid under Medicare’s fee expansion and (2) gaps in health care ever, have expressed concern about schedule range from solo practitioners in quality measures relative to the National Medicare’s aggregation of data for underserved areas who may serve as many Strategy for Quality Improvement in all payers, noting that insurers have as 8,000 patients in a county to physicians Health Care.20 adequate sample sizes to measure and in large, multispecialty medical groups and ACOs with large staffs, significant Medicare’s Value-Based, Physician Payment Modifier: Improving the Quality and Efficiency of Medical Care page 8 resources, and sophisticated, interoperable before the data’s release to the public; and • CMS could encourage physicians in solo, electronic health records that support qual- (3) an appeals process for physicians who small, and medium-sized practices to ity improvement initiatives. disagree with their evaluations. participate in learning collaboratives at the regional level, perhaps in collabora- The ACA requires HHS, when applying 2. Supporting Physicians’ Quality tion with private payers, local initiatives, the VBPM, to account for the special cir- Improvement Processes or the Center for Medicare and Medicaid cumstances of physicians or groups of phy- Unlike physicians in entities such as Innovation Center or perhaps by using a sicians in rural areas and other underserved ACOs in Medicare’s Shared Savings hybrid approach that involves Medicare communities. Thus, CMS must strive to Program, many physicians in solo, small, data and data from private payers or develop and apply Medicare’s value-based or medium-sized practices who receive other qualified entities. Section 10322 of physician payment modifier in such a way Medicare compensation on a strictly a fee- the ACA requires the Secretary of HHS that it is acceptable to all fee-for-service for-service basis lack the knowledge, staff, to establish a process that allows quali- physicians serving Medicare beneficiaries and resources to undertake ongoing quality fied public and private entities’ use of across the country without imposing an improvement initiatives or develop learn- standardized extracts of Medicare claims undue burden on solo physicians in under- ing health care systems. Participants at the data to evaluate and report on the per- served communities. March 2012 meeting suggested that perhaps formance of service providers and sup- CMS could provide financial incentives pliers on measures of quality, efficiency, The challenge in selecting and applying and/or consider one or more of the follow- and effectiveness of resource use.26 quality measures for Medicare’s VBPM lies ing approaches to help physicians in solo, in supporting the needs of small offices small, and medium-sized practices partici- that struggle to provide care but without pate in collaborative learning initiatives: II. Issues Related to the Selection encouraging them to engage in outdated of Quality and Cost Measures for management practices, such as the use of • Even though Medicare does not con- Medicare’s Value-Based Modifier paper records that could impede innova- tract with physician groups, CMS could for Physician Payment tion and efficiency. set regional or national standards for As of 2013, the ACA requires publication in performance targets and allow physi- the Federal Register of measures of resource • Following an approach adopted by cians in solo, small, and medium-sized use and quality and the analytic measures that BlueCross BlueShield, CMS could select practices to create “virtual” peer groups CMS will use to determine Medicare’s VBPM. performance measures for the VBPM that engage in collaborative efforts for The law does not specify the manner in which that are meaningful for solo practitio- performance improvement and thus quality of care and costs are to be compared ners and practices with no more than 50 “rise and fall together.” Small practices but does stipulate the following with respect physicians and staff. in Arkansas, for example, could be to how quality and costs are to be evaluated encouraged to join a quality improve- for purposes of Medicare’s VBPM:27 • Another approach could use a vertically ment collaborative to participate in a integrated family of measures for the program to reduce overimaging for • Quality of health care is to be evaluated VBPM that could operate at the indi- lower back pain under Medicare’s “to the extent practicable” on the basis of vidual level as well as at higher levels VBPM program. a composite of appropriate outcomes or (ACOs). Participants at the March 2012 other measures established by the Secretary meeting noted that interactions between • CMS could enlist the help of the Quality of HHS and is to reflect quality of care. Medicare’s VBPM fee-for-service physi- Improvement Organizations (QIO) cians and quasi-prospective payment with which Medicare contracts in each • Costs of health care are to be evaluated systems such as ACOs have received state to support the participation of “to the extent practicable” on the basis insufficient attention. solo, small, and medium-sized practices of a composite of appropriate measures in collaborative quality improvement of costs established by the Secretary of Representatives of physician organizations initiatives related to Medicare’s VBPM. HHS. The composite is to eliminate in attendance at the meeting emphasized Medicare’s 10th statement of work for the effect of geographic adjustments in that, regardless of the measures selected, QIOs emphasizes transformational payment rates and to take into account the following factors are important to phy- change in health care and the creation risk factors such as socioeconomic and sicians: (1) transparency and the ability to of topic-specific learning and action net- demographic characteristics, ethnic- drill down into the data used to measure works to spread best practices and spark ity, the health status of individuals, and a physician’s performance; (2) an oppor- change through peer-to-peer learning other factors deemed appropriate by the tunity for physicians to review the data and sharing of solutions.25 Secretary of HHS. Medicare’s Value-Based, Physician Payment Modifier: Improving the Quality and Efficiency of Medical Care page 9 CMS’s selection of quality and cost mea- tent, and performance indicators used in • Promoting effective communication sures for use in Medicare’s VBPM is a the QRURs. As CMS moves expeditiously and coordination of care matter of great concern and interest to to respond to the ACA, it will phase in physician organizations, commercial health the modifier through 2017, although the insurers, and other stakeholders. CMS structure now undergoing development • Promoting the most effective preven- tion and treatment practices for the has indicated that Medicare’s Physician may prove less than optimal over the leading causes of mortality, starting Feedback Program and Medicare’s VBPM long term. The phased approach will help with cardiovascular disease will operate in a complementary manner. signal a proposed course of action with The Medicare Improvements for Patients allowances for the possibility of adjust- and Providers Act of 2008 authorized ments between 2015 and 2017. • Working with communities to promote wide use of best practices to foster Medicare’s Physician Feedback Program healthy living for the purpose of providing physicians A. Selecting Quality Measures for with confidential feedback on the resourc- Medicare’s VBPM es used to provide care to Medicare To send a consistent message to physi- • Making high-quality care more afford- able for individuals, families, employers, beneficiaries. Section 3003 of the ACA cians about priorities, the quality measures and governments by developing and reauthorized and expanded the program. that CMS selects for Medicare’s VBPM in disseminating new health care delivery CMS has used an iterative approach under 2015 and beyond should be aligned with models the program to measure and provide con- the goals and priorities of the National fidential feedback to physicians on their Strategy for Quality Improvement in comparative performance in terms of According to the Agency for Healthcare Health developed by HHS under Section resource use and quality of care.28 Research and Quality, the primary issues 3011 of the ACA. The three broad aims for consideration when selecting measures of the strategy submitted to Congress in of the quality of health care are whether In Phase I of Medicare’s Physician March 2011 follow:31 the measures are “good” and whether Feedback Program in 2009, CMS sent a they are appropriate for the intended small sample of physicians in 12 metro- 1.Better care. Improve overall quality of audience.32 A good measure demon- politan areas prototype feedback reports care by making health care more patient- strates attributes such as standardization, focusing on measures of resource use. In centered, reliable, accessible, and safe. comparability, availability, timeliness, rel- Phase II in 2010, CMS sent QRURs—fea- evance, validity, experience, stability, and turing per capita resource use measures 2.Healthy people/healthy communi- evaluability. One option for categorizing and quality-of-care measures to 36 medical ties. Improve the health of the U.S. a measure as good is its endorsement by groups and to the approximately 1,600 population by supporting proven inter- the NQF or the MAP. To promote trans- individual physicians affiliated with those ventions that address behavioral, social, parency and quality improvement among groups in the same 12 metropolitan areas. and environmental determinants of health care providers, CMS could select In Phase III in 2011 and 2012, CMS is health in addition to delivering higher quality measures for Medicare’s VBPM sending confidential QRURs to 20,000 quality care. that are timely, valid, and important to physicians serving Medicare beneficiaries health care providers. To resonate with in Iowa, Kansas, Missouri, and Nebraska 3.Affordable care. Reduce the cost of consumers, measures of health care quality to enable them to compare their perfor- high-quality health care for individuals, must capture aspects of health care valued mance to the average care and costs of families, employers, and government. by consumers. Medicare patients of other physicians in their specialty in the four states.29 To advance these three broad aims, 1. Selection of Quality Measures for the the 2011 National Strategy for Quality VBPM in 2015 As of January 1, 2015, Section 3007 of the Improvement in Health specifies that Donabedian’s well-known paradigm for ACA requires CMS to apply Medicare’s public and private partners should initially the evaluation of the quality of health care VBPM to payments of selected fee-for- focus on the following six priorities: identifies the following three basic types service physicians and physician groups of measures:33 serving Medicare beneficiaries. CMS has indicated that, in 2015, it plans to • Making care safer by reducing harm caused in the delivery of care • Process measures of quality, which link Medicare’s VBPM to Medicare indicate what processes a health care claims-based and other data from the 2013 QRURs sent to physicians under • Ensuring that each family and each provider has adopted to maintain or person is engaged as a partner in his or improve a patient’s health (e.g., the Medicare’s Physician Feedback Program.30 her care percentage of patients up to age 75 who CMS is continuing to test the design, con- Medicare’s Value-Based, Physician Payment Modifier: Improving the Quality and Efficiency of Medical Care page 10 receive LDL-C screening, the percent- In 2015, data for quality measures sup- measure for Medicare’s VBPM would be age of patients who receive B-blocker porting Medicare’s VBPM will likely be consistent with CMS’s future adoption of treatment after a heart attack) drawn from a subset of the 2013 QRURs PCMH-related measures. data.38 One outcome measure of quality • Structural measures of quality, which that CMS could consider for the VBPM, 2. Selection of Next-Generation Quality indicate the available capacity, systems, though not currently in the QRURs, is a Measures for the VBPM and processes for providing high-quality measure of patient experiences. Unlike • With each year, the measures of health health care (e.g., whether an organiza- clinical outcome measures, measures of care quality become more precise and tion uses electronic health records, the patient experiences and satisfaction do more complex, and the next genera- ratio of providers to patients) not require risk adjustment; instead, they tion of measures will span health care emphasize the importance of “patient- settings and present a more complete • Outcome measures of quality, which centered” care that is applicable to all picture of care.41 indicate the impact of health care inter- types of medical practices and specialties. ventions on patients’ status (e.g., the In addition, patients will appreciate and • Medicare’s VBPM may increasingly rely rate of surgical complications) find it easy to understand measures of on clinical data submitted via electronic patient experiences. health records. CMS is exploring ways Clinical outcome measures of the effects to improve the collection of clinical of health care interventions on patients’ CMS could test the viability of patient data and to encourage the adoption and health status are often considered the experiences as a quality measure by incor- use of electronic health records.42 “gold standard” in measuring quality. porating the measure into the initial phases Such measures require risk adjustment for of modifier implementation. NCQA has • Medicare’s use of risk-adjusted clini- different characteristics within a popula- developed an optional patient experience cal outcome measures of health care tion (e.g., patients’ health status) that are reporting program to help medical prac- quality is an important future target. beyond health providers’ control34 as well tices capture data on patients’ experiences As noted earlier, clinical outcome mea- as a high degree of adoption of electronic (related to access, information, communica- sures require risk adjustment and a high health records. For that reason, clinical tion, coordination of care, self-management degree of adoption of electronic health outcome measures of quality are much support, and shared decisionmaking).39 records. Several physician specialty more challenging to implement than This year, NCQA gave credit to physicians organizations—including the American process or structural measures of quality. (including some particularly small physi- Academy of Ophthalmology, American Moreover, some participants at the March cian practices) for voluntarily reporting on College of Cardiology, and Society of 2012 meeting questioned the feasibility of patient experiences. CMS could adopt a Thoracic Surgeons—are committed to attributing outcome measures, as opposed similar approach for Medicare’s VBPM. the development of patient registries to process measures, to specialists. that will help them monitor clinical NCQA expects to incorporate its patient outcomes of care. Electronic health The vast majority of quality measures experiences measure into its algorithm records linked to patient registries offer available for use under Medicare’s VBPM for the patient-centered medical home the means for active surveillance and are process measures. The 2010 QRURs (PCMH).40 The PCMH is a model for early detection and reporting of adverse under Medicare’s Physician Feedback primary care that has drawn the endorse- outcomes in real time and thus hold Program provided confidential informa- ment of the American Medical Association, enormous potential for improving the tion on physicians’ quality of care by American College of Physicians, and quality and value of health care in a way using a core set of 12 broadly applicable numerous specialty societies as a means to never before possible. process measures for ambulatory care— attract and retain primary care physicians, a subset of the NCQA’s according to improve quality, and lower overall costs. As • Section 3003 of the ACA requires CMS to NCQA HEDIS stands for Healthcare of 2010, the adoption of PCMH processes develop Medicare-specific episode group- Effectiveness Data and Information Set was far greater among the largest medical ing software for Medicare’s Physician (HEDIS®) measure set—that could be groups (those with more than 140 physi- Feedback Program in order to address the calculated with Medicare claims data.35, 36 cians) and practices owned by large entities limitations of proprietary episode groups, such as hospitals, all of which are likely which have limitations when applied to The 2012 QRURs provide confidential to have the resources required to institute people with several chronic conditions. information to fee-for-service physicians PCMH processes; small or medium-sized Many Medicare beneficiaries live with on similar types of measures based on practices do not command the staff and several chronic conditions (e.g., cardiovas- 2010 Medicare claims data and enhanced resources needed to support the imple- cular disease, diabetes, chronic obstructive claims-based quality information submit- mentation of such processes. CMS’s adop- pulmonary disease). ted by physicians to the PQRS.37 tion of patient experiences as an outcome Medicare’s Value-Based, Physician Payment Modifier: Improving the Quality and Efficiency of Medical Care page 11 m Episodes of care represent a group • Overuse (i.e., the provision of tests and Asthma, and Immunology; American of health care services for a health interventions that have no clinical ben- Academy of Family Physicians; American condition (e.g., hip fracture, diabetes) efit yet carry associated risk) College of Cardiology; American over a defined period. Episodes of Gastroenterological Association; American care may occur in a single setting, • Misuse (e.g., medical errors such as Society of Clinical Oncology; American may include both hospital and physi- adverse reactions to drugs, hospital- Society of Nephrology; and American cian services, or may involve the con- acquired infections, surgical injuries) Society of Nuclear Cardiology. In fall tinuum of health care services. The 2012, eight additional physician organiza- hope is that Medicare’s episode-based tions that have joined the Choosing Wisely® approaches to performance measure- • Underuse (e.g., lack of access to pre- initiative will release similar evidence- ventive care for leading chronic diseases ment, accountability, and payment such as cardiovascular disease and dia- based lists of five common tests or proce- will foster greater coordination of betes or the failure to administer beta- dures that have no proven benefit.49 care, reducing fragmentation and blocking drugs to people experiencing a “Estimates suggest that as much as costs associated with the overuse and heart attack) duplication of services.43 $700 billion a year in healthcare costs do not improve health outcomes.” Eliminating the overuse of tests and treat- m Nonetheless, episodes are often dif- ments that may harm patients’ health Peter Orzag, Director of White House ficult to define because of different could potentially translate into safer and Office of Management and Budget, opinions as to which services should higher-quality care while reducing health May 2009 interview with National Public be grouped together. Many Medicare Radio care costs. Jack Wennberg of Dartmouth’s beneficiaries live with several chronic Center for the Evaluative Clinical Sciences conditions such that questions arise has estimated that up to one-third of the over which physician has primary In a related discussion at the March over $2 trillion spent annually on health responsibility for a patient’s care.44 2012 meeting, one participant noted the care is expended on unnecessary or redun- dant tests and procedures.46 complexity of differentiating and reward- m CMS requested proposals from ing physicians’ inactivity versus activity, contractors to develop a prototype particularly when patients expect certain One approach to calculating costs associ- Medicare-specific episode grouper for services. For example, a physician could ated with overuse could involve match- six of nine conditions and has since receive a payment bump under Medicare’s ing the 45 tests and procedures that the selected one of the episode groupers VBPM for not ordering an unnecessary American Board of Internal Medicine for use. It intends imminently to pro- procedure. (ABIM) Foundation’s Choosing Wisely® vide more information about the epi- initiative has identified as overused per sode grouper in a national provider 1. Selection of Cost Measures for the physician utilization data. The initiative is call and plans to test and validate the VBPM in 2015 part of a multiyear effort undertaken by initial grouper software in 2012. As noted, CMS is engaging in an iterative the ABIM Foundation to help physicians become better stewards of finite health process to develop the modifier structure B. Selecting Cost Measures for care resources.47 At a widely publicized and its application. The process involves Medicare’s VBPM press conference on April 4, 2012, the an evaluation in 2015 of the quality and The New England Healthcare Institute per capita costs of care for Medicare’s ABIM Foundation, Consumer Reports, and has defined waste in health care as VBPM that will rely on Medicare claims- nine physician organizations released an “healthcare spending that can be eliminat- based and other data from the 2013 evidence-based list of 45 tests or proce- ed without reducing the quality of care.”45 confidential QRURs developed under dures that have no proven benefit for When selecting cost measures for Medicare’s Physician Feedback Program.50 many patients and sometimes cause more Medicare’s VBPM, policymakers will harm than good.48 need to address the use of health care • The confidential 2010 QRURs in Phase services that increase expenditures without I of Medicare’s Physician Feedback Each of the following nine physician improving patients’ or populations’ health Program provided comparative infor- organizations participating in the Choosing outcomes. mation to fee-for-service physicians on Wisely® initiative identified five tests or procedures that patients and their phy- their resource use in terms of (1) aver- The Institute of Medicine’s (IOM) seminal age annual costs per capita (risk-adjust- sicians should question: the American 2001 report Crossing the Quality Chasm ed and price-standardized) attributable College of Radiology; American College of identified the following three dimen- to the physician’s Medicare patients Physicians; American Academy of Allergy, sions of health care quality: and (2) per capita costs for specific Medicare’s Value-Based, Physician Payment Modifier: Improving the Quality and Efficiency of Medical Care page 12 categories of services. The data for the • As noted, CMS has begun development III. Methodological Issues cost measures were all Medicare Part A of a Medicare-specific episode grouper Related to Medicare’s Value- and B claims submitted by all provid- for Medicare’s Physician Feedback Based Modifier for Physician ers who treated patients attributed to a Program (as mandated by Section 3003 Payment given physician, including providers not of the ACA) and plans to test and Participants at the March 2012 meeting part of the given physician’s medical validate the initial grouper software in discussed a wide range of methodological practice group. 2012.52 issues pertaining to the development and implementation of Medicare’s VBPM, par- • The confidential 2012 QRURs in Phase m Especially in the case of Medicare ticularly as related to the development of III of Medicare’s Physician Feedback beneficiaries living with several chronic composite measures of cost and quality, Program provide information to physi- conditions, questions arise with respect risk adjustment of health care outcomes cians in Iowa, Kansas, Missouri, and to attributing responsibility to a pro- and resource use, and performance mea- Nebraska on (1) per capita spending vider and the costs to be included or surement (e.g., attribution of health care for various types of services; (2) aver- excluded. The broader a cost measure, quality and costs, benchmarking, peer age Medicare spending per patient the greater is the variance and the larger group comparisons, and sample size). (total per capita cost) in 2010; and (3) is the required sample. average Medicare spending per patient A. Composite Measures of Quality for patients with several chronic con- m Participants at the March 2012 meet- and Costs ditions.51 All cost data in the 2012 ing suggested that CMS might have to A composite measure is a combined QRURs have been payment-standard- use different episode groupers for each metric that incorporates several mea- ized and risk-adjusted to account for medical specialty to ensure valid cost sures into a single score.56 As noted, the differences in patients’ age, gender, data. They noted the various subspe- ACA requires construction of Medicare’s Medicaid eligibility, and history of cialties (e.g., eight subspecialties within VBPM to be based on (1) “to the extent medical conditions. Cost information ophthalmology) and pointed out physi- practicable” a composite of health qual- is shown for each physician’s Medicare cians’ enormous variation in resource ity measures and (2) “to the extent prac- patients in various categories (i.e., total use depending on the types of patients ticable” a composite of cost measures. patients for whom the physician filed who make up their case mix. The ACA does not specify the manner any claim, patients whose care the phy- in which the composite measures of sician directed, patients whose care the • The cost and utilization of health care quality and costs are to be compared in physician influenced, and patients to resources vary tremendously among Medicare’s VBPM. whose care the physician contributed). health care providers across the United States.53 At present, it is impossible to How might CMS construct composite know whether such regional variation • One option discussed by meeting par- is attributable to health care practices measures of quality and costs in Medicare’s ticipants is for per capita cost measures VBPM per the ACA’s mandate? for Medicare’s VBPM to include total that constitute “wasteful” spending or Medicare (or health care) expenditures, demographic conditions and are war- ranted.54 In late 2010, at the request of • The Medicare Physician Feedback including Part D expenses. Program’s QRURs have displayed mea- the Secretary of HHS, the IOM began conducting a study to identify the fac- sures of quality and costs as separate 2. Selection of Next-Generation Cost measures.57 Some participants at the tors that may explain geographic varia- Measures for the VBPM March 2012 meeting suggested that the tion in health care spending across the • One possible approach to the selection country.55 The committee performing development of a composite measure of next-generation cost measures for of health quality and cost would be sim- the study is examining how geographic the VBPM could identify key drivers of ilar to the development of a composite variation may or may not be related to cost in different specialties (e.g., using measure of a car’s styling and engine factors such as cost of care, supply of a method such as that pioneered by power—adding together two uncorre- care, quality of care, and health out- Howard Beckman at Focused Medical lated measures would just create noise. comes; diversity within patient popula- Analytics) and then determine the best If CMS is able to separate measures of tions, including the populations’ health practices related to those key drivers. physician quality and cost in Medicare’s status, access to care, and insurance Data from the evidence base of key VBPM, it could avoid the introduction coverage; and physicians’ decisions on drivers could be used to improve the of “noise” into estimates. what care to deliver. quality and reduce the cost of care. Medicare’s Value-Based, Physician Payment Modifier: Improving the Quality and Efficiency of Medical Care page 13 • Convened by the American Medical • Risk adjustment of costs in • Medicare’s Physician Feedback Program Association, the Physician Consortium Medicare’s VBPM. has set a precedent for using different for Performance Improvement58 has rules for the attribution of quality and created a framework for developing m The ACA specifies that measures cost. The current group of QRURs composite measures of health care qual- of costs established by HHS for includes two indicators of physician ity that are evidence-based, patient-cen- Medicare’s VBPM are to eliminate the quality for Medicare beneficiaries: (1) tered, outcome-focused, and testable. effect of geographic adjustments in pay- measures calculated by CMS that rely CMS could consider the framework ment rates and account for risk factors solely on Medicare administrative claims in developing a composite measure of such as socioeconomic and demograph- and (2) Physician Quality Reporting quality for Medicare’s VBPM. ic characteristics, ethnicity, the health System measures submitted to CMS by status of individuals, and other factors PQRS program participants. For PQRS B. Risk Adjustment of Clinical deemed appropriate by the Secretary of quality measures, physicians self-identify Outcomes and Health Care Costs HHS. specific Medicare beneficiaries as their As a necessary precursor to comparing patients. For claims-based quality and m Since 2003, CMS has used the cost measures, the physicians’ Medicare clinical outcomes and health care costs, Hierarchical Condition Categories patient panel includes all Medicare risk adjustment involves the use of a sta- (HCC) model to risk-adjust resources to beneficiaries for whom an eligible physi- tistical process to adjust for differences in account for the case mix of a physician cian filed at least one professional claim patient characteristics that are beyond the practice under the Medicare Advantage in 2010. Each Medicare beneficiary’s control of a health care provider. program.60 The process adjusts for dif- relationship with each physician to ferences in diagnostic history of the whom the beneficiary is attributed is • Risk adjustment of clinical out- practice’s Medicare beneficiaries but comes in Medicare’s VBPM. categorized according to the amount not for differences in the severity of ill- of contact the physician had with the Measures of clinical outcomes are risk- ness; the ICD-9 coding system, which beneficiary (i.e., physician directed the adjusted to ensure that differences in provides the foundation for the HCC beneficiary’s care, influenced the care, or patient characteristics that are beyond model, generally does not measure contributed to the care). a health care provider’s control (e.g., severity levels reliably. patient’s age, medical condition, co- morbidities) do not unfairly affect the In the future, as part of its work in Should Medicare’s VBPM use single- provider’s performance results with developing Medicare-specific episode provider attribution or multiple-provider respect to outcome measures. From the grouping software as mandated by attribution? payer’s perspective, risk adjustment is Section 3003 of the ACA, CMS will essential for ensuring that physicians do continue to study methods of risk- • Single-provider attribution assigns the not have an incentive to avoid observ- adjusting costs at the per capita level patient or episode to the provider who ably high-risk patients. and for specific episodes of care. provides the greatest percentage of patient visits or total costs. m Building on earlier work on risk-adjust- C. Performance Measurement ing hospital mortality and re-admission 1. Attribution of Quality and Costs of m Some minimum number of contacts rate measures,59 CMS is planning and Care to Health Care Providers between the physician and patient over implementing risk-adjustment strategies Attribution is the process used to deter- a specified period of time could be for quality measures for incorporation mine which health care provider or required. into Medicare’s VBPM. Factors used to providers are to be held accountable for m One issue that may arise with attribu- risk-adjust these measures have included the quality and costs of health care.61 tion is how to deal with patients who primary and secondary diagnoses from Attribution associated with Medicare are not local for long periods of time the index hospitalization and condition patients’ care in Medicare’s VBPM will (e.g., snowbirds) but for whom a physi- categories that account for co-morbid- pose a challenge. cian may held be responsible. ities derived from previous-year inpa- tient, outpatient, and physician claims. Physicians and others at the March 2012 Adjustment has been by condition, and meeting suggested that CMS might con- models have been validated. sider using different rules for the attri- bution of quality and cost measures in Medicare’s VBPM. Medicare’s Value-Based, Physician Payment Modifier: Improving the Quality and Efficiency of Medical Care page 14 • Multiple-provider attribution assigns continuous improvement.63 For low-per- tant than physicians’ resource scores patients or episodes of care to more forming physician practices, CMS might in changing both physician and patient than one physician or to other health consider linking benchmarks to improve- behavior. There will be far less variation providers. ment rather than to attainment. in physician quality measures than in cost measures. On the other hand, the dollar m Participants at the March 2012 meeting Comparative benchmarks may be defined in measures will be subject to a high degree noted that, if the goal is to encourage terms of the best performers in a given peer of random variation. One approach sug- physicians to work as part of a team, group (e.g., the 90th percentile), low perform- gested for CMS with regard to Medicare then attribution should be at the team ers (e.g., 10th percentile), or peer group norms VBPM could reward (1) physicians with level rather than at the level of the indi- (e.g., the 50th percentile).64 Each approach has high quality scores but without high costs vidual physician. Mechanisms will need its strengths and weaknesses. and (2) physicians with low costs but to be developed to link physicians to without low quality scores. specific groups, possibly including self- • Average performance benchmarks for identification (as in the ACO program). quality give health care providers deemed 3. Defining Appropriate Peer Groups “worse than average” a relatively attain- for Comparisons of Quality and Cost m One argument for multiple-provider able goal for which to strive. In confiden- Measures attribution is that Medicare’s elderly and tial QRURs recently sent to physicians Benchmarking requires identifying appro- disabled beneficiaries—especially those in Phase III of Medicare’s Physician priate peer groups of health care providers with several chronic conditions —tend Feedback Program, CMS designated the for comparisons. What are appropriate to see several health care providers.62 middle ground of mean or median perfor- peer groups for comparisons in Medicare’s mance as the benchmark for quality. VBPM? Should Medicare’s VBPM rely on patient-based attribution or episode-based • High-performance benchmarks for • Defining specialists and subspecial- attribution? quality acknowledge the best perform- ists remains a challenge for CMS. In ers but may seem unattainable to low the QRURs provided in Phase II of • Patient-based attribution assigns per performers. Low-performance bench- Medicare’s Physician Feedback Program capita costs to one or more provider(s) marks may give low performers an in 2010, CMS used the Health Care and holds the provider(s) accountable incentive to improve but provide no Financing Administration (HCFA) for the entire spectrum of a patient’s incentives for others. Moreover, the medical specialty code submitted by a care. At least at the outset, as noted ear- values at the extreme upper and lower medical professional on his or her 2007 lier, Medicare’s VBPM will use patient- ends of peer group distributions are Carrier Medicare claims to determine based attribution. All patient costs will general less statistically reliable than the specialty of peers to whom perfor- be assigned to one or more physicians. values near the middle. In the confiden- mance and resource use would be com- tial reports sent to physicians in Phase pared.66 If a medical professional listed • Episode-based attribution assigns dis- I of Medicare’s Physician Feedback different specialties in different claims, crete episodes of care (whether in a Program, CMS used high-performance CMS used the medical specialty cited single or multiple settings of care) to benchmarks. in the majority of the medical profes- specific provider(s). As noted earlier, sional’s claims. CMS is developing a Medicare-specific Variation in costs among physicians is episode grouper and hopes to include much greater than variation in quality of • Determining which physicians in a par- episode-based costs in future QRURs care. Moreover, health care costs differ ticular specialty care for sicker patients in Medicare’s Physician Feedback by geographic region.65 The reasons for is another challenge in determining Program, but episode costs are not cur- the variation are not entirely clear and appropriate peer groups for compari- rently required to be a cost component are currently under investigation by the sons. Different specialists and subspe- of Medicare’s VBPM. Institute of Medicine. cialists treat different types of patients, and, in some cases, the resources used 2. Selecting Benchmarks for Quality and If CMS seeks to leverage Medicare’s to treat patients reflect physicians’ Cost Measures VBPM to encourage greater quality and case mix (e.g., cardiologists who treat Benchmarking in the health sector value in the health care system, it might a large number of patients with cardiac involves measuring and comparing the adopt a strategic approach to devising failure and order a large number of performance of health care providers or payment incentives for physicians affected echocardiograms versus cardiologists organizations against that of other provid- by the modifier. Physicians’ quality scores whose patients are not so ill and do not ers or organizations in order to permit for the VBPM will likely be more impor- require as many echocardiograms). Medicare’s Value-Based, Physician Payment Modifier: Improving the Quality and Efficiency of Medical Care page 15 • For episodes of care, one possibility of Medicare payment incentives under VI Endnotes for creating comparisons might be the the modifier could increase. 1. Mathematica Policy Research. “Quality and Re- development of clusters of Episode source Use Reports (QRRs). Frequently Asked Questions” [Internet], 2010 [cited April 6, 2012]. Treatment Groups (ETGs®). ETGs®, • Could CMS give physicians credit Available at: http://www.cms.gov/Medicare/ which were introduced in the 1990s and under Medicare’s VBPM for report- Medicare-Fee-for-Service-Payment/Physician- FeedbackProgram/downloads//2010_QRUR_ reportedly allow adjustments for case ing enriched claims or other data? Is FAQ.pdf. mix, use claims-based information for there an option for weighting certain 2. U.S. Department of Health and Human Services. measuring and comparing health care measures in the early years and moving Centers for Medicare and Medicaid Services. providers based on the cost of patient toward clinical data in later years? “Section 1023 of the Patient Protection and Af- fordable Care Act: Availability of Medicare Data treatment episodes.67 for Performance Measurement.” Background • Given that, at least in the short term, paper for September 20, 2010, listening session [Internet], 2010 [cited April 14, 2012]. Available IV. Phasing in Medicare’s Value- claims data and data reported via at: http://www.cms.gov/Medicare/Medicare- Based Modifier for Physician Medicare’s PQRS will be the source Fee-for-Service-Payment/PhysicianFeeSched/ Downloads/092010_Listening_session_back- Payment of information, is it likely that other grounder.pdf. Beginning in 2015, the ACA requires longer-term sources of data (e.g., from 3. Kaiser Family Foundation, 2011. Medicare’s VBPM to be phased in over electronic health records, patient regis- 4. Center for Medicare Advocacy. “HealthReform: a two-year period. The Secretary of HHS tries) will be available? Linking Medicare Payment to Quality Outcomes” may exercise discretion with respect to [Internet], 2011 [cited April 11, 2012]. Avail- able at: http://www.medicareadvocacy.org/ the specific physicians and physician V. Conclusion InfoByTopic/QualityOfCare/10_06.24.Refor- groups to which the VBPM will apply in This report has identified some of the mAndQuality.htm. that year. In any event, CMS’s actions in observations, concerns, and suggestions 5. Smith R. “CMS Finalizes ACA Hospital Value- 2015 will serve as the foundation for the related to a complex program designed to Based Purchasing Program” [Internet] [cited April 4, 2012]. Health Industry Washington Watch. Available broader implementation of Medicare’s promote value-based health care. While at: ttp://www.healthindustrywashingtonwatch. VBPM in 2017. In 2017, CMS is required the March 2012 meeting participants com/2011/04/articles/regulatory-developments/ cms-finalizes-aca-hospital-valuebased-purchasing- to apply the VBPM to payments to virtu- described the challenges likely to emerge program/. ally all physicians paid under Medicare’s during implementation of the VBPM, they 6. Fiegl C. “Medicare PQRS: “Quality Reporting fee schedule; therefore, CMS may elect expressed some consistent messages. or Else” [Internet], 2012 [cited April 4, 2012]. to increase the number of physicians to Amednews, Available at: http://www.ama-assn.org/ amednews/2012/02/06/gvsa0206.htm. whom the VBPM applies in 2016. Several public and private sector efforts 7. “Outcome. Physician Quality and Reporting are underway to improve health care qual- System (formerly PQRI)” [Internet], 2012 [cited Participants at the March 2012 meeting ity and reduce costs. Alignment of these April 4, 2012. Available at: http://outcome.com/ pqrs.htm. offered several observations with respect to efforts will help minimize the associated the sequencing of the phased implementa- burden and maximize the impact for all 8. McCanne D. “How Does the Affordable Care Act Define ACOs?” [Internet] [cited April 7, 2012]. tion of Medicare’s VBPM for physicians: concerned. Available at: http://pnhp.org/blog/2010/10/28/ how-does-the-affordable-care-act-define-acos/. • It would be helpful to fee-for-service • While CMS is mandated by law to insti- 9. Kaiser Family Foundation, 2011. physicians serving Medicare beneficia- tute the VBPM in short order, a phased 10.U.S. Department of Health and Human Services. ries if CMS provided them with explicit approach will permit the testing of an Centers for Medicare and Medicaid Services. “The Official WebSsite for the Medicare and Medic- information in advance about what initial structure that will lay the founda- aid Electronic Health Records (EHR) Incentive they could earn or would forgo under tion for an iterative process of improve- Programs” [Internet]. Available at: http://www. cms.gov/Regulations-and-Guidance/Legislation/ Medicare’s VBPM. ment of the modifier. EHRIncentivePrograms/index.html?redirect=/ EHRIncentivePrograms/10_PathtoPayment.asp. • When CMS rolls out Medicare’s VBPM The message most clearly articulated by 11.Network for Regional Health Care Improvement. for fee-for-service physicians in 2015, meeting participants was that CMS must “Where Regional Health Improvement Collabora- tives Are Located” [Internet] [cited April 7, 2012]. it will enjoy some flexibility with the provide clear, actionable, and timely guid- Available at:http://www.nrhi.org/locations.html. size of the differential payment for ance to all physicians committed to the 12.National Quality Forum. “The ABCs of Mea- value. Especially at the outset, given application of Medicare’s VBPM. surement” [Internet], 2012 [cited April 7, 2012]. Available at: http://www.rwjf.org/files/re- the uncertainties about the VBPM’s search/71008.pdf. functioning, it may be advisable to offer About the Author 13.National Quality Forum. “Measuring Perfor- physicians only a small payment incen- Kerry B. Kemp is an independent health mance” [Internet], 2012 [cited April 1, 2012]. tive under Medicare’s VBPM. As mea- policy analyst and writer in Washington, Available at: http://www.qualityforum.org/Mea- suring_Performance/Measuring_Performance. surement improves over time, the size D.C. aspx. Medicare’s Value-Based, Physician Payment Modifier: Improving the Quality and Efficiency of Medical Care page 16 14.National Committee for Quality Assurance. ternet] [cited April 4, 2012]. Available at: http:// aspx. “NCQA Programs: Accreditation, Certification, www.gpo.gov/fdsys/pkg/FR-2010-07-13/ Distinction, and Recognition” [Internet], html/2010-15900.htm. 44. Kelley R. “Where Can $700 Billion in Waste Be 2012. Available at: http://www.qualityforum. Cut Annually from the U.S. Healthcare System?” org/Measuring_Performance/Measuring_ 29.U.S. Department of Health and Human Services. [Internet], October 2009 [cited April 15, 2012]. Performance.aspx. Centers for Medicare and Medicaid Services, Available at: http://www.factsforhealthcare.com/ 2012. whitepaper/HealthcareWaste.pdf. 15.Center for Medicare Advocacy, 2011. 30.U.S. Department of Health and Human Services. 45.Robert Wood Johnson Foundation. “The State of 16.U.S. Department of Health and Human Services. Agency for Healthcare Research and Quality. Health Care Quality in America.” [Internet] [cited Centers for Medicare and Medicaid Services. “Selecting the Right Measures for Your Report” April 8, 2012]. Available at: http://www.rwjf.org/ Agency for Healthcare Research and Quality. [Internet], 2012 [cited April 13, 2012]. Available files/research/currentstateofquality.pdf. “Working for Quality” [Internet], 2012 [cited at: http://www.talkingquality.ahrq.gov/content/ April 4, 2012]. Available at: http://www.ahrq. create/select.aspx. 46.Kelley R, 2009. gov/workingforquality/ - nqs. 47.American Board of Internal Medicine (ABIM) 31.U.S. Department of Health and Human Services. 17.Ibid. Agency for Healthcare Research and Quality. Foundation. “Choosing Wisely®” [Internet] [cited “Types of Quality Measures” [Internet] [cited April 8, 2012]. Washington, D.C.: ABIM Foundation. 18.Center for Medicare Advocacy, 2011. April 1, 2012]. Available at: http://www.talk- Available at: http://www.abimfoundation.org/ ingquality.ahrq.gov/content/create/types.aspx. Initiatives/Choosing-Wisely.aspx. 19.U.S. Department of Health and Human Services. Centers for Medicare and Medicaid Services. 32.Ibid. 48.“Do You Need That Test?” New York Times (edi- “Multi-Stakeholder Group Input on Quality torial) [Internet], 2012 [cited April 9, 2012 ]. Avail- Measures” [Internet], 2012 [cited April 11, 2012]. 33.Mathematica Policy Research, 2010. able at: http://www.nytimes.com/2012/04/09/ Available at: http://www.healthcare.gov/law/ opinion/do-you-really-need-that-medical-test. resources/reports/quality12012011a.html. 34.U.S. Department of Health and Human Services. html?_r=2&partner=rssnyt&emc=rss. Centers for Medicare and Medicaid Services. “The 20.Measure Applications Partnership. “Coordination Physician Feedback Program & Quality and Re- 49.American Board of Internal Medicine (ABIM) Strategy for Clinician Performance Measurement” source Utilization Reports (QRURs)” [Internet], Foundation, 2012. [Internet] [cited April 12, 2012]. Washington, 2011 [cited April 5, 2012]. Available at: http:// D.C.: National Quality Forum, 2011. Available at: www.cms.gov/Medicare/Medicare-Fee-for- 50.U.S. Department of Health and Human Services http://www.qualityforum.org/Setting_Priorities/ Service-Payment/PhysicianFeedbackProgram/ Centers for Medicare and Medicaid Services, July Partnership/Measure_Applications_Partnership. Downloads/QRUR_Presentation.pdf. 13, 2010. aspx. 51.U.S. Department of Health and Human Services. 35.Ibid. 21.Ibid. Centers for Medicare and Medicaid Services, 2011. 36.U.S. Department of Health and Human Services 22.Quality Alliance Steering Committee. “Ag- Centers for Medicare and Medicaid Services, July 52. Damberg CI, Sorbero ME et al, 2009. gregating Physician Performance Data across 13, 2010. 53.Institute of Medicine. “Geographic Variation in Health Plans” [Internet], 2011 [cited April 7, Health Care Spending and Promotion of High- 2012]. Available at: http://www.rwjf.org/files/ 37.National Committee for Quality Assurance. “NCQA Distinction in Patient Experience Re- Value Care” [Internet], 2012 [cited April 8, 2012]; research/72031physician.pdf. Availableat: http://www.iom.edu/Activities/ porting” [Internet] [cited April 22, 2012]. Wash- 23.Brown P, Arkansas Foundation for Medical Care. ington, D.C.: NCQA, 2012. Available at: http:// HealthServices/GeographicVariation.aspx. “Medicare 10th Statement of Work: New QIO www.ncqa.org/LinkClick.aspx?fileticket=x9jnmE 54.Van Ochten K. “Why Geographic Variation in Program Aims for Bold Goals, System-Level oT2tw%3d&tabid=1429. Medicare Spending Matters” [Internet], 2010 Transformation” [Internet], December 2011 [cited [cited April 7, 2012]. HealthLeaders Media. Available April 10, 2012]. Available at: http://www.afmc. 38.National Committee for Quality Assurance. “NCQA’s Patient-Centered Medical Home at: http://www.healthleadersmedia.com/page-1/ org/LinkClick.aspx?fileticket=t3RqWELsPAM%3 FIN-259708/Why-Geographic-Variation-in- D&tabid=683&mid=1567. (PCMH)” [Internet]. Washington, D.C.: NCQA, 2011. Available at: http://www.ncqa.org/ Medicare-Spending-Matters. 24.U.S. Department of Health and Human Services. tabid/631/default.aspx. 55.Institute of Medicine, 2012. Centers for Medicare and Medicaid Services, Sep- tember 20, 2010. 39.Rittenhouse DR, Casalino LP et al. “Small and 56.Physician Consortium for Performance Improve- Medium-Size Physician Practices Use Few Patient- ment, Convened by the AMA. “Measures Devel- 25.U.S. Department of Health and Human Services. Centered Medical Home Processes” [Internet], opment, Methodology, and Oversight Advisory Centers for Medicare and Medicaid Services. “Medi- June 2011[cited April 12, 2012] . Health Affairs Committee: Recommendations to the PCPI Work care FFS Physician Feedback Program: Value-Based 30(8):1575-1584 . Available at: http://content. Groups on Composite Measures” [Internet], 2010 Payment Modifier. Background” [Internet], 2011 [cit- healthaffairs.org/content/30/8/1575.full?keytype [cited May 5, 2012]. Available at: http://www. ed April 4,2012]. Available at: http://www.cms.hhs. =ref&siteid=healthaff&ijkey=Cq.8ITSlIscIM ama-assn.org/resources/doc/cqi/composite- gov/Medicare/Medicare-Fee-for-Service-Payment/ measures-framework.pdf. PhysicianFeedbackProgram/index.html?redirect=/ 40.National Quality Forum. “Measuring Performance. PhysicianFeedbackProgram. How Measures Will Serve Our Future” [Internet], 57. U.S. Department of Health and Human Services. 2012 [cited April 2, 2012]. Available at: http://www. Centers for Medicare and Medicaid Services, 2011. 26.Mathematica Policy Research, 2010. qualityforum.org/Measuring_Performance/ABCs/ How_Measures_Will_Serve_Our_Future.aspx. 58.Physician Consortium for Performance Improve- 27.Rau J. “Medicare Moves to Tie Doctors’ Pay ment, Convened by the AMA, 2010. to Quality and Cost of Care.” Washington Post 41.Mathematica Policy Research, 2010. [Internet], 2012 [cited April 15, 2012]. Available 59.U.S. Department of Health and Human Services. at: http://www.washingtonpost.com/national/ 42.Damberg CI, Sorbero ME et al. “Exploring Centers for Medicare and Medicaid Services. health-science/medicare-moves-to-tie-doctors- Episode-Based Approaches for Medicare Perfor- “Listening Session: Physician Feedback Program pay-to-quality-and-cost-of-care/2012/04/14/ mance Measurement, Accountability, and Pay- and Implementation of the Value-Based Payment gIQAFq3IIT_story.html. ment” [Internet], 2009 [cited May 4, 2012]. Avail- Modifier for Fee-for-Service Medicare” [Internet], able at: http://aspe.hhs.gov/health/reports/09/ September 24, 2010 [cited May 5, 2012]. Available 28.U.S. Department of Health and Human Services. mcperform/report.shtml - _Toc239573506. at: http://www.cms.gov/Medicare/Medicare- Centers for Medicare and Medicaid Services. Fee-for-Service-Payment/PhysicianFeedback- “Proposed Rule: Medicare Program; Payment 43.National Quality Forum. “Mutliple Chronic Con- ditions Framework” [Internet], 2012. Available at: Program/downloads//092410_Listening_Ses- Policies under the Physician Fee Schedule and sion_Feedback_Program_Transcript.pdf. Other Revisions to Part B for CY 2011.” Federal http://www.qualityforum.org/Projects/Multiple_ Register 75(133):40040-40709, July 13, 2010 [In- Chronic_Conditions_Measurement_Framework. 60.Ibid. Medicare’s Value-Based, Physician Payment Modifier: Improving the Quality and Efficiency of Medical Care page 17 61.Pantely Susan E. “Whose Patient Is It? Patient 63. Kay JF. “Health Care Benchmarking” [Internet]. 66.Mathematica Policy Research, 2010. Attribution in ACOs.” Miliman Healthcare The Hong Kong Medical Diary [Internet], 2007 [cited Reform Briefing Paper [Internet], 2011 [cited May 5, 2012]; 12(2):22-7. Available at: http:// 67.Ingenix.”What are ETGs?” [Internet], 2012 [cited May 5, 2012]. Available at: http://www.ama-assn. www.fmshk.org/database/articles/06mbdrflkay. April 14, 2012]. Available at: http://www.optu- org/resources/doc/cqi/composite-measures- pdf. minsight.com/content/File/What_are_ETG.pdf. framework.pdf. 64.Mathematica Policy Research, 2010. 62. Damberg CI, Sorbero ME et al, 2009. 65.Institute of Medicine, 2012.