In April 2016, the Centers for Medicare and Medicaid Services (CMS) enacted a major policy change by mandating for the first time that providers participate in an alternative payment model. The Comprehensive Care for Joint Replacement (CJR) program requires nearly 800 hospitals and thousands of physicians across 67 markets to accept bundled payments for major joint replacement of lower extremity (MJRLE). Hospitals must bear financial risk for care episodes beginning with elective MJRLE admission and extending 90 days after discharge. In July, CMS announced its intention to add hip and femur fracture to CJR and expand mandatory bundles to patients admitted for acute myocardial infarction (AMI) or coronary artery bypass graft surgery in 98 markets.1
These mandatory bundles build upon the Bundled Payment for Care Improvement (BPCI) program, an ongoing CMS demonstration in which 539 hospitals and physician practices are voluntarily participating under “Model 2,” and accept bundled payment for episodes that encompass acute hospitalization and postacute care.
Although data from independent evaluations are lacking, early evidence from BPCI Model 2 suggests that bundled payment may benefit patients and providers.2,3 Medicare's own evaluation after the second year of BPCI demonstrated an approximate 8% decrease in per episode spending for MJRLE bundles with stable-to-improved quality.3 Patient-reported functional outcomes, such as return to mobility and pain, also seemed to improve among a subsample.
Despite these results, however, one major concern about bundled payment—raised again in response to the recent CMS study—is that overall Medicare spending may not decrease if providers respond by increasing the volume of procedures.4 Such a compensatory response could occur because in contrast to population-based payment models such as Accountable Care Organizations, under bundled payment providers retain the incentive to do more because they are paid on a per-episode basis. For example, additional analysis of the CMS study demonstrated that even though per-episode spending for MJRLE was lower among BPCI participants, both case volume and total MJRLE episode payments were higher than in nonparticipant hospitals. The feared attributes of this volume response are potential demand inducement by physicians and hospitals attempting to add volume, as well as inappropriate patient selection that leads to more procedures for healthy or low-risk patients and decreases access for higher-risk patients.
In this perspective, we argue that although policymakers should be vigilant about unintended consequences such as inappropriate patient selection, neither volume increases nor a resulting lack of spending reductions inherently signals policy failure. Instead, if procedures become higher-quality and lower-risk under bundled payment, volume increases may actually be patient-centered.
VOLUME INCREASES DO NOT NECESSARILY SIGNAL POLICY FAILURE
Although hospital spending on MJRLE decreased more slowly in BPCI participant hospitals (6.4% vs 7.7% among nonparticipants) and case volume increased, data suggest that the procedure episode itself likely became cheaper, safer, and better.2,3 It is not inherently perverse for more patients to receive a procedure as it becomes higher value. In fact, we may expect patients to appropriately demand more of it. For example, high-risk patients may perceive increased benefits relative to risks and be more inclined to pursue surgery, driving up volume.
Despite criticism that mandatory bundled payment may lead to compensatory volume responses, possibly from providers inducing demand for MJRLE, and preferentially operate on healthy or low-risk patients, data differentiating between appropriate and inappropriate volume increases remain weak. Existing evidence relies on analyses of geographic markets in which hospitals do not uniformly participate in bundled payment and shifts of low-volume procedures from nonparticipating to participating hospitals—rather than an overall increase in the number of MJRLE performed—could explain the majority of volume increases.5 Thus, inferences about policy failures driving volume increases cannot simply be extended to markets with mandatory bundled payments at all hospitals in a geographic area, like those under CJR.
Vigilant monitoring for inappropriate patient selection that reduces access to care or worsens disparities6 remains critical to ensuring that patients benefit from bundled payment. However, in the absence of conclusive evidence about which patients experience the greatest improvements in risk and benefit, our intention is to highlight that volume increases, and the resultant lack of reductions in overall Medicare spending, are not necessarily negative.
MAXIMIZING VALUE THROUGH SHARED DECISION-MAKING
One mechanism to ensure that volume increases for MJRLE or other conditions are appropriate and patient-centered is to require shared decision-making (SDM), perhaps as part of a systematic process to reduce overuse using appropriateness use criteria and personalized risk and informed consent as well.7 SDM is critical for helping patients understand and weigh risks and benefits of elective MJRLE surgery in the context of their preferences. A Cochrane Collaborative review of 86 studies demonstrated that SDM decision aids resulted in increased knowledge, more accurate risk perceptions, and up to 20% fewer patients selecting invasive surgical options or aggressive treatment.8 The application of decision aids to MJRLE resulted in 26% fewer replacement surgeries and 38% fewer knee replacements.8
Medicare already requires SDM as part of its payment policies in other clinical scenarios. For example, physicians must document a lung cancer screening counseling and SDM visit before Medicare will reimburse for a low-dose screening chest computed tomography. The requirements also include the use of ≥1 decision aids, to include benefits and harms of screening, impact of comorbidities, and ability or willingness to undergo diagnosis and treatment, and follow-up testing. Furthermore, in December 2016, Medicare announced the Shared Decision Making Model demonstration project, which will test a structured 4-step SDM process within accountable care organizations. Hip and knee osteoarthritis will be among the 6 preference-sensitive conditions included.
Medicare could similarly require an SDM visit with a tailored, approved decision aid before MJRLE surgery to promote appropriate higher value surgeries while deterring inappropriate ones. To enhance patients’ ability to make informed decisions, the SDM process could also require transparency of performance data, such as hospital- and surgeon-specific procedural volume, complication rates, and functional outcomes. Data transparency would also strengthen the incentives for physicians and hospitals to prioritize quality, patient satisfaction, and access to care.
Early results suggest that bundled payment may increase the value of MJRLE through lower costs and stable to improved quality. However, amid Medicare's effort to advance bundled payment through an expansion of mandatory programs, there are concerns that compensatory volume increases may represent policy failure via inappropriate demand inducement by physicians and hospitals and/or inappropriate selection of patients. On one hand, it is critical to monitor for unintended consequences of MRJLE under bundled payment, particularly given the recognized racial disparities in access associated with the procedure.6 On the other hand, it is important to recognize that as we await more definitive evidence, shifts toward low-risk procedures or slower decreases in Medicare spending under mandatory bundles may reflect increased value and patient-centeredness rather than ineffective or problematic policy. One strategy for increasing the chances that bundled payment achieves intended results is to incorporate requirements such as SDM and data transparency into policy design.
The authors thank Mark V. Pauly for comments on an earlier version of this manuscript, for which he was not compensated financially.
1. Notice of proposed rulemaking for bundled payment models for high-quality, coordinated cardiac and hip fracture care (Accessed on October 1, 2016 at https://www.cms.gov/Newsroom/MediaReleaseDatabase/Fact-sheets/2016-Fact-sheets-items/2016-07-25.html
2. Navathe AS, Troxel AB, Liao JM, Nan N, Zhu J, Zhong W, Emanuel EJ. Cost of Joint Replacement Using Bundled Payment Models. JAMA Intern Med. Published online January 03, 2017. doi:10.1001/jamainternmed.2016.8263.
3. Dummit LA, Kahvecioglu D, Marrufo G, et al. Association Between Hospital Participation in a Medicare Bundled Payment Initiative and Payments and Quality Outcomes for Lower Extremity Joint Replacement Episodes. JAMA. Published online September 19, 2016. doi:10.1001/jama.2016.12717.
4. Fisher ES. Medicare's Bundled Payment Program for Joint Replacement: Promise and Peril? JAMA. Published online September 19, 2016. doi:10.1001/jama.2016.12525.
5. Alexander D. How do doctors respond to incentives? Unintended consequences of paying doctors to reduce costs. Available at: http://scholar.princeton.edu/sites/default/files/dalexand/files/alexander_jmp.pdf
. Accessed October 12, 2016.
6. Ibrahim SA, Kim H, McKonnell KJ, et al. The CMS Comprehensive Care Model and Racial Disparity in Joint Replacement. JAMA
7. Milford CE, Hutter MM, Lillemoe KD, et al. Optimizing appropriate use of procedures in an era of payment reform. Ann Surg
8. Arteburn D, Wellman R, Westbrook E, et al. Introducing decision aids at Group Health was linked to sharply lower hip and knee surgery rates and costs. Health Affairs