Whatis the problem? In economics, the law of marginal returnsstates that at a certain threshold, additional inputs produce a decreasingamount of outputs. Put quite simply, more is not always better. In the case ofUS healthcare, this is astoundingly true. Place health care expenditures on thex-axis and health outcomes on the y-axis, and what you get is a display of thisvery principle. There has been inconclusive evidence that increased health carespending actually increases health outcomes.
1 Figure 1: Shows total expenditure per capita in 2010 vs life expectancy atbirth. Source: OECD 2010, “Health care systems: Getting more value for money”,OECD Economics Department Policy Notes, No. 2.2 Althoughsometimes the benefit of healthcare spending is clear cut, such as when acancer therapy enables a patient to enter permanent remission, other times thebenefit is quite hazy, such as an extra dose of chemotherapy for a terminallyill patient. In 2016, the US spent more money on health care per capita, $10,348,than any other country.3 Unfortunately, despite the US’s high percapita health care spending, it ranks poorly in infant mortality, obesity andother quality metrics compared to the rest of the OECD.
4 One of theonly outcomes the US ranks well in is number of MRI machines. Whatis the source of the problem? One explanation is that the current system ofhealthcare rewards volume, such as the number of MRIs ordered, regardless of theactual health outcome. Fee-for-service (FFS) payment systems create a financialincentive for physicians to order as many tests, perform as many procedures,and provide as many services as possible. This generates a market failure inhealthcare, or more specifically, a principal agent problem. Physician andpatient goals are misaligned, with physicians seeking volume, whereas patientsare seeking quality healthcare that will improve their health. This is adeep-rooted systematic problem with historical incentive structures thatproduce overutilization, not an individual problem.
Fee-for service paymentmodels negatively affect physicians, patients, and the entirety of the USeconomy. Whatcan be done? Apromising value-based model known as MACRA (Medicare Access and CHIP ReauthorizationAct) has emerged and is already in stages of implementation. This paper willexamine the fee-for-service market failure, as well as value based care models,and how they will realign physician and patient incentives. Background: Whatis Fee-for-Service and why does it matter? Fee-for-servicepayment models have traditionally been the dominant form of payment in the US.For each service or procedure, a physician receives payment regardless ofquality or outcome. This type of payment model has largely shaped the way inwhich medicine is practiced in the US. In 1980, approximately half of allphysicians were compensated by fee-for-service and of the remaining half,roughly 30 percent received a mixed form of compensation with incentivessimilar to fee for service8. Another study shows that over 90% ofprimary care practice revenue is generated through fee-for-service.
6A separate study showed that in 2013, nearly 95% of all provider visits usedfee for service payment methods.5 Impactof Fee-for-Service on Healthcare StakeholdersAlthoughnot all physicians enter medicine for the sake of financial wealth, financialincentives do have an impact on their decision making process. With the knowledge that volume will be thedriver of their compensation, physicians are more likely to prescribe or ordera test than not in the face of an unclear situation. Fee-for-service impactsphysicians in a very direct fashion, shaping the decision process andutilization of resources for physicians. As a result of influencing physicianbehavior, fee-for-service payment models have important implications forpatients. The danger in fee-for-service is not only the high costs associatedwith this style of practicing, but also the potential negative effects onpatients of overutilization.
The principal agent problem created by FFS causespatients to be vulnerable to excessive and unnecessary medical services that arecostly and potentially harmful with no benefit. A study in 2006 found thatMedicare patients with alternative payment plans rather than thefee-for-service payment plans had a shorter average length of stay and a lowertotal cost per hospitalization.7 Although it is difficult toascertain that fee-for-service is the culprit, the dangerous incentivestructure created by FFS aligns with the result seen. Fee-for-service has thepotential to increase costs as well as cause harm through overutilization. Results: Solutionsoffered by the ACA: Therehave been attempts at implementing quality or value based payment models that seekto uproot FFS payment models with varying success.
The Affordable Care Act in2010 introduced a number of elements into the US healthcare ecosystem thatattempted to move away from sheer volume of services, to more quality,coordinated care. TheAccountable Care Organization (ACO) was created by the ACA and are groups ofdoctors, hospitals, and healthcare providers that give coordinated quality careto Medicare beneficiaries. Under the ACA, each ACO manages a minimum of 5,000beneficiaries. If ACOs are able to demonstrate certain quality benchmarks, theyare able to receive shared saving payments.9 Many of the patientswithin ACOs are still under fee-for-service, but the overall ACO programrepresents a shift towards quality and value of care. There are also ACOprograms for Medicaid and pediatric services. TheCenter for Medicaid and Medicare Innovation (CMMI) created by the ACA alsoestablished a number of bundled payment programs that seek to pay physiciansfor an episode of care. CMMI created the Medicare Bundled Payments for CareImprovement Program, an initiative that pays providers a set payment perpatient episode.
9 By paying per episode, bundled payment programseliminate the incentive to drive up volume of services as in FFS. Patientcentered medical homes are primary care practices that meet certain certificationrequirements. Much like an ACO, a patient centered medical home is a team basedmodel that seeks to provide coordinated care by having the primary carephysician lead the treatment model. A PCMH has five characteristics: comprehensivecare, patient-centered care, coordinated care, accessible services, and qualityand safety. It attempts to provide value through this coordination of care, andbringing together.9 Althoughmany of these care models have shown promise, especially in the realm of carecoordination, none of them have displaced fee-for service. Many of theseprovisions were introduced with the ACA in 2010, and fee-for-service continuesto dominate.
Solutionsoffered by MACRAOneof the most recent policies regarding physician payment is the Medicare Accessand CHIP Reauthorization Act of 2015 (MACRA), and it has been one of thebiggest shifts towards value seen in the US healthcare system. MACRA repeals the long standing SustainableGrowth Rate for Medicare reimbursement, requires that doctors pick either theMerit-Based Incentive Payment System (MIPS) or a separate advanced alternativepayment model (APM) for Medicare payment, and seeks to reward providers basedon value.9 This requirement will begin in 2019, although manyorganizations have already begun preparing for the shift it represents. UnderMIPS, providers are given a composite performance score that is based on 4categories.
These categories are quality, cost performance, clinical practiceimprovement activities, and meaningful use of certified EHR technology oradvancing care.10 Initially in the 2017 performance year, qualitymetrics will make up 60% of the composite score, improvement activities 15%,and advancing care 25%. By the 2019 performance year which affects 2021payments, that proportion will shift to 30% quality metrics, 30% costperformance, 25% advancing care information, and 15% improvement activities. From2017-2019, the weight of quality goes down and the weight of cost goes up.10This score will determine whether providers will receive positive, neutral or negativeadjustments.
This payment adjustment will depend on the year and start at 4% in2019, then incrementally increase to 9% in 2022. So, if a provider is able toscore highly, the payment adjustment could award up to 9%, whereas a poor scorecould result in a 9% penalty. MIPS seeks to encapsulate many of the value basedprograms that used to exist for Medicare such as physician quality reportingprogram, value based payment modifier and the Medicare EHR incentive program.10Many physicians will still operate on fee-for-service, but will have addedincentive to perform on these separate quality measures.
This helps toalleviate the principal agent problem, through giving physicians financialincentive to produce valuefor patients, not just volume. Figure shows howpayment adjustments will increase until 2022. Taken from CMS report on MACRA “Pathto Value”.10So,although providers will still be paid through fee-for-service largely, MACRAhas the potential to drive practice significantly towards value as seen by theplus or minus 9% that could impact revenue.
MIPS will not apply to first yearMedicare Participation providers, providers in APM models, or providers belowthe low volume threshold.10 Alternativepayment models is the second pathway through MACRA, and providers whoparticipate in APMs will automatically receive a 5% payment bonus. Unlike MIPS,APMs often offer a completely different model for payment that can potentiallyshift away from fee for service. Although CMS estimates that most providerswill initially opt into MIPS, there exist APMs such as the bundled care paymentinitiative, the Medicare Shared Savings Program, and Next Generation ACOs whichproviders could choose to pursue in place of MIPS.
10 Analysisof MACRASmaller practices will facenegative penalties the most under MACRA. CMS estimates that in year one of MACRA, 2019, 87% ofclinicians in solo practice will face negative payment adjustments throughMIPS. For groups with 10-24 clinical professionals, 59.4% are expected to facenegative payment adjustments. For groups of over 100 clinical professionals,only 18.3% of groups are expected to receive negative payment adjustments.11Figure takenfrom CMS MACRA proposed rule.
11 Shows how different practice sizeprovider groups will be impacted under MIPS. Different specialties aremore likely than others to face negative penalties under MACRA. Based on CMS’s proposed rule of MACRA in2016, different specialties face big differences in the percentage ofphysicians that would receive a negative adjustment. In cardiology, 62.1% ofphysicians would receive a positive adjustment, and in pediatrics, 79.
3% ofspecialties would receive a positive adjustment. On the other hand, 98.4% ofchiropractors would receive a negative adjustment, and 68.8% of psychiatristswould receive a negative adjustment. The differences in adjustment are shown ina self-generated graph below.
11Figuregenerated from data in the CMS MACRA Proposed rule document.11 Many physicians still havea poor grasp of what MACRA entails.A survey conducted jointly by KPMG and AMA in May of 2017 attempted to uncoverhow physicians understood MACRA.13 They surveyed 1,000 physiciansacross different specialties. According to the survey, only about 51% ofphysicians were even somewhat knowledgeable about MACRA. 90% of physicians alsofelt that the MIPS requirements were slightly or very burdensome.13Another survey for health solutions completed by Deloitte showed that nearly80% of physicians preferred fee-for-service or salary based compensationcompared to value models.
79% of physicians disagree with tying compensation toquality of care.12 Discussion: Fee-for-servicehas dominated the provider payment landscape for decades. As physician paymentmodels begin to transition from entirely fee-for-service, to value-basedadjustors, providers will need to re-evaluate the way they practice and whatdata they collect to meet the demands of this new legislation. MACRA attemptsto tie fee-for-service payments in Medicare to quality or value, as an adjustorbut not as a replacement. Regardless, MACRA represents a significant transitiontowards a value based model and seeks to create incentives to realign physicianand patient incentives.
Limitationsof MACRAOneof the inherent problems with MACRA lies in the sheer amount of informationreporting required by physicians. Of the four categories that determine themodifier in MIPS, physicians need to report at least 6 quality measures,including one outcome measure. The problem with using quality measures acrossthe US for the same hospitals is that hospitals simply experience variation onso many levels that make this kind of competition unfair. Certain hospitals incertain geographic, demographic or risk areas will inherently have differentbase values for these kinds of quality and outcomes regardless of the qualityor value of care the provider actually gives. Given this, providers whoactually treat a healthier population will be rewarded as opposed to those thatmay take on unhealthier patients.
As an example, a very experienced and renown heartsurgeon may not perform strongly on many of these quality metrics, as he or shemight be dealing with some of the most complex and probabilisticallyunfavorable patients. His or her low quality metrics in this case, are not arepresentation of lack of quality care, but rather represent the initial patientpopulation. In this way, fee-for-service represents a seemingly purer form ofcompensation, based on inputs rather than based on health outcomes. In thiscase, there is a huge incentive to not treat the unhealthiest patients, as theywill potentially lead to worse quality metrics and hence a penalty incompensation. Alsoas seen in the analysis of MACRA, different specialties have different likelihoodsof being penalized by MACRA.
Different types of practitioners have differentrisk pools of patients, have different levels of effectiveness in managing andtreating illness, and will hence have different levels of quality metrics.Comparing outcomes in a field like oncology to a field such as dermatologymight give drastically different proportions of favorable outcomes. Accountingfor these differences also becomes paramount in a value based care model.
Beyonddifferences in the kinds of patients providers actually treat, the utilizationof data for the purpose of physician compensation also comes with limitations.In any case where there is reporting of metrics there is potential for abuse orgaining of the system. If provider groups are able to artificially inflate someof their quality metrics through a) accepting only healthy populations or b)providing care that specifically targets those numbers without actuallyimproving patient care, there is potential to take advantage of the value basedmodifiers in place. Out of the 6 quality measures needed to be reported forscore calculation, there are usually close to 200 measures to choose from. Acertain provider that chooses 6 completely different quality measures fromanother provider might result in entirely different score calculations notbecause actual quality of care is different, but the quality measuresthemselves inherently create differences. Furtherthan that, data is not always an accurate representation of the clinicalpicture. Although many people become astounded with the potential of”data-driven solutions” as an end all be all savior to all problems, thereality is that data is filled with bias in initial collection, in how it isreported, and what it actually represents.
Despite the seemingly complex andconvoluted process and guidelines of MACRA, the determination of the MIPScomposite score seems overly simplistic, ignoring the existence of many ofthese biases. But what can be done? Physicians need to understand the paymentmodel, and the more complex calculation of an adjustment is, the more difficultit will be for information transparency. As seen in the analysis of MACRAsection, many providers do not even fully understand MACRA, and adding layersof complexity would only serve to further widen this gap in understanding. Complex models would be ignoring the fact thatphysician comprehension is also a key factor in the success of these programs.
A balance must be struck between the complexity of the model and the extent towhich physicians can operate and understand these models. Physicians aretrained to understand the body, biological mechanisms, and clinical features,not payment and statistics. ConclusionItis clear that MACRA is a step in the right direction towards value-based care.
It provides direct incentives for physicians to provide value and quality notjust volume. MACRA aims to realign patient and physician incentives, andeliminate the principal agent problem produced by fee-for-service, increasingquality of care for patients. Despite its shortcomings, MACRA has realpotential to fix this market failure, and as MACRA begins to take effect, itwill be important to observe how this new value based legislation will affectthe US healthcare system.