The most important defects in Bernie Sanders’ “single payer” bill, S 1804, are:
- The absence of budgets for institutional providers (hospitals and nursing homes) and clear authority to set uniform fee schedules for individual providers; and
- Section 611(b), which requires the maintenance of the “reform activities” authorized by the Affordable Care Act and MACRA (the Medicare Access and CHIP Reauthorization Act of 2015) for the elderly (the most important one being accountable care organizations, which closely resemble HMOs), and the extension of these “reforms” to the non-elderly.1
In this paper I discuss only the second defect – the “payment reform activities” that Section 611(b) perpetuates and extends. This paper focuses on the seven most important of these.
These and related proposals are often referred to collectively as “value-based payment” (VBP) proposals. What all of these “reforms” have in common is they are worsening racial and income health disparities. These “reforms” have that effect because they all require that doctors and hospitals be given bonuses and penalties based on crude measurement of the two components of their “value” or “merit” – that is, their cost and quality. The crudeness of today’s cost and quality measures guarantees that providers who treat a disproportionate share of the sick and the poor will be penalized while providers who treat a disproportionate share of the healthy and wealthy are rewarded.
A paper in a recent edition of JAMA entitled, “How value-based Medicare payments exacerbate disparities,” observed: “In this [VBP] game, the losers are more likely to be physicians who care for poorer or sicker patients, and, in turn, their patients. ‘We are literally taking money from providers that serve the poor and giving it to providers that serve the rich,’ said Karen Joynt Maddox, MD, MPH, a cardiologist and health services researcher at the Washington University School of Medicine in St. Louis.” If VBP reforms were at least cutting costs, their proponents might argue that the worsening of disparities is an acceptable price to pay. But they aren’t cutting costs either.
When the cost of implementing the “reforms” is taken into account, some of them, possibly all of them, are raising costs.
The seven most important “reforms” authorized by the ACA and MACRA
- the physician Value-based Payment Modifier program and its successor, the Merit-based Incentive Payment System (MIPS) authorized by MACRA,
- “medical homes,” also known as “patient-centered medical homes” (PCMHs),
- bundled payments (payments for particular procedures such as hip replacement that cover out-patient and post-acute care, not just hospital costs),
- the hospital readmission reduction program (there are several, one each for particular diagnoses, such as heart failure),
- the hospital VBP program (which measures mortality for three diagnoses, adverse events, patient “satisfaction,” and cost), and
- the hospital acquired condition program (designed to reduce adverse events in hospitals such as falls and infection).
Proponents of the ACA and MACRA place their hopes for cost containment (and, to some degree, quality improvement) primarily in ACOs, PCMHs and bundled payments. Although the MIPS program is enormous (it will affect most doctors), expert opinion has recently begun to turn against MIPS. Research supporting this conclusion is so solid that the Medicare Payment Advisory Commission (MedPAC) recommended last January that Congress repeal MIPS. So, if we prioritize our homework, we should focus on ACOs, PCMHs, and bundled payments.2
ACOs are very poorly defined. They are defined in aspirational terms, such as, “ACOs are groups of providers who agree to be held accountable for cost and quality.” In a nutshell, this means hospital-clinic chains (most ACOs contain hospitals) agree with an insurer (Medicare, Aetna, etc.) to share insurance risk and to submit to “quality measurement,”which means the insurer measures a tiny handful of medical services (for example, the percent of diabetics who have their blood pressure under 140/90).
The only reliable cost data we have on ACOs are the 500 or so that participate or have participated in one of four ACO demonstrations conducted by CMS:
- *The Physician Group Practice (PGP) demonstration (2005-2010),
- the Pioneer ACO program (2012-2016),
- the Medicare Shared Savings Program (MSSP, 2012 to date), and
- the Next Generation ACO program (2016 to date).
The following table summarizes the results of these four programs.
Net savings and losses of Medicare’s four ACO programs
|Pioneer (32 ACOs in year one, 8 by year five):|
|Next Generation (18 ACOs):|
|Sources: For the PGP demo, see p. 64 of Evaluation of Group Practice Demonstration. For the Pioneer, MSSP and Next Generation programs, see Tables 8-6, 8-3, and 8-7 respectively of Chapter 8 of the MedicarePaymentAdvisory Commission’s June 2018 report to Congress.|
The PGP demonstration saved just 0.3 percent, the Pioneer program saved a few tenths of a percent, the MSSP program has lost a few tenths of a percent, and the Next Generation program, in its first year, saved two-tenths of a percent. If we count the costs the ACOs incurred to pay for the interventions that were supposed to save money (which are reported to be 1 to 2 percent of their Medicare spending), all four of the ACO programs raised total health care spending.
CMS reports that the ACOs improve “quality” based on their performance on three dozen quality measures. But we have no assurance that the improvement on those few quality measures wasn’t achieved by “teaching to the test” (shifting resources away from patients whose care was not measured). Even worse, we do have solid evidence indicating ACOs worsen income and racial disparities.
The Medicare physician Value-based Payment Modifier program was the subject of a paper published last February in Annals of Internal Medicine The authors concluded this program could “exacerbate health care disparities.”Two well-known experts writing in an accompanying editorial did not mince words: They said this study showed that “value-based payment” schemes have a “reverse Robin Hood effect” (that is, they take from the poor and give to the rich), and they characterized the study as the “final nail in the coffin” of pay-for-performance (a synonym for VBP) for doctors.
The MIPS program was modeled after the physician Value-based Payment Modifier program. MedPAC voted last January to recommend to Congress that it repeal MIPS because it is so inaccurate and places an unacceptable burden on doctors (Chapter 15).
Medical homes (PCMHs)
As is the case with ACOs, the PCMH is poorly defined, and reliable research comes primarily from PCMH demonstrations conducted by CMS. The PCMH is essentially a primary care clinic that agrees with an insurer to accept some insurance risk, submit to quality measurement, buy electronic medical records, stay open 24/7, and carry out activities that allegedly advance vaguely defined aspirations such as “team-based care,” “patient engagement,” “focus on the whole person,” and “coordination of care across the medical neighborhood.”
All three of CMS’s PCMH demos have failed. The three demos are:
- the Multi-payer Advanced Primary Care Practice (MAPCP) demo,
- the Comprehensive Primary Care Initiative (CPCI), and
- the Federally Qualified Health Center (FQHC) Advanced Primary Care Practice demo.
Here are their results:
- The FQHC demo raised Medicare’s costs and had a tiny positive effect on a half- dozen process measures and damaged staff morale;4
- the CPCI had no impact on expenditures and “minimal effects on quality;”5 and
- the MAPCP demo raised costs and had minor and mixed effects on quality.6
As is the case with ACO results, these PCMH results do not take into account unreimbursed expenditures primary care clinics incur to start up and maintain PCMHs. The limited research indicates PCMH expenditures are enormous. According to a paper in Annals of Family Medicine, the average annual cost of maintaining a PCMH is at minimum $105,000 per physician. The portion of this amount that is paid for by Medicare and other insurers is not known.
3 The poor results of the Medicare PCMH demos are consistent with earlier research on PCMHs. A 2013 literature review of 19 studies found “no evidence for overall cost savings” A 2014 analysis of the first three years of a PCMH experiment in Pennsylvania reported no savings.https://jamanetwork.com/journals/jama/fullarticle/1832540 A 2014 report on the Veterans Health Administration’s PCMH program found the program is raising costs.
Here is a quote from the RAND report on the FQHC demo on a very understudied problem –the impact of VBP on physicians and nurses:
[W]e also found that clinicians and staff of demonstration FQHCs reported significant reductions in overall professional satisfaction and corresponding increases in stress, burnout, chaos, and likelihood of leaving their practices. These findings suggest that sites with high levels of medical home structures and processes at baseline
were less able to withstand any additional stress associated with participation in the FQHC APCP Demonstration than were sites with comparatively fewer medical home attributes at baseline. That is, having medical home structures at baseline mightitself have been stressful, eroding sites’ capacity to withstand further stress. (pp. xxi-xxii)
This finding is consistent with a report on the first national demonstration of the PCMH concept conducted by the American Academy of Family Physicians between 2006 and 2008. Nutting et al. reported, “The work is daunting and exhausting and occurring in practices that already felt as if they were running as fast as they could.”
Bundled payment (BP) contracts are typically signed between an insurer and a hospital. They “bundle” into one payment numerous services typically billed separately for hospitalized patients – hospital, physician, and post-acute care services (nursing homes, rehab facilities, home care). Again, reliable data come primarily from CMS’s BP experiments.
The data indicate BPs are having almost no effect on cost and mixed effects on quality. Here is how Burns and Pauly summarized the evidence on BPs in a paper published last March in the Milbank Quarterly: “The Bundled Payments for Care Improvement Initiative reported‘modest reductions in Medicare episode payments for select clinical episode groups with isolated instances of quality declines and fewer instances of increased quality.’ These findings held for one set of conditions (orthopedic) but not another (cardiovascular surgery). For the most widely adopted bundled payment model…, spending reductions were due to lower use of post-acute care.” These lower post-acute care costs appear to be achieved both by the hospital demanding lower rates from nursing homes (which merely rewards gigantism) and, at least in some cases, harm to patients.
A recent review of five types of BP contracts between CMS and hospitals concluded bundled payment “was not associated with significant changes in Medicare payments, clinical complexity, length of stay, emergency department use, hospital readmission, or mortality.”
As is the case with ACOs and PCMHs, the data we have on the impact of BPs on cost do not include the cost to providers of participating in BP contracts.7 The costs appear to be high.
According to representatives of “the largest-volume orthopedic surgery facility in California,” the cost of participating in California’s unsuccessful Integrated Health Plan Bundled Payment Demonstration was “large and ongoing.”
As is the case with ACOs and other VBP payment schemes, limited evidence suggests BPs harm sicker patients. For a riveting tale of how one 90-year-old woman with dementia was mistreated by a hospital-clinic chain that signed a BP contract with CMS, see this op-ed by Timothy Hoff, a professor of health policy and long-time believer in managed care, in the Washington Post
Hospital VBP programs
I am lumping into this section brief discussions of three CMS hospital programs – the Hospital Readmission Reduction Program (HRRP), the VBP program, and the Hospital Acquired Condition (HAC) program. There is no evidence that these programs are cutting costs.
CMS began the HRRP and HAC programs in 2012, and the Hospital Value-Based Purchasing Program in 2015. According to a paper on these programs by Charles Kahn et al., “by 2017 hospitals risk losing as much as 6 percent of base operating payments … under all three pay-for-performance programs combined ” (p 1282) Kahn et al. reported that these three programs produce the “reverse Robin Hood effect” – they disproportionately penalize teaching hospitals, which typically treat sicker and poorer patients, and hospitals that treat a high proportion of patients on Supplemental Security Income and Medicaid.8 The HRRP program may be harming patients directly. According to 2018 study, penalties imposed on hospitals with “excess” readmissions of heart failure patients led to a reduction in such readmissions and an increase in
Unlike the other “reform activities” discussed here, the HAC program appears to have achieved concrete benefits for some patients by lowering the prevalence of some adverse events. It is not clear what this achievement cost the hospitals. It seems likely that these same results could have been achieved simply by paying hospitals to carry out whatever interventions they used to lower the adverse- event rate. Paying for these services, as opposed to imposing punishment for not delivering them, would have prevented the HAC program from punishing safety-net hospitals.
VBP programs are not saving money, some or all of them are raising costs, and they are harming poorer and sicker patients.
There are several reasons why VBP programs don’t save money, including: (1) their proponents greatly overestimate the role that excess quantity of medical services, as opposed to the price of those services, plays in driving up US health care costs, and VBP programs focus primarily on quantity; (2) determining when a service is unnecessary is often difficult even for the doctor in the examining room, and even more difficult for supervisors and analysts hundreds or thousands of miles away; and (3) VBP interventions are not free and their costs may outweigh whatever savings were achieved in the form of reduced utilization of medical services.
VBP nostrums inflict harm on poorer and sicker patients because risk adjustment of cost and quality scores (adjustment of scores for factors beyond provider control) is very crude. To give you a sense of how crude, consider this description of the inaccuracy of CMS’sHierarchical Condition Categories (HCC) risk adjustment method, the most studied and perhaps the most sophisticated risk adjuster anywhere in the world today. According to MedPAC’s June 2014 report to Congress, the HCC overpays for the healthiest 20 percent of Medicare beneficiaries by 62 percent and underpays for the sickest 1 percent by 29 percent (see first column in Table 2-1 page 30). The result of crude risk adjustment is that providers who treat a disproportionate share of the sick and the poor are punished and have even fewer resources than they should have, and the threat of punishment in turn creates an incentive to avoid the sick and to deny services to the sick who can’t be avoided.
The research I have reviewed here should convince reasonable people not only that Section 611(b) should be removed from S 1804, but that Congress, state legislators, and regulators should refuse to endorse any proposals that purport to reward “value” unless those proposals have been rigorously tested and demonstrated to be safe and effective.
1 Section 611(b) reads: The ACA and MACRA together authorized dozens of “reforms” that will allegedly influence cost or quality for the better. But experts and the media justifiably focus on just a few of them that affect the most people. They are:
APPLICATION OF CURRENT AND PLANNED PAYMENT REFORMS.—Any payment reform activities or demonstrations planned or implemented with respect to such title XVIII [this is the title in the Social Security Act that created Medicare] as of the date of the enactment of this Act shall apply to benefits under this Act, including any reform activities or demonstrations planned or implemented under the provisions of, or amendments made by, the Medicare Access and CHIP Reauthorization Act [MACRA] of 2015 (Public Law 114–10) and the Patient Protection and Affordable Care Act (Public Law 111–148).
2 For readers eager to do more research and develop your own typology of the “reforms” in the ACA, here’s a link to
a 2105 report by the Commonwealth Fund summarizing the programs authorized by the ACA that allegedly lower
cost, improve quality, or both. http://www.commonwealthfund.org/publications/issue-briefs/2015/may/aca-
payment-and-delivery-system-reforms-at-5-years MACRA merely borrowed the ACA VBP “reforms” and added
MIPS, which is based on the physician VBP program endorsed by the ACA. Thus, it is roughly accurate to say that if
you understand the “reforms” in the ACA, you understand what’s wrong with Section 611(b).
3 The poor results of the Medicare PCMH demos are consistent with earlier research on PCMHs. A 2013 literature review of 19 studies found “no evidence for overall cost savings” https://annals.org/aim/fullarticle/1402441/patient-centered-medical-home-systematic-review A 2014 analysis of the first three years of a PCMH experiment in Pennsylvania reported no savings. https://jamanetwork.com/journals/jama/fullarticle/1832540 A 2014 report on the Veterans Health Administration’s PCMH program found the program is raising costs https://www.healthaffairs.org/doi/abs/10.1377/hlthaff.2013.0893.
Sources: For the PGP demo, see p. 64 of Evaluation of Group Practice Demonstration.
For the Pioneer, MSSP and Next Generation programs, see Tables 8-6, 8-3, and 8-7 respectively of Chapter 8 of the MedicarePaymentAdvisory Commission’s June 2018 report to Congress.
4 See RAND evaluation file:///C:/Users/kiprs/Downloads/RAND_RR886z2.pdf p. 153.
5 See final evaluation by Mathematica Policy Research.
6 See final evaluation by RTI International..
7 Judging from the failure of two prominent private-sector attempts to create BPs, the start-up costs are very high. See this article on the failure of the PROMETHEUS project and this one on the failure of the Integrated Health Association Bundled Payment Demonstration.
8 Kahn et al. report: “Major teaching hospitals are 1.60, 2.58, and 4.04 times more likely than nonteachinghospitals to receive a penalty in the Hospital VBP, the HRRP, and the HAC Reduction Program, respectively….Hospitals with the highest Medicare DSH [disproportionate-share payment] patient percentage … are much more likely … to receive a Hospital VPB program penalty….[H]ospitals with a Medicare DSH patient percentage of 50-65 percent (compared to hospitals with the lowest Medicare DSH patient percentage) are 1.5 times more likely toreceive a HAC penalty.” (pp. 1285-1286)