This is the second of a two-part series on MedPAC’s October 4 decision to recommend the repeal of the MIPS program. In Part One , I gave the MedPAC staff credit for urging the commission to support repeal of MIPS, and I criticized their irrational proposal to replace MIPS. I said MedPAC is stuck in a vicious cycle – they recommend “reforms” without evidence, and when the reforms don’t work, they recommend evidence-free tweaks that don’t work either. I referred to this vicious cycle as a “tar pit.”
In this essay I attempt to explain how MedPAC created this intellectual tar pit. I begin by describing the three most important “reforms” in MACRA – pay-for-performance, ACOs, and “patient-centered medical homes.” Then I review the decisions MedPAC made, starting in 2003, that led them to endorse those “reforms.” We will see a pattern: MedPAC adopts “reform” proposals based on opinion, not evidence, and MedPAC never works out the details of their evidence-free proposals but instead foists that responsibility on Congress or CMS.
The three pillars of MACRA
If I were asked to explain MACRA (the Medicare Access and CHIP Reauthorization Act) to someone who wasn’t familiar with it, I would start like this: “MACRA imposes a pay-for-performance (P4P) scheme on all doctors who participate in Medicare’s fee-for-service program. This program is called the MIPS program. Doctors who want to escape the MIPS program must join either an ACO or a ‘patient-centered medical home (PCMH).’”
Those of you who are familiar with MACRA will have noticed that I left out the handful of small-bore “bundled payment” programs that doctors could enroll in to escape MIPS. But those programs apply to relatively small pools of patients with specific diseases, not patient populations in the tens of thousands as ACOs and PCMHs do.
If you accept my summary description of MACRA, then you must also accept this statement: If P4P, ACOs, and “medical homes” don’t work, MACRA can’t work.
P4P must work at the level of the individual doctor if MIPS is to work, and it must work at the group level if ACOs and “homes” are going to work as advertised.  And ACOs and PCMHs must work if doctors are going to have some place to run to escape MIPS, and if Medicare is going to save money on the ACOs and PCMHs that doctors are expected to run to. Not one of the three nostrums essential to MACRA’s success – P4P, ACOs, and PCMHs – has worked (they do not lower costs and have mixed effects at best on quality). Yet MedPAC enthusiastically endorsed all of them.
MedPAC endorsed the Three Nostrums in rapid succession between 2003 and 2008: It endorsed P4P for all providers (including individual doctors) in reports issued between 2003 and 2005; it was the first proponent of the ACO (it invented the term in 2006); and in 2008 MedPAC endorsed “medical homes.” MedPAC endorsed each of the Three Nostrums without a shred of evidence that they would work.
Monkey see, monkey do
In the wake of the “HMO backlash” of the late 1990s, chastened managed care proponents looked for less visible, and therefore less annoying, tactics than those pioneered by HMOs. Pay-for-performance was the first of these new tactics to emerge, and MedPAC was among the first to jump on the P4P bandwagon.
MedPAC put Congress on notice in Chapter 7 of their June 2003 Report to Congress http://www.medpac.gov/docs/default-source/reports/chapter-7-using-incentives-to-improve-the-quality-of-care-in-medicare-june-2003-report-.pdf?sfvrsn=0 and Chapter 4 of their June 2004 report http://www.medpac.gov/docs/default-source/reports/June04_Entire_Report.pdf?sfvrsn=0 that they intended to endorse P4P on “quality” for all providers – individual physicians as well as institutional providers like hospitals and nursing homes. They formally endorsed P4P for physicians, hospitals and home health agencies in a report http://medpac.gov/docs/default-source/reports/Mar05_EntireReport.pdf?sfvrsn=0 issued in March 2005. “We come to this year’s recommendations by determining that quality measures can be used to distinguish among hospitals, home health agencies, and physicians,” said the commission in the 2005 report. “Where necessary, adequate risk adjustment is available. Data needed to take these measurements can be collected without undue burden on providers or the program.” (p. 184)
Twelve years later, at its October 4, 2017 meeting, MedPAC’s staff urged the commissioners to reverse these positions as they applied to individual doctors. It wasn’t true, it turns out, that quality can be measured at the individual level, that “adequate risk adjustment is available,” and that “data …can be collected without undue burden.” The staff did not, however, remind commissioners that 12 years earlier MedPAC had told Congress just the opposite. The commission should have been alerted to that fact, and the commissioners should have investigated how their predecessors made such a serious mistake.
Had the commissioners bothered to do that, they would have discovered that the 2005 decision to recommend P4P for individual physicians and other providers was adopted because General Motors, Leapfrog and other business and insurance industry groups were experimenting with P4P. The decision was simply a case of monkey see, monkey do. MedPAC admitted as much. Here is how MedPAC described their “research” on P4P in their June 2003 report: “Through our interviews we find that many purchasers and plans are experimenting with incentives for improving quality.” (p. 111) And in their March 2005 report they said this: “To determine whether it is feasible for Medicare to pay for performance we consulted with quality experts, providers, researchers, purchasers, CMS, the NQF, and accreditors. It is their hard work and enormous progress in improving quality measurement that provide the foundation for these recommendations.” (p. 184) Note that this “foundation” was not research, but gossip picked up from interviews with “experts” in business and agencies charged with promoting managed care.
Research on (as opposed to gossip about) P4P was almost non-existent in 2005. A 2006 edition of Medical Care Research and Review devoted entirely to the emerging P4P fad stated, “P4P programs are being implemented in a near-scientific vacuum.”  In the same edition of that journal, Glenn Hackbarth, a former HMO executive who chaired MedPAC at that time, offered this lame explanation for why MedPAC endorsed P4P in such a vacuum:
Why is MedPAC confident that P4P is the proper thing to do, especially given the limited amount of hard evidence on its impact? Two reasons. First, there is overwhelming research documenting the poor performance of our health care system…. The status quo is unacceptable…. Second, there is abundant evidence that health care providers respond to incentives. For people with substantial experience in health care delivery and policy, like the MedPAC commissioners, it does not seem like much of a leap to conclude that P4P is a step in the right direction. 
Hackbarth’s justification for endorsing P4P boiled down to, “Things are so bad, experts like us are justified in recommending pretty much whatever we want.” Even his claim about the evidence on system “performance” (“overwhelming research” demonstrates “poor” quality is rampant) was faith-based. As the Institute of Medicine put it in Crossing the Quality Chasm (the New Testament of the managed care movement) in 2001, “The concern about quality arises more from fear and anecdote than from facts; there is little systematic evidence about quality of care in the United States.” (p. 231)
In the case of the ACO, MedPAC was the first monkey that others copied. The label “accountable care organization” was invented during a conversation between commissioners and Elliott Fisher, the “father of the ACO,” at a MedPAC meeting in 2006 following a presentation by Fisher on an algorithm he invented that lumped doctors and their patients arbitrarily with nearby hospitals. Several commissioners indicated they doubted Fisher’s vaguely defined groups were anything more than artificial constructs. Fisher had no answer for them. And most importantly, he had no evidence that even ACOs with some internal cohesion could cut costs and improve quality. Nevertheless, the commission went on to become a fervent advocate of the amorphous ACO concept.
MedPAC’s endorsement of the equally amorphous PCMH in its June 2008 report to Congress was another evidence-free decision. The report identified several vague and unproven criteria for “homes,” including use of electronic medical records and email to communicate with patients, and it recommended that P4P be inflicted on PCMH doctors. The report implied that MedPAC’s endorsement was based on interviews with “experts,” but even that was not clear. What is clear is that MedPAC was once again making decisions about abstractly defined “reforms” based on folklore, not evidence.
Thus, by the summer of 2008, MedPAC had notified Congress that lawmakers should impose upon the traditional Medicare program all three of the MACRA pillars – P4P, ACOs, and PCMH’s. Congress took that advice. It passed several bills instructing CMS to begin linking reimbursement to “performance” on both cost and quality (the Physician Group Practice Demonstration, an early test of the ACO, is one example) and in March 2015 Congress enacted MACRA, the managed care proponent’s dream legislation, to replace the failed Sustainable Growth Rate (SGR) formula. MedPAC’s wishes had come true.
Why can’t MedPAC admit its mistakes?
As I stated in my last post on this subject, the October 4 meeting at which MedPAC staff recommended the repeal of MIPS would have been an ideal time for the commission to ask how they made the mistake of endorsing P4P at the individual physician level in their 2003, 2004 and 2005 reports to Congress. The commission’s decision at the October 4 meeting to urge the repeal of MIPS, mainly on the ground that measuring individual physician “merit” accurately is not possible, was a 180-degree reversal of their existing position on that topic. Why not admit their mistake and learn from it? Why not set aside some time to ask, “How did our predecessors endorse such a bad idea and why are we reversing it only now?”
The commission didn’t do that.
I submit the reason the commission refuses to admit error, and refuses to investigate how they commit errors, is that they understand that an honest investigation will force them to terminate two MedPAC traditions that minimize their accountability. The first of these is one I have already discussed – MedPAC’s deeply ingrained habit of endorsing “reforms” based on opinion as opposed to evidence. The second tradition I blame is MedPAC’s habit of articulating its proposed “reforms” as abstractly as possible and leaving to Congress and CMS the unpleasant task of making their evidence-free proposals work in the real world.
Both traditions encourage intellectual laziness and lack of accountability. If you don’t have to think through the details of your abstract proposals because you think your job is merely to throw out grand ideas from 80,000 feet, and if you don’t have to present evidence for your proposal or explain away evidence that contradicts your proposal, it’s easier to persuade yourself that your proposal will work. It’s also easier to say, “It’s not my fault,” when CMS is unable to implement your grand idea down on Earth where doctors and patients live.
Here is an example of both traditions at work. This example (some of which I have already quoted) is from MedPAC’s chapter in their March 2005 report in which they recommended measuring “quality” at the individual physician level:
[A]dequate risk adjustment is available…. Data needed to take these measurements can be collected without undue burden on providers or the program…. Expanded use of IT would also increase the ability to measure and reward good performance. In sum, adequate measurement tools are available to begin paying for performance…. The Congress should instruct the Secretary [of HHS, which in effect means CMS] to design a pay-for-performance program that rewards both improvement and attaining or exceeding certain benchmarks. This approach will encourage all providers to respond…. The program should be budget neutral.… Further, we would expect the Secretary to define the specific parameters of this program, such as the weights assigned to different measures and the mechanism for distributing the funds among providers. (p. 184)
You see how that works? You make statements about risk adjustment, the burden of collecting data, the role that IT could play in P4P and other conditions necessary for P4P to work that aren’t true or are grossly exaggerated and for which there is no evidence. Then, on the basis of this happy talk, you recommend P4P in the abstract, you throw out a few guiding principles (rewards and punishments should be dished out for both improvement from baseline and attaining some absolute thresholds, and the P4P scheme should be a zero-sum game so as not to increase the cost to Medicare). And then you declare that CMS can figure out the rest.
An honest investigation by MedPAC of its erroneous endorsement of the notion that physician “merit” can be measured accurately at the individual physician level would force the commission to confront these two traditions – the habit of valuing opinion over evidence, and the habit of never articulating proposals clearly and concretely and leaving it to someone else to wrestle with the details. That would be more than just embarrassing. It would force MedPAC to terminate those dysfunctional traditions. And that in turn would make it very difficult for MedPAC to continue to promote the faith-based managed care diagnosis (“overuse” due to the fee-for-service system) and the faith-based managed care solution (antidoting the FFS incentive by shifting insurance risk to providers, and micromanaging them just in case shifting insurance risk doesn’t do the trick).
Addiction to the managed care diagnosis and solution is, in the end, MedPAC’s fundamental problem. Until MedPAC’s staff and commissioners are ready to confront that addiction, they will resist recognizing the roles that disdain for evidence and love of abstraction play in maintaining that addiction. And until they explicitly condemn those traditions, they will not extricate themselves from the MACRA tar pit.
 Here are a few more details about MACRA. MACRA, enacted in 2015, will force all doctors who treat patients insured by Medicare’s traditional fee-for-service program to choose between two programs: A pay-for-performance scheme called the Merit-based Incentive Payment System (MIPS) that will allegedly measure the value (that is, both the cost and quality) of individual doctors; and a scheme called the “alternative payment model” (APM) program that relies on ACOs, “medical homes” and bundled payments. Doctors who remain in or refuse to join one or several of the APM programs will be given a score of somewhere between zero and 100 for their “total performance” as measured by hundreds of “quality” measures and a cost measure, and depending on their score they could receive up to a 9-percent increase or decrease in their Medicare payments.
 Dan Berlowitz et al., “Introduction,” Medical Care Research and Review, 2006; 63 (Supplement) 118S. Over the last decade a large body of research on P4P has emerged and it does not support P4P. As a literature review http://annals.org/aim/article/2596395/effects-pay-performance-programs-health-health-care-use-processes-care published in 2017 put it, “consistently positive associations with improved health outcomes have not been demonstrated in any setting.”
 Glenn Hackbarth, “Commentary,” Medical Research and Review, 2006; 63 (Supplement) 118S.
 The four CMS ACO programs I’m referring to are the Physician Group Practice (PGP) Demonstration (2005 to 2010), the Pioneer program (2012 to 2016), the Medicare Shared Savings Program (a permanent program begun in 2012), and the NextGen program (begun in 2016.) Here is the abysmal track record of these four programs (the figures take only Medicare’s claims costs into account; they do not take into account the costs to ACOs of attempting to reduce Medicare’s claims costs nor the cost to CMS of running these complex programs): The PGP demo raised costs by 1.3 percent; the Pioneer and NextGen programs cut costs by a few tenths of a percent; and the MSSP program, by far the largest of the four, has increased costs by a few tenths of a percent. The latest report https://oig.hhs.gov/oei/reports/oei-02-15-00450.asp from the OIG, its deceptive title notwithstanding, indicates that the MSSP program raised Medicare’s costs by $300 million over the three years 2013 to 2015. CMS’s 2016 data for its NextGen ACOs indicate that they cut Medicare’s costs by seven-tenths of a percent that year (my calculations based on CMS data). For specific savings rates on CMS’s Pioneer and MSSP programs by year, see my comment here http://thehealthcareblog.com/blog/2017/01/28/simulated-acos-vs-real-world-acos/. MedPAC staff estimate the cost to ACOs of starting and running ACOs is 1 to 2 percent of their Medicare spending. This estimate seems low to me, but low as it is it swamps the tiny savings achieved by the Pioneer and NextGen programs and substantially raises the cost of the MSSP program.