By KIP SULLIVAN

MACRA (the Medicare Access and CHIP Reauthorization Act) is a mess. It is extremely difficult to comprehend, it is based on assumptions that defy commonsense and research, and it may raise costs.

The Medicare Payment Advisory Commission (MedPAC) would never say what I have just said, but MedPAC definitely understands MACRA’s defects. The transcripts of MedPAC’s October 8, 2015 and January 15, 2016 meetings indicate that members and staff perceive daunting impediments to the implementation of MACRA. But those transcripts also suggest that MedPAC won’t tell Congress to rewrite or repeal MACRA. Rather, the evidence suggests MedPAC will mince words. It appears MedPAC will send CMS and Congress a few wishes dressed up as “principles” and wait for MACRA’s inevitable failure before offering more useful advice.

Before I attempt to explain MACRA, let me first convey to you MACRA’s mind-numbing complexity by quoting four commissioners. Each statement below is followed by the last name of the commissioner who made it, the date the statement was made, and the page number of the transcript where the statement appears.

“[T]he complexities involved are just absolutely enormous and mind-boggling.” (Hall, October 8, 2015, 158-159)

“I think … Congress has given us a very tricky set of semantics to work with…. What does ‘in’ mean? [Laughter]….Now, is that in our scope of discussion this morning – what these words [in MACRA] should mean?” (Nerenz, January 15, 2016, 86, 89-90)

“[W]hen you go through the law, in some parts of it it’s very hard to figure out how they fit together….” (Chairman Crosson, January 15, 2016, 90)

“I understand this is very complicated, but I think one of our first principles should be to try to simplify this, because it’s very complicated for providers to try to understand what path to take…. I understand that’s probably heroic….I think there’s a lot of aversion because folks don’t understand it….” (Thomas, January 15, 2016, 144)

Even the staff, who must explain MACRA to the commissioners, find themselves at a loss for words. Mark Miller said at the January 2016 meeting, “I don’t feel like I have a real strong handle on all these definitional issues….You know, we’ve had huge internal conversations. I’m surprised no one has resigned yet.” (94)

A bird’s-eye view of a Rube Goldberg device

To comprehend MACRA, it helps to understand the motives of its chief authors, including Representatives John Boehner and Nancy Pelosi. Like other managed care proponents, Boehner et al. thought the fee-for-service (FFS) method of paying for Medicare services was raising Medicare’s costs. Having made that evidence-free diagnosis, they decided the solution is (a) to turn the FFS incentive upside down by shifting insurance risk onto the backs of doctors and hospitals and, to minimize the possibility that risk-shifting will induce providers to short-change patients, (b) to measure the quality of a tiny fraction of all medical services delivered to patients.

Congress could have implemented this “solution” by simply instructing CMS to stop using pure FFS and to use instead pure capitation or some hybrid of capitation and FFS for all providers. Such a bill would have raised serious questions about its ability to lower costs without harming patients, but at least the questions would have been relatively easy to understand.

But instead of stating forthrightly they wanted to terminate FFS Medicare and replace it with a system resembling a giant federal HMO, Congress ordered CMS to implement a sprawling Rube Goldberg device with two compartments: One, known as the Merit-based Incentive Payment System (MIPS), would irritate doctors who stayed in FFS Medicare so much they would look for an alternative; and the second, the Alternative Payment Model (APM) program, would provide them with an alternative that would allegedly improve quality and lower costs.

But MACRA is so devoid of details about each compartment that it is impossible to predict anything. We know Boehner et al. wanted the first compartment, MIPS, to inflict large penalties and rewards, but we don’t know whether the MIPS compartment will be sufficiently irritating to drive its occupants out. Even if we knew how irritating MIPS will be, we wouldn’t be able to compare it to the APM program’s power to irritate because we don’t know what the APM entities will look like. In fact, at this date we can’t even be sure APMs will exist.

For all those reasons and more, we also don’t know whether either compartment will save money.

MIPS: Measuring the unmeasurable

The MIPS program was designed for doctors who can’t or won’t join the ACO-medical-home bandwagon. MIPS inflicts huge rewards and penalties (up to 9 percent of the average annual Part B payout per physician) based on a single “value” score (a score derived by the capricious jamming of crude quality and cost measures into one number) for each doctor. As commissioner Gradison put it, MIPS was designed for “the laggards, the people who stand for the status quo,” the knuckle-draggers who just won’t get with the Managed Care 2.0 program. (January 15, p. 133)

Evaluating how the laggards will view MIPS vis a vis the APM program is impossible because we have no idea how CMS will enforce the requirement that CMS produce a single score representing both the cost and quality of services provided by individual physicians. With the exception of a few services, measuring cost and quality accurately at the individual doctor level is not possible. There are two intractable impediments: Determining which doctor patients “belong” to (the attribution problem); and adjusting measures of physician cost and quality for factors outside of physician control such as patient health status (the risk-adjustment problem). The attribution and risk-adjustment methods used today are crude even for large groups; with the possible exception of a few of the simplest process measures, they are worthless at the individual physician level.

MedPAC is well aware of at least the risk-adjustment problem. Here is how MedPAC staffer Kate Bloniarz expressed her “concerns” about MIPS to the commission at its October 2015 meeting:

First is the challenge posed by assessing clinician performance at the individual level. The MIPS … is designed to produce an individual-level payment adjustment. But many quality and resource use measures are not reliable at the individual clinician level, and it is a particular challenge for outcomes measures. Based on CMS’s experience to date with individual-level payment adjustments, most clinicians will likely look average, and the Medicare program will only be able to reliably identify persistent outliers. (pp. 93-94 of the October 8 transcript)

At the January 2016 meeting Bloniarz repeated this criticism. She told the commission that CMS’s Physician Quality Reporting System (PQRS), which MIPS must use, could not distinguish 80 percent of the physicians from one another. (p. 74, January 15 transcript, and slide 10 of presentation)

If CMS cannot detect differences at the individual physician level for 80 percent or more of the doctors who serve Medicare patients, it seems likely that the pay-for-performance scheme Boehner et al. had in mind for the laggards in MIPS won’t have any effect on the vast majority of them. Which calls into question the ability of MIPs to “reform” doctors – to get them to “coordinate care” and do all those other wonderful things ACOs and “medical homes” were supposed to make them do. [1]

Ms. Bloniarz also expressed her concern that the reporting requirements for doctors under both MIPS and the APM program are “likely to be overly complex and further burden both providers and CMS….” (p. 94)

Sand castles on top of sand castles: The APM program

The second compartment in MACRA – the APM program – is difficult to describe for these reasons:

• MACRA does not describe the entities that will qualify as APMs;

• MACRA sets standards for these entities that cannot be met by the vast majority of conceivable entities, including today’s most faddish APMs – ACOs and “medical homes”;

• MACRA says APMs must expose doctors to “risk above a nominal level,” but doesn’t say what “nominal” means; and

• MACRA says doctors who join APMs will earn a 5 percent bonus on revenues received through APMs, but the law fails to define “revenue” (options include revenue from Part A, Part B, Part D, all or some of those parts, or revenue received for particular services).

Here is how Ms. Bloniarz explained the problem at the commission’s October 2015 meeting:

The first question is: What is an APM? The pool of APMs are all models under the Center for Medicare and Medicaid Innovation (except for innovation awards), models tested under the pre-existing Medicare demonstration authority, a demonstration required by law, and the Medicare shared savings program.

 Then there is statutory language that further winnows down the number of APMs that can qualify clinicians for the incentive payment. They must meet three criteria

…: they must use certified eHR technology; they must have comparable quality measures to MIPS; and they must either bear financial risk above a nominal amount or be a medical home that has been certified for expansion by the Office of the Actuary. And I’ll note here that that certification has not occurred.

A key takeaway is that not all APMs will be eligible APMs for which clinicians can receive the incentive payment. They will need to meet the criteria set out in the law, and presently, very few models are likely to do so. (pp. 94-95)

I suspect the reason Congress did not populate the APM compartment with clearly defined entities is that Boehner et al. knew full well that the first iteration of “value-based” entities – the ACOs and “medical homes” authorized by the Affordable Care Act – have not panned out. We know that MedPAC understands that. In its reports to Congress, and in statements by staff and members, MedPAC has clearly indicated they understand that ACOs and “homes” are saving little or no money and are having at best minor and mixed effects on quality.

It is fair to say, then, that MedPAC understands that Congress has essentially instructed CMS to build a new layer of undefined, unproven APMs on top of the existing layer of poorly defined and unproven ACOs and “homes,” in other words, a second layer of sand castles over an existing layer of sand castles.

MedPAC’s reaction to the MACRA mess

If you were a member of MedPAC and you had to advise CMS and Congress on the implementation of the MIPS and APM programs, what would you do? You know that CMS can’t measure “value” in MIPS anywhere near accurately enough to provide useful feedback to the laggards who remain in MIPS. And you know CMS has to populate the APM compartment by 2017 even though virtually no examples of the fabled APMs exist, and you know the first-generation APMs that might serve as templates for CMS’s are not working.

I suspect most ordinary people would tell Congress that it must stop writing managed care fantasies into law and hoping CMS can make them work. I suspect most people would tell Congress to go back to the drawing board and propose legislation that describes flesh-and-blood entities (not abstract managed care fantasies) that actually have a track record documented by empirical research (not happy talk). The very least MedPAC should do is define clearly the crippling defects in MACRA and tell Congress to postpone MACRA’s implementation to give CMS and MedPAC time to develop more informed recommendations.

But it appears MedPAC has no intention of doing that. Not one of the 17 commissioners suggested at the October or January meetings that MACRA should be repealed, amended, or postponed. It appears, rather, that MedPAC will send CMS and Congress a list of abstract wishes for the APM program dressed up as “principles.”

When the commissioners arrived for their January 2016 meeting, the staff (presumably with Chairman Crosson’s blessing) asked them to discuss 12 “principles” that would guide CMS in its doomed effort to make sense of the APM compartment (see slides 12 through 14 here) The principles said nothing about the MIPS program.

The 12 APM principles will look very familiar to anyone who has followed the debate about ACOs – they look just like the elements in the flabby “definition” of ACO. [3] These principles would be worthless if they were applied even to the existing layer of sand castles; they tell us nothing about why CMS’s current crop of ACOs and “homes” are performing so poorly. It is extremely difficult to comprehend, therefore, why MedPAC is wasting its time formulating these vague principles for the second layer of sand castles – the APMs that don’t exist yet.

Congress and CMS don’t need to hear any more abstract and wishful remarks about MACRA. They need to hear useful feedback. MedPAC should tell Congress MACRA is an unworkable mess and must be repealed or amended.

[1] In addition to the CMS’s inability to measure quality at the individual physician level, Bloniarz should also have listed MIPS’ bizarre attribution scheme as another mind-bending impediment to the implementation of MACRA. Unlike ACOs and “medical homes,” where insurers “attribute” patients to doctors based on how often patients visited doctors in the past, MIPS requires that doctors “attribute” themselves to patients by selecting from several possible descriptions of their relationship to each patient, such as “lead” physician and “supportive” physician, and entering a code for this relationship on each claim form.

[2] Chairman Crosson, for example, has observed that MACRA’s APM compartment “is not designed to essentially work well with what exists right now because what exists right now isn’t getting us necessarily to where you want to be.…” (October 8, 2015, p. 119) Commissioner Naylor has questioned why MACRA is built on “programs or models that have not yet been proven or, in some cases, have been proven not to be working.” (January 15, 2016, p. 150) Other commissioners have endorsed the radical notion (radical by the shoot-first-aim-later standards of the managed care movement) that APMs should be subjected to research before CMS certifies them.

[3] For example, principle 1 says CMS should authorize only those APMs that have been shown to be “successful [at] controlling cost, improving quality, or both,” principle 6 says the entity “must assume risk,” and principle 8 says APMs must have “sufficient number of beneficiaries to detect changes in spending or quality.”