In my last post here, I reported that the Medicare Payment Advisory Commission (MedPAC) believes Medicare ACOs and “medical homes” are unlikely to qualify as “alternative payment models” under MACRA (the Medicare Access and CHIP Reauthorization Act).
In this article I review the evidence supporting MedPAC’s pessimistic assessment of “patient-centered medical homes” (PCMHs). I will review evidence that PCMHs are not cutting Medicare costs but are probably inflicting financial and emotional stress on many PCMH clinics. In a future post I will review the evidence on Medicare ACOs. Down the road I hope to comment on bundled payments. PCMHs, ACOs, and bundled payments are the three templates available to CMS to fashion APMs.
“Medical homes” may be raising Medicare’s costs
Within five years after the PCMH fad appeared, CMS had begun three demonstrations to test the ability of PCMHs to lower Medicare costs while simultaneously improving quality. The first of these demos, the FQHC Advanced Primary Care Practice Demonstration , ended in October 2014. The other two, the Comprehensive Primary Care (CPC) Initiative and the Multi-payer Advanced Primary Care Practice (MAPCP) Demonstration, will end this year. 
CMS has released evaluations of the first two years of the FQHC and CPC demos , and an evaluation of the first year of the MAPCP demo . CMS has delayed the second-year evaluation of the MAPCP demo. 
The findings presented in these evaluations are not good news for PCMH supporters. In a recent report , the Kaiser Family Foundation (KFF) offered this summary of the results of the three PCMH demos: “Among the office-based multi-payer models (MAPCP and CPC) and the FQHC/APCP model, little to no savings have been generated after accounting for the outgoing Medicare expenditures in care management fees” (see Table 2 of the report).
KFF could have been a tad harsher in their judgment: The evaluations suggest the three demos together are raising Medicare’s costs when Medicare’s payments to PCMHs are taken into account. The evaluations of the CPC and MAPCP demos have reported no statistically significant change in Medicare spending, while the evaluation of the FQHC demo reports that FQHC “homes” are raising Medicare costs by a statistically significant amount. Here is what the RAND Corporation said in its evaluation of the FQHC demo in its July 2015 second-year evaluation: “As of the end of the demonstration’s ninth quarter, we know that costs for demonstration sites are higher than for comparison sites…. These cost findings … suggest that it is unlikely that overall costs associated with the demonstration at its completion will be lower for demonstration FQHCs than for comparison FQHCs.” (p. xvii)
The last bit of bad news for PCMH proponents is that the PCMH demos have failed to improve quality of care much or at all. According to the KFF report, “Minimal differences in quality were found between CPC practices and comparison primary care practices, as well as between FQHCs participating in the APCP model compared with FQHCs not participating.” Differences in quality between PCMHs and non-PCMHs have not been reported yet for the MAPCP demo.
To sum up, the evaluations of the three PCMH Medicare demos indicate that the “medical home” cannot serve as a viable MACRA APM because there is no evidence yet that it can lower Medicare’s costs without harming patients.
PCMHs inflict financial stress on doctors
Whether PCMHs can serve as a viable APM will depend not just on whether they save Medicare money, but whether PCMHs raise or lower the revenues of the doctors who “transform” their clinics into PCMHs. If the entities CMS certifies as APMs are less attractive to doctors than the Merit-based Incentive Payment System (MIPS) (the other “track” or compartment within MACRA), doctors are likely to stay out of APMs and hunker down in the MIPS program, regardless of whether the APM in question cuts Medicare’s costs.
We need to know, then, how much it costs doctors to set up and run PCMHs.
So, let me ask you: When MedPAC, the Patient-Centered Primary Care Collaborative, and other PCMH proponents endorsed PCMHs, how much did they think it would cost doctors to set up and run PCMHs? That’s a joke – a trick question. MedPAC and other “medical home” advocates didn’t have the slightest idea when they endorsed “homes” what it would cost clinics to “transform” into PCMHs. PCMH proponents assumed the costs of the PCMH were non-existent or so trivial they could be ignored.
If we cannot find even crude estimates of PCMH-related costs in the manifestoes of “home” proponents, where might we turn for useful information? You might think the evaluations of the three CMS “home” demos would help, but you’d be wrong. The contractors CMS hired to evaluate the CPC and MAPCP demos have made no attempt to determine what costs PCMH clinics incur and whether the revenues clinics receive from CMS offset or exceed those costs. Only RAND, in its evaluation of the FQHC demo, has made that effort, and they struck out. There appear to be two reasons RAND came up empty. The first is the vacuous definition of “medical home” (I will discuss this issue in a future post). The second is that PCMH clinics are so overwhelmed by the busywork CMS imposes on them they don’t have time to determine what the PCMH adventure is costing them.
To illustrate the latter problem, let me quote from RAND’s second-year evaluation of the FQHC demo. At page 99 of the report, RAND poses two questions you’d think would have been answered long ago by PCMH advocates. The first question was, “How much does it cost to implement and sustain the various features of an APC Practice?” (In RAND’s report, “APC practice” [APCP] is synonymous with “PCMH.” That’s because, for purposes of the FQHC demo, CMS defined an APCP to be a PCMH certified as a Level 3 “home” by the National Committee for Quality Assurance [NCQA]).
Here’s the second question RAND posed: “Is $6 per beneficiary per month (PBPM) sufficient to offset those costs? Is there a return on investment and, if so, how quickly are costs recouped?” (Six dollars PBPM is what CMS pays the participating FQHCs for each Medicare beneficiary assigned to FQHC clinics by CMS’s strange “attribution” method.)
Here is how RAND answered the first question:
[T]his research question refers to costs realized by the FQHC sites to implement and sustain an APCP. To fully answer this question, it would be necessary to … collect detailed data on resources utilized to implement and sustain an APCP…. Such data would include the time that clinic staff spend on the additional patient-level activities that are required to be recognized as an APCP, facilities or technology acquired or maintained specifically because of their APC activities, administrative costs required to shift to and sustain an APCP, and any other materials used to deliver care through an APCP….
Given the burden faced by FQHCs to complete achievement of NCQA Level 3 within a short period of time – and, furthermore, to successfully transform their practices – we have not identified any FQHC sites that have been willing to allow this type of data evaluation. While we will continue to gather data from site leaders about their perception of the financial incentives associated with the demonstration, the evaluation team does not anticipate being able to have adequate data to meaningfully address this question. (p. 99)
Here is how RAND answered the second question: “Because of the limitations noted above in the discussion of [the first question], the evaluation team is not likely to be able to address these questions quantitatively.” The only definite statement RAND was willing to make was that not a single FQHC employee would say the $6 was enough. “None committed to a statement that the payment was adequate for the task,” RAND reported. (p. 126)
Commonsense and a tiny body of peer-reviewed research tells us PCMHs require clinics to incur substantial costs. Last fall, MacGill et al. http://www.annfammed.org/
Do the “care management fees” CMS is paying PCMHs exceed $105,000 per physician per year, or $150,000 per physician per year, or whatever the true number is? Will those fees plus “shared savings” bonuses CMS will pay to the PCMHs in the CPI demo starting in 2015 exceed that number? Nobody knows. Anecdotal evidence http://www.aafp.org/news/2012-
PCMHs inflict emotional stress on doctors as well
The evaluation of one of the three PCMH demos contains evidence that setting up and running PCHMs is not merely expensive but also stressful, largely because of the numerous administrative chores CMS imposes on participating clinics. Here is how RAND described the stress CMS and NCQA impose on FQHC clinics:
Although FQHCs were enthusiastic about the hoped-for transformation to APCPs, many were overwhelmed with the need to change so much so fast while simultaneously completing complex forms required for NCQA recognition applications and still caring for complex patients…. [F]or some, the time pressure compromised efficiency. Ongoing requests for sites to participate in webinars, calls, surveys, and site visits occurred concomitant with the need to respond to six domains of transformation heralded by NCQA…. Some sites described teams that were exhausted and frustrated, others described frenzy, apathy, or a sense of having gone astray. (p. 45)
MedPAC perceives these problems
Although MedPAC was one of the earliest and most influential institutions to jump on the “medical home” bandwagon (they did so in 2008 without offering a shred of evidence for their decision), MedPAC staff and commissioners have expressed disappointment in PCMHs on several occasions in the last two years. At the March 6, 2014 meeting, more than half of the 17 commissioners criticized the PCMH fad.
Then-chairman Glenn Hackbarth began the criticism that day by calling the PCMH requirements “gold-plated.” “I’ve become increasingly concerned about the medical home demonstrations on a number of different grounds,” he began. “I am a little bit worried that the medical home model has … become gold-plated … in order to meet all of the NCQA requirements…. [A] lot of bells and whistles … have been added. … [M]y impression is that not all of them have really been validated as adding value, but they add cost….” (pp 252-253 of the transcript ). A few minutes later Hackbarth identified “electronic medical records and … 24-hour coverage” as examples of the “long list of [PCMH] requirements” he questions.
At least eight other commissioners added their own criticisms that day. For example, Scott Armstrong (CEO of Group Health Cooperative) said that requiring doctors to give their patients e-mail access would greatly increase “virtual visits” and that many emails from patients are “useless.” (pp. 275-276) Jon Christianson noted that large hospitals that are buying up clinics may not pass on “care coordination fees” to PCMH clinics. (p. 279) Michael Chernew referred to “home” requirements as a “hassle.” (p. 288)
One month later, at its April 2014 meeting, one of MedPAC’s staff testified that the evidence on PCMHs is “mixed,” and another stated that the VA’s “medical home” experiment has failed to improve quality (see pp. 74-75 of the transcript of that meeting http://www.medpac.gov/
I have described three reasons why MedPAC’s pessimistic assessment of the PCMH as an APM prototype is correct: Medicare’s PCMH demos are showing that PCMHs cannot cut Medicare costs and are having little impact on quality; PCMHs are expensive, and are probably costing doctors more than they are getting back in the form of “care management fees” and “shared savings”; and PCMHs can add to the emotional stress that is already at epidemic levels among physicians.
For all these reasons, we may conclude that the PCMH cannot play the role of an APM under MACRA. But that doesn’t mean CMS won’t force the PCMH (or something quite like it with a new and equally saccharine name) to play that role. CMS is not acting entirely rationally these days. It appears CMS intends to cram the square MACRA peg through the round hole of reality regardless of the consequences.
 Readers should be aware the law and CMS make a distinction between “initiatives” and “demonstrations.” I use only “demonstration” here.
 The first-year evaluation of the MAPC demo was released in January 2015 along with the first-year evaluation of the CPC demo. Although the second-year evaluation of the CPC demo was released in April 2016, the second-year evaluation of the MAPC demo has still not been released. On April 25, I sent an email to RTI International, the author of the first-year evaluation, asking why the second-year evaluation has not been released. As of April 29, I have not received an answer.
 MacGill et al. stated that they did not include the cost of the following activities in their estimate of $105,000 per doctor per year: “[M]aintaining EHRs …, providing culturally and linguistically appropriate services, collecting patient information, collecting and documenting clinical data, conducting comprehensive health assessments, or using electronic prescribing….” (p. 430). They also reported, “We did not evaluate the startup costs of PCMH implementation, its revenue impact, or its associated opportunity costs.” (p. 434)