What a Primary-Care Focused Advanced APM Would Look Like

The Medicare Access and CHIP Reauthorization Act's (MACRA's) Advanced Alternative Payment Models (AAPMs) offer incentives to physicians who provide high-quality and cost-efficient care as it applies to a specific clinical condition, care episode, or population. Examples include the Comprehensive Primary Care (CPC) Plus and the Oncology Care Model.

MACRA also calls for physicians to design their own AAPMs based on how care is actually delivered. In December 2017, the Physician-Focused Payment Model Technical Advisory Committee (PTAC), which provides recommendations to the Health and Human Services (HHS) Secretary, voted to recommend that HHS test the Advanced Primary Care APM developed by the American Academy of Family Physicians, which has more than 84,000 actively practicing physicians as members. The AAFP expects that primary-care physicians currently not in CPC Plus regions would be most interested in participating in the APC-APM. Any fee-for-service Medicare beneficiary not attributed to another APM could receive care under the APC-APM, so the potential impact is huge.

Physicians Practice recently asked three AAFP representatives to talk about the payment model they have proposed and the PTAC’s response. We spoke with Michael Munger, MD, AAFP president; Amy Mullins, MD, medical director for quality improvement; and Kent Moore, senior strategist for physician payment.

Physicians Practice: Dr. Munger, when you addressed the PTAC committee evaluating your proposal, you said that the undervaluation of primary-care services and fragmentation driven by our current payment system are well understood. You even noted that payment experts on the PTAC panel have pointed out that building APMs on flawed physician fee schedules would simply perpetuate current inequities in payment. Why did you think that was an important point to stress?

Munger: In family medicine, we are primarily getting payment based on who is right in front of us and how much we can do to them. Furthermore, if you look at the relative valuation of an E&M [evaluation and management] code vs. some of the procedural codes, that is where you start to see inequities. For instance, look at the value placed on spending 25 minutes seeing an established patient about three or four chronic conditions vs. doing a colonoscopy in the same length of time. The procedural code for the colonoscopy is worth almost three times more than the E/M visit focused on chronic care management. Also, many aspects of family medicine are not done face-to-face. I can’t drop a code around coordinating care or managing conditions between visits unless a certain threshold of time and other requirements are met. To build anything on our current system is really going to miss out on the comprehensiveness of what we deliver in family medicine.

PP: MACRA established two payment pathways. Physicians can participate via the Merit-based Incentive Payment System (MIPS) or in AAPMs. Why was it important that the MACRA statute include the capability for physicians to propose their own tailored payment models?

Mullins: It is important that we realize there are not enough AAPMs available for physicians to participate in. As the law is structured, MIPS is a complicated program. We believe the intent of the law is to move physicians and clinicians into the AAPM track, but the onramp to the AAPM track is narrow and steep right now. There needs to be an AAPM that will be available for the majority of primary-care physicians to participate in as an alternative to MIPS.

PP: Could you talk about the design of the Advanced Primary Care Alternative Payment Model and its goals? As I understand it, it would move physicians out of fee for service and into a prospective payment, coupled with performance-based incentives that hold physicians accountable for quality and costs.  Could you talk about how it is similar to or different from CPC Plus?

Munger: The scale is much larger. I myself participate in CPC Plus, but it is limited to 18 geographic regions now. This is building on that, but it is different because it really does recognize the importance of a prospective, risk-adjusted population-based payment. You have the revenue streams recognizing what we are doing in primary care and helping with our infrastructure cost, yet still preserving some fee for service for those procedural things outside of a prospective global payment.

Mullins: We see this as a continuum from CPC Plus - almost the next step up from that model. Also, CPC Plus has a lot of reporting requirements. We are hoping the model we have designed is somewhat less burdensome to the participants.

PP: Is one benefit of your APC-APM that it can also be adapted for use with other payers and populations?

Mullins: To be a successful AAPM, you have got to have more than CMS on board. You really need all the payers to participate and treat this program the same in their contracts, because when you are seeing patients in your office, the last thing you want to do is look to see what kind of insurance they have and try to figure out if they are a participant in this program. You want to treat your patient as a patient, not as someone who carries a particular kind of insurance. It is difficult to achieve wide adoption of a payment model if it is too complex and has a high reporting burden. This is magnified if physicians are not getting paid for doing this work on all of their patients. You need all the payers to participate.

PP: Have you engaged with other payers to get their feedback on participating in this model?

Mullins: We have been talking with the major payers about this model. We have let them know this is something that has been presented to the PTAC and recommended for testing. We are hopeful they will be willing to participate.

PP: Is there a timeline for when the Secretary is expected to act on the PTAC recommendation to test the model?

Moore: The short answer is no, there is no timeline built into the regulation or the statute. We are expecting the PTAC to issue its formal, written report to the Secretary sometime soon. That is the only milestone we have at this point.

PP: The PTAC had a few critiques of your model. At the PTAC hearing in December, one of PTAC’s members, the Brookings Institution’s Kavita Patel, MD, said the Preliminary Review Team had concerns that the proposed model would increase payments for primary-care practices without sufficient assurances that there would be proportionate savings for payers such as CMS, and that there could be real challenges measuring those savings. How did you respond to that critique?

Moore: Measurement would involve evaluating the extent to which the model does in fact save Medicare and other payers money, as we hypothesize it would. Part of the reason we advocate rolling this out nationally is to provide a greater opportunity to have comparison groups in terms of evaluation. Also, our model envisions prospective attribution of patients so that both the payers and practices will know in advance which patients are in the program. Therefore, it should be relatively easy to identify what those patients have cost historically and compare that to what they cost going forward.

PP: Dr. Patel also mentioned that under the model, if a practice underperforms, it would have to pay some or all of the incentive back, and this puts the government in the position of performing collections on money already paid out and puts participants with weak balance sheets at significant financial risk. What is your response to that?

Munger: CMS and CMMI [the Center for Medicare and Medicaid Innovation] are already doing that. With CPC Plus, I receive my performance-based payment at the first of the year, and the accounting will occur in the first quarter of the following year, and I will have to pay back part or all of it if I don’t hit the performance goals.

PP: Did the PTAC also say it believed more work needed to be done on quality measures to demonstrate or incentivize improved care coordination?

Mullins: We did have a long conversation with them about the quality measures. Our goal was to use the measures that the Core Quality Measures Collaborative developed [through] public and private payers as well as physicians in order to have that harmonized, parsimonious set of measures that multiple payers had already agreed upon. Because we want multiple payers to participate in this AAPM, we wanted to use something they were already familiar with. The other reason we chose those measures is because they are in MIPS, and to be in an AAPM, you have to base payment on quality measures that are comparable to measures in MIPS. We realize this may not be the ideal set of measures. It is a set that is going to be updated, and it is necessarily in a state of flux, because measurement is always changing.

PP: Dr. Munger, in your comments during the PTAC meeting, you mentioned that the model allows for addressing social determinants of health, which affect health outcomes and cost, and facilitate a true longitudinal assessment of patient needs. Why did you think it was important to stress that aspect?

Munger: You cannot do effective population health management unless you are addressing social determinants of health. I keep in mind that the most important risk factor for hospital readmission is ZIP code. The ability to risk-stratify your population to look for individuals who have food insecurity or transportation needs or who don’t feel safe in their neighborhood lets you target your resources better. It also allows patients to better participate in their own chronic disease management.