Getting ACOld shoulder, or how a carrot becomes a stick: ACO Proposed Rule could negatively affect non-ACO health care providers

As described in a previous post, PPACA section 3022 (a.k.a. Public Law 111–148 § 3022, or § 1899 of the Social Security Act, to be codified at 42 U.S.C. § 1395jjj) authorizes the formation of Accountable Care Organizations (ACOs) to implement a Medicare “shared-savings program” (MSSP).  At the risk of contributing to all the noise around ACOs, below are some observations about a “proposed rule” issued by the Centers for Medicare and Medicaid Services (CMS) on April 7, 2011, and how the rule as currently proposed could have some negative consequences for entities that do not participate in an ACO, particularly when this proposed rule is considered in conjunction with other pre-regulatory activity. 


When reading below, please remember that the bulk of this post relates to a proposed rule, and that if you are truly concerned about possible consequences, it is absolutely worth submitting comments on the proposed rule. 

So, what's an ACO?

As proposed by CMS, ACOs would generally be groups of health care providers that are centered around primary care physicians, and Medicare beneficiaries would be “assigned” to them to allow the government to monitor health care quality and cost. Note that this post discusses only the PPACA Medicare ACO concept, and does not address possible future private sector ACOs (which could nevertheless take strong cues from the Medicare ACO concept). 

The MSSP will essentially pay ACOs to: 1) Reduce the overall cost of care provided to Medicare beneficiaries; 2) maintain or improve the quality of that care; and 3) provide quality data to Medicare. MSSP payments could in some cases be significant: Up to 10 percent of what the reimbursement would be for a similar group of patients not assigned to an ACO. However, CMS’s current proposed rule could significantly affect an organization even if it does not participate in an ACO. How? Read on....

The ACO/MSSP program must by law start on or before January 1, 2012. Various government agencies responsible for regulating ACO activities, including CMS, have issued proposed regulations or official requests for comment on how to implement the ACO/MSSP program, as described in more detail below. It is likely (although not explicitly stated by statute or the proposed rule) that quality data collected through the ACO/MSSP program may be used to influence future Medicare payment policy, and the program could to some degree serve as a model for private payer reimbursement policy changes. In fact, the FTC and Department of Justice request for comment on antitrust issues (see below) suggests this latter possibility.

This post addresses four main issues below: Current regulatory activity around ACOs, who can form an ACO, how Medicare beneficiaries are assigned to it, and how an ACO gets paid (along with potential ripple effects). 

Why all the exposition? Well, without it, the main point here—i.e., that if the proposed rule gets implemented as currently proposed, ACOs could become some sort of “carrot stick” to entice Medicare-enrolled health care providers to join ACOs, then essentially force all the others who haven't to join them—wouldn’t make much sense (but if you’re already familiar with the proposed rule, please skip ahead to “The Stick” below). So, let’s dive in.

Current Regulatory Activity

Regulators from the Department of Justice (DOJ), Federal Trade Commission (FTC), and Internal Revenue Service (IRS) have issued “requests for comment” on possibly modifying enforcement of antitrust laws for certain Medicare ACO operations and collecting input on potential implications for tax-exempt organizations. Substantively responding to these requests for comment could result in improved clarity about the legality of operating ACOs, which could contribute to making their operation more efficient in the future. Written responses to the DOJ, FTC, and IRS requests are due on or before May 31, 2011. See http://www.gpo.gov/fdsys/search/pagedetails.action?granuleId=2011-9466&packageId=FR-2011-04-19&acCode=FR for the FTC and DOJ request for comment; see http://www.irs.gov/irb/2011-16_IRB/ar07.html for the IRS request for comment.

CMS and the HHS Office of Inspector General (OIG) are also soliciting comments on how they should execute their statutory authority to waive application of various civil and criminal laws that would otherwise apply to—and thus, make illegal—many ACO operations. See 76 Fed. Reg. 19655 (April 7, 2011); http://www.gpo.gov/fdsys/search/pagedetails.action?granuleId=2011-7884&packageId=FR-2011-04-07&acCode=FR. Comments on those issues are due by June 6, 2011 at 5 p.m. EDT.

CMS has also issued a proposed rule to implement the MSSP on or before January 1, 2012. See 76 Fed. Reg. 19528 (April 7, 2011), available at http://www.gpo.gov/fdsys/search/pagedetails.action?granuleId=2011-7880&packageId=FR-2011-04-07&acCode=FR. Comments on the proposed rule are due by June 6, 2011 at 5 p.m. EDT. 

The bulk of comments below address CMS’s proposed ACO/MSSP rule. 

Who can form an ACO, and how would it be organized?

As described in PPACA and CMS’s proposed rule, ACOs would be organizations consisting of at least primary care physicians and other practitioners (the proposed rule would limit these other practitioners to physician assistants, nurse practitioners, and clinical nurse specialists), and may be integrated health systems, hospitals that employ primary care physicians and other practitioners, physician and practitioner group practices or networks, joint ventures between hospitals and physicians and other practitioners, or other arrangements authorized by CMS. See Social Security Act § 1899(b)(2) (to be codified at 42 U.S.C. § 1395jjj(b)(2)).

The currently proposed structure of an ACO is relatively flexible, but ACOs would need to meet a raft of legal, organizational, financial, administrative, and record-keeping requirements that with few exceptions would likely make them more complex than existing practices. See Proposed 42 C.F.R. § 425.24, 76 Fed. Reg. at 19654. As provided by PPACA, ACOs will be required to enter into a three-year agreement with CMS to operate as ACOs.

CMS’s proposal to “assign” beneficiaries to an ACO

Unlike an HMO, under an ACO, beneficiaries would be free to use any Medicare provider they choose, and beneficiaries would not enroll an in ACO: CMS would assign them to one. CMS is proposing a relatively complex process to assign beneficiaries to ACOs, but it is essentially based on where each beneficiary gets the “plurality” of his or her primary care services as measured by the aggregate Medicare allowed amounts paid on behalf of that beneficiary for certain HCPCS codes specified in Proposed 42 C.F.R. § 425.4. Aside from the hopefully improved quality of care, beneficiaries would likely not even notice that they have been assigned to an ACO until they have been told so. Similarly, under the current proposed rule, the only way for a beneficiary to get out of an ACO is to get a “plurality” of his or her “primary care” services from a physician who does not belong to an ACO.

By statute, ACOs need to serve at least 5,000 beneficiaries, and CMS is proposing penalties if ACOs serve fewer than 5,000 beneficiaries during a year. See Proposed 42 C.F.R. § 425.6, 76 Fed. Reg. at 19645. ACOs would also be required to notify beneficiaries that the ACO’s “providers” (generally, Part A health care providers) and “suppliers” (generally, Part B health care providers) are participating in an ACO.

Calculating “shared savings” payments


The Carrot

To calculate MSSP payments, CMS is proposing to look at Medicare expenditures for each ACO’s patients at the beginning of an ACO agreement and adjust that amount every year to account for Medicare expenditure inflation; this is called the “benchmark”. The proposed rule suggests this is a per-capita figure, but for simplicity, we'll use it in the aggregate sense here, because that is how CMS will effectively use it to calculate MSSP payments and remittances to CMS. The benchmark is designed to project what the average expenditures would be for a similar patient group outside of an ACO. See Proposed 42 C.F.R. § 425.7, 76 Fed. Reg. at 19645–6 (there are also some other adjustments and calculations, but they’re too complicated to be relevant here). 


The proposed rule provides that if the total cost of care—for all services, not just primary care—for an ACO’s patients falls below the benchmark in a given year of the ACO’s agreement with CMS, then the ACO may be eligible to receive a “shared savings payment” at the end of that year. If the total cost of care is above benchmark, then (with several exceptions) the ACO may need to pay CMS the difference between the benchmark and the cost of care (a “loss”). 

Generally, CMS is proposing two “Tracks” to participate in the MSSP as an ACO. “Track 1” would allow an ACO to not assume any risk for increased costs in years 1 or 2 of an ACO agreement (called the “one-sided model” in the proposed rule), but would limit “shared savings” payments to 7.5% of the benchmark. In year 3, this would increase to 10%, but the ACO would also be liable to CMS for up to 5% of the benchmark if costs go significantly above the benchmark (this is called the “two-sided model”). Under “Track 2”, the maximum payment in each year of the agreement would be 10% of the “benchmark”, but the ACO could be liable for up to 5%, 7.5%, and 10% of the benchmark if costs go significantly above the benchmark in years 1, 2, and 3, respectively (i.e., it’s under the “two-sided model” the whole time). After completing its first agreement, an ACO may only participate in Track 2 in the future. In both Tracks, “loss” payments would be payable to CMS in full within 30 days of CMS assessing them.

The proposed rule goes into substantially more detail about how the benchmarks, savings, and losses are calculated, and there are many hoops to jump through to get the maximum payment, as well as minimum loss and savings thresholds to trigger savings or loss payments, but here’s the upshot: Under the “one-sided model”, an entity could get up to 50% of its “shared savings” plus any extra payments, net 2 percent of the benchmark. For (fictionalized) example, in Track 1’s “one-sided model” period, for an ACO with an aggregate benchmark of $100,000, to get the maximum 7.5% payment of $7,500 the ACO would need to meet all quality requirements—and there are many—in the proposed rule and demonstrate savings of 17% over the benchmark (i.e., (($17,000 - $2,000) * 50%) = $7,500; see 76 Fed. Reg. at 19613, 19646–7). However, if 41% or more of the ACO’s beneficiaries visit a rural health clinic (RHC) or federally-qualified health center (FQHC) at least once in the year upon which payment is based, the ACO would get a 2.5 percentage point bonus ($2,500 in this case) above the quality-adjusted savings it earned and would only need to demonstrate 12% savings over the benchmark (i.e., (($12,000 - $2,000) * 50% = $5,000) + $2,500 = $7,500; see id.).

Under the “two-sided model” (i.e., year 3 of Track 1 or all of Track 2) with the same hypothetical $100,000 benchmark, to get the maximum 10% payment ($10,000), an ACO would get up to 60% of its “shared savings” for meeting all quality measures and other requirements, and would therefore need to demonstrate 16.67% savings over the benchmark ($16,667 * 60% = $10,000; no “net” requirement and a higher savings ratio), or only 8.33% of the benchmark if 41% or more of its beneficiaries visit an RHC or FQHC at least once in the year upon which payment is based, which results in a 5 percentage point bonus (i.e., (($8,333 * 60%) = $5,000) + $5,000 = $10,000). 

The Stick (or carrotstick if you prefer)

All sounds good (or at least marginally so) so far, right? Well, here’s the rub: under CMS’s proposed rule, as noted above, although beneficiaries are assigned to an ACO based on primary care services only, with no minimum threshold for a quantity of primary care services, CMS is proposing that “savings” and “losses” will be computed based on all care furnished to a beneficiary, regardless of where the beneficiary receives that care. 

Thus, under the proposed rule, primary care physicians that belong to an ACO could have a strong incentive to not refer patients to facilities or specialists outside of the ACO or that they view as inefficient. The proposed rule would penalize ACO primary care physicians for dropping “at risk” patients, but it does not per se prevent ACO primary care physicians from changing their referral patterns. 


Granted, under current law (CMP/AKS/Stark; take your pick), directing referrals to specific people or groups to maximize one’s own revenue would likely be very problematic (essentially, the physician would get a kickback from the ACO for referring a patient to a specific entity; ACOs would pay kickbacks to health care providers who withhold certain care, etc.), but much of what an ACO is supposed to do would likely be problematic under current law, which is one reason that CMS and OIG are soliciting comments on how to carve out exceptions that would allow ACOs to lawfully operate. Bear in mind, however, that if physicians cannot direct their referrals to other health care providers in their ACOs or other providers they view as efficient, the entire purpose of the ACO system would be undermined and ACO physicians would risk losing a substantial mount of money. Thus, it is likely that these sorts of preferred referrals would need to be carved out for the ACO program to have any chance of success.


This is all well and good if you intend to join an ACO and take on all the administrative responsibilities and financial risk. However, the problem is that if you’re a health care provider and you’re not in an ACO or don’t plan to join one, the ACO program is not a carrot so much as a stick. Why is that? 


If you’re a health care provider and you don’t join an ACO, you run the risk of losing referrals from ACO physicians in the area unless you reduce the amount you charge Medicare for each patient. If you do that without joining an ACO, however, you lose out on the bonus payments, which could help recover some of the reduction in Medicare reimbursement (which could get promptly eaten up—or surpassed—by administrative costs unless the ACO is large enough, but that's another story). 

Conclusion

So, what to do about all of this? Tell CMS what you think. Almost everything above comes from a “proposed rule”, meaning that it’s just a proposal. CMS wants and needs your input. If you are concerned about the issues raised here, responding to the CMS/OIG fraud and abuse request for comment as well as commenting on the CMS ACO reimbursement rule would likely be a good way to help CMS and OIG strike a balance between achieving the aims of ACOs and protecting beneficiaries from unscrupulously directed care. Instructions for submitting comments are in the Federal Register notices linked to above. 

When preparing comments, it may help to consider the following: CMS is statutorily authorized to determine how to assign beneficiaries to ACOs (see 42 U.S.C. § 1395jjj(c)) or to use “other” payment models (see 42 U.S.C. § 1395jjj(i)(3)), so comments on this proposed rule could, theoretically, help CMS decide how to adjust the beneficiary assignment process and/or remove or limit the negative incentive to refer outside of an ACO or to fail to join an ACO.


If you have a lawyer you work with on a regular basis for regulatory issues such as this, I would encourage you to work with him or her to draft and submit comments that make sense for you. If you don't have someone like that, please contact me, and I may be able help you find the right person to help you address your concerns.

All that said, where Medicare goes, other payors often follow. Thus, although the full impact of ACOs is unknowable at this stage, responding to the proposed rule and other requests for comment could not only directly affect how ACOs are implemented for Medicare, but could influence private reimbursement for years to come.

© 2011 Alex M. Hendler. All Rights Reserved.

No mere "mandatectomy" for PPACA/HCERA: Off with their heads!

On Monday, Jan. 31, Senior U.S. District Judge Robert Vinson in the Northern District of Florida ruled that the "individual responsibility" tax penalties (the so-called "individual mandate") of PPACA/HCERA (collectively, the "Act") are unconstitutional as written. Furthermore,  he held that because the "mandate" cannot rationally be severed from the rest of the Act, the whole mess of a law needs to be thrown out and Congress needs to start over. Ilya Shapiro offers a comprehensive take on the decision here, along with a link to the decision itself.

I've finally had a chance to read through the decision, and it's a good read in general; a particularly good read when you consider that it's a court decision. If you ever need a primer on the history and evolution of Commerce Clause jurisprudence, this is a great place to start. 

Judge Vinson fully accepts the "inactivity" argument against the "mandate", which is the rather myopic—and perhaps legally correct, albeit not economically correct—view that because someone makes a present decision not to purchase health insurance, this is economic "inactivity", and the future economic impact of potentially uncompensated care can only be regulated once it actually occurs or is imminent. That is, it appears that Judge Vinson would have no legal problem with a law that financially penalizes an uninsured person who goes to the hospital and can't afford to pay, but there is a big legal problem with prospectively penalizing that person.

Clearly, this stance does not work for an insurance system, which relies on people paying into the system before they get sick or injured, but that is really not the judge's concern: His concern is whether Congress has the power to compel people to do this or face immediate financial consequences (his answer is no, at least not the way they did it here). Judge Vinson also firmly rejects the notion that health insurance and health care are in any way "unique", and thus the same rules that apply to any other activity (or "inactivity") should apply here; the rules being that Congress can regulate only "activity", and this is not it, no matter how "unique" it might be. 

(Whatever the merits of his Commerce Clause analysis, Judge Vinson is 100% wrong in footnote 14 about young people not being able to buy scaled-down "catastrophic" insurance plans:  PPACA § 1302(e) (124 STAT. 168) explicitly provides for this; this mistake is certainly not a dealbreaker in terms of the opinion's overall validity, but it certainly reveals that we all make mistakes, as I'll discuss a bit below.)

Of course, this decision will be appealed, but there are some other important issues worth noting:



A tax is a tax is a tax. Unless it's not....

As much as it pains me to admit it, I may have been wrong about there being a lack of Constitutional issues with the "mandate." As Mr. Shapiro and Judge Vinson have noted, courts have so far universally rejected the assertion that the PPACA § 1501 penalty provision (as amended by PPACA § 10106 and HCERA § 1002) is a "tax" (see also p. 4, n. 4 of Judge Vinson's decision). So what, exactly, is wrong with calling this thing a "tax"?

As I have freely admitted, I am by no means a tax scholar, but some researchers have very cogently and convincingly explained that the "penalty" imposed by the law acts more like a "capitation tax" than an income tax because it potentially imposes an obligation upon a taxpayer to pay per-person dollar amounts, and as such, the Constitution demands that it must be "apportioned" among the States, which it isn't (i.e., the amount of total tax collected from residents of any given State is not related to the State's population as a proportion of the national population). Thus, if the government successfully argued that the "penalty" were a tax, it would be almost necessarily conceding its unconstitutionality.

However it seems that Congress would be fully within its rights to levy a 100% income tax, then refund it to people to reward behavior that it deems desirable. This already happens to some extent with the mortgage interest deduction, hybrid car credits, and child tax credits. Of course, a 100% income tax would be politically ruinous and people would storm the Capitol with pitchforks plastic forks and torches flashlights (welcome to the post-9/11 world...).

Thus, if Congress had, for example, enacted an across-the-board income tax increase (not very popular in an election year) or imposed a new "excise tax" (like Medicare or Social Security) then offered refundable tax credits roughly equivalent to—or perhaps even identical to or exceeding—the amount of the new tax paid, less the penalty amounts currently in the Act, this would almost certainly be a "tax". For example, the following would probably be legally (if not politically) acceptable: enact a 2.5 percentage-point income tax increase for every taxpayer in every bracket, refund all of it to everyone as a tax credit, but reduce the credit by the greater of $[X] per person in the household who does not have qualifying insurance or the entire 2.5 percent.

This is functionally almost indistinguishable from the Act as it is currently written, but formally, the difference is enormous. The bottom line seems to be this, based on the materials I've read on the topic: Congress can tax your income and give the money back for whatever reason it wants, e.g., if you buy a hybrid car, a home, or health insurance, or you have a child; Congress cannot, however, simply take money from you if you choose not to buy a hybrid car, a home, or health insurance, or have a child. This latter bit seems to be what is happening here, and I'm coming around to the idea that Congress should not be able to do this outside the income tax regime.

A major benefit of requiring this sort of coercive regulation—and I do mean that in the nicest possible way—to go through the income tax wringer is that not only does this give Congress the clear authority to act, but Congress will likely think very, very carefully about it before doing anything. Few constituents would ask their Representative or Senator to support a tax hike. Which leads us to....

Political Compromises?


Earlier versions of the Act contained a "severability" clause, meaning that if a court struck down any portion of the law, Congress would intend for the rest of it to go forward, but Congress dropped this clause before passing the law. See pp. 66-68 of Judge Vinson's opinion.

Why drop the clause? The bill likely could not have gotten endorsement from key supporters unless linked to the "mandate", e.g., insurers would never have gone along with the bill without the "mandate," which makes perfect sense.

In any event, the government in this case argued—quite honestly—that large portions of the Act could not function as intended without the individual mandate. Judge Vinson therefore reasoned that if the individual mandate is unconstitutional as written, all the portions of the Act that rely upon it must therefore be discarded; because the Act is so big and complicated, he reasoned, it goes beyond the judge's role to pick and choose which provisions are inextricably bound to the "mandate" and which are not, so he threw out the whole thing. Yes, that's right, all of it.

Was this the best way to go? That depends what he was trying to accomplish, but you have to admit that this is one surefire way to bump this controversy to higher courts.

How to fix it

Now that the midterm elections are over, Congress could likely head off the leading potential Constitutional problem (although I'm sure that people vehemently opposed to the bill will look for others) with a comparatively simple bill to do the following:

1) Amend the Act to add a severability clause. This is not foolproof, but it could help prevent the whole law from being thrown out if courts find another problem with one of its provisions.

2) Repeal the "individual responsibility" penalty provisions (PPACA § 1501 and amendments thereto). Despite my earlier argument that these are functionally equivalent to Constitutionally-permissible taxes, after becoming more familiar with the case law and scholarly research, the formal difference is dramatic, and possibly even insurmountable.

3) The hard part (politically): Institute a new income or excise tax, say, a flat rate of 2.5 percentage points of taxable income (i.e., the maximum current "penalty" as of 2014), and set aside the proceeds in a trust fund that helps subsidize uncompensated care or the purchase of insurance for those otherwise unable to afford it.

4) The slightly easier part: Amend PPACA § 1401 (26 U.S.C. § 36B) to expand the existing refundable tax credits for qualifying insurance coverage to completely wipe out taxpayer's tax liability under the new tax if they purchase qualifying insurance or are unable to afford qualifying insurance. For those able, but unwilling, to get insurance, reduce the refundable credit by an amount equal to what the § 1501 penalty (as amended) would have been, not to exceed the amount of the new tax. This is analogous to the mortgage interest deduction or the child tax credit, the only possible twist being that instead of receiving an larger deduction or credit for engaging in desired behavior, someone would be receiving a smaller credit for engaging in undesired behavior. This would likely still be controversial, but substantially less so than the current situation.

Conclusion

There is no guarantee that any of this would be politically or legally feasible, but in my view these fixes would moot the Commerce Clause issue upon which this and other "mandate" cases are based, putting the Act on considerably more solid footing and allowing implementation to go forward.

Why should implementation go forward? The Act is far from an ideal solution to the complete market failure we have had in the health care/health insurance industries, but it is a step towards some meaningful reform. Hopefully politics will not continue to impede progress in that direction. 


© 2011 Alex M. Hendler. All Rights Reserved.