Lambert here: I’ve previously posted about HR676 (Conyers) vs. S1804 (Sanders) on provider payments and phase-in, and about HR676 vs. S1804 on incentivizing institutional providers with profit (don’t). Now Kip Sullivan, a grizzled and authoritative veteran of the single payer battles from 2009-2010 and well before, looks at cost containment in HR1384 (Jayapal) and S1129 (the new Sanders bill). Sullivan’s bottom line:
I fear the public will give us only one more bite at the apple during my lifetime. If we screw it up again with legislation with no cost containment in it, the division over how to solve the US health care crisis may remain unresolvable for decades.
This is an important post.
By Kip Sullivan, a member of the Health Care for All Minnesota Advisory Board and of the Minnesota chapter of Physicians for a National Health Program. Originally published at The Deductible.
Two bills that are called “Medicare for all” bills by their supporters have just been introduced in Congress. On February 27, Representative Pramila Jayapal introduced the Medicare For All Act of 2019, HR 1384 , in the House of Representatives. On April 10, Senator Bernie Sanders introduced a bill bearing the same name in the Senate, S 1129. The cost-containment section in Representative Jayapal’s bill will cut health care costs substantially without slashing the incomes of doctors and hospitals. Senator Sanders’ bill cannot do that.
In this article, I explain the differences in the cost containment sections of the two bills and call upon Senator Sanders to correct two defects in his bill that minimize its ability to reduce costs. Defect number one: S 1129 authorizes a new form of insurance company called the “accountable care organization” (ACO). Defect number two: S 1129 fails to authorize budgets for hospitals. Representative Jayapal’s bill, on the other hand, explicitly repeals the federal law authorizing ACOs, and it authorizes budgets for individual hospitals.
I write this essay as both a long-time organizer, writer and speaker for a single-payer (the older name for “Medicare for all” system) and a strong supporter of Senator Sanders. Bernie’s enthusiastic support for a “single payer” solution to the American health care crisis has added millions of new supporters to the single-payer movement. But precisely because he is now the most recognizable face of the single-payer movement, it is extremely important that all of us, whether we’re already in the single-payer movement or we just long for a sane and humane health care system, encourage Bernie to fix the defects in his bill.
To explain the two defects in S 1129, I must first explain why a single-payer bill like Representative Jayapal’s will be effective at cutting the high cost of American health care. I begin by explaining the origin and meaning of the “single payer” label. I will then describe the two defects in S 1129 in more detail.
The Origin of “Single Payer”
The “single payer” label entered the lexicon in 1989 to describe a universal coverage proposal published in the January 1989 edition of the New England Journal of Medicine by Physicians for a National Health Program (PNHP). The “single payer” label was invented within months after the article’s publication to characterize the distinguishing feature of PNHP’s proposal: One government agency would reimburse doctors and hospitals directly; in other words, the payments would not flow first to insurance companies so they could take 15 to 20 percent off the top before passing on the rest to doctors and hospitals. In addition to the one-payer feature, the PNHP proposal contained three other essential features – the use of budgets for hospitals (annual lump sum payments akin to budgets for fire departments), the establishment of a uniform fee schedule for doctors, and limits on drug prices.
The advantage that PNHP’s proposed single-payer system has over other cost containment proposals is that it reduces costs primarily by reducing price of medical services (as opposed to the quantity of medical services Americans receive), and it reduces prices mainly by reducing excessive administrative costs that drive prices up. Administrative costs currently eat up about one-third of total US health care spending. Research has demonstrated that three of the four features of the PNHP proposal – one payer, budgets for hospitals, and uniform fee schedules for doctors – could reduce the share of total spending devoted to administration from over 30 percent to 15 to 20 percent. Drug price controls could reduce total spending by another 3 to 5 percent, for a total savings of somewhere between 15 to 20 percent. Fifteen to 20 percent of today’s US health care bill, which is $3.5 trillion, is an enormous sum – approximately $500 to $700 billion saved every year. Note that this does not include additional savings that might flow from reduced fraud (it’s easier to detect fraud with one payer), reductions in the compensation paid to some specialists, and more equitable distribution of hospitals and equipment. 
Within a few years after the publication of PNHP’s 1989 proposal, the “single payer” label had been widely adopted by health policy researchers around the world to distinguish systems like Canada’s, in which tax dollars flow directly from provincial governments to doctors and hospitals, from systems like Germany’s in which tax dollars and compulsory premium payments flow first through insurance companies.
By 1990 the modern American single-payer movement had begun to form. The first single-payer legislation was introduced in the Ohio legislature in 1990, in the US House of Representatives in 1991, and in the US Senate in 1992. Those bills all contained the four basic elements of a single-payer system as PNHP proposed it: One payer (a government agency) replaces multiple insurers (insurance companies and self-insured-employer plans), and the one payer negotiates annual budgets with hospitals, uniform fee schedules with doctors, and price ceilings with drug companies. Insurance companies could sell insurance for services not covered by the legislation, but not for services that were covered. 
Throughout the 1990s and early 2000s, the traditional fee-for-service (FFS) American Medicare program (to be distinguished from the Medicare Advantage program in which hundreds of insurance companies participate) was often cited as an example of a single-payer system even though it did not fit the ideal described by PNHP and other experts. The most obvious similarities: traditional FFS Medicare was the sole payer of providers (doctors and hospitals) who treated Medicare beneficiaries; and it paid providers directly – the payments did not pass first through insurance companies. Thus, strictly speaking, the FFS Medicare program qualified as a single-payer program. However, unlike the ideal single-payer system, traditional Medicare has never had the authority to negotiate annual budgets for each of America’s 5,500 hospitals, nor to negotiate limits on drug prices. And, of course, it doesn’t cover all Americans.
Because the traditional Medicare program was a single-payer program, although less than an ideal one, and because the public is familiar with Medicare, single-payer proponents often substituted “Medicare for all” for “single payer.” For example, recent iterations of HR 676 , a single-payer bill modeled on PNHP’s proposal that was introduced in the US House of Representatives between 2003 and 2018, have been entitled “Expanded and Improved Medicare for All Act.” Polling data indicate that explaining single-payer by comparing it to Medicare raises public support for (and perhaps understanding of) the concept. A 2007 AP-Yahoo poll illustrated the difference between public support for universal coverage under a “single payer” versus a program “like Medicare.” When asked whether “[t]he United States should continue the current health insurance system … or should adopt a universal health insurance program in which everyone is covered under a program like Medicare that is run by the government and financed by taxpayers,” 65 percent chose “a program like Medicare.” However, when asked if they considered themselves a single-payer supporter, only 54 percent said they did.
Adulteration of “Single Payer” and “Medicare for All”
Like all bumper-sticker labels designed to describe complex proposals, the phrases “single payer” and “Medicare for all” cannot by themselves convey all the elements of a complete single-payer system. The terms are therefore easily muddled by those who want to hitch significantly different proposals to the single-payer bandwagon, and by those who oppose single-payer systems.
With the rise of single-payer’s popularity in the last three years, much confusion has been visited on that term as well as “Medicare for all.” The confusion arises from two sources: “Medicare for some” advocates – people who want to expand Medicare to some but not all Americans; and universal coverage advocates who bestow the “single payer” label on bills that are missing two of the essential features of a single-payer system – one payer and budgets for hospitals. In this essay, I’m focusing on the latter problem – the introduction of bills that are labeled “single payer” or “Medicare for all” by their authors and supporters, but which are missing the one-payer feature and hospital budgets. The most prominent of these bills is S 1129, the bill Senator Sanders introduced on April 10, and its predecessor, S 1804 , which he introduced in September 2017.
S 1129 contains the clause that every true single-payer bill contains – a clause stating that on a certain date after the bill becomes law insurance companies may not sell policies that duplicate the coverage of the Medicare For All program. If you read nothing else in the bill, that clause would lead you to think S 1129 was a classic single-payer bill – a bill that replaces today’s insurance industry (about 1,000 insurance companies) with one government payer, in this case, the federal Department of Health and Human Services (HHS). But elsewhere in the bill we come upon a section that indicates insurance companies will be replaced by a hybrid of existing insurance companies plus hospital-clinic chains, a hybrid called “ACOs.”  There are more than 1,000 ACOs in business today.
Defect 1: ACOs Cannot Control Costs
Replacing 1,000 insurance companies with 1,000, or more likely several thousand, ACOs cannot reduce administrative costs substantially or at all.  Under such a system, ACOs would generate almost the same overhead costs insurance companies generate now (15 to 20 percent of premium payments), and hospitals and clinics would incur roughly the same administrative costs billing multiple payers (i.e, the 1,000-plus ACOs).
Congress included in the Affordable Care Act of 2010 (aka Obamacare) a section (Section 3022) requiring CMS to establish an ACO program within the traditional FFS Medicare program. It is not clear why Congress did that. Congress was warned in 2008 by the Congressional Budget Office (CBO) that ACOs would not save money for Medicare. The simplest way to describe ACOs is to say they are HMOs in training. Like HMOs, they are corporations that own or contract with chains of hospitals and clinics; they have the equivalent of enrollees; they attempt to keep their “enrollees” from seeking care outside their networks; they bear insurance risk (that is, they are paid on a per-enrollee basis and in exchange are obligated to provide medically necessary services to their enrollees); and because they are risk-bearing organizations, they generate overhead costs similar to those created by traditional insurance companies.
In fact, precisely because ACOs resemble insurance companies, nearly half of them already have contracts with insurance companies to help them carry out insurance-related tasks. The largest insurance companies – Aetna Humana , and United Healthcare , for example – are already deeply embedded in the ACO industry.
The only significant differences between ACOs and HMOs are (1) ACO “enrollees” are assigned to ACOs (usually without their knowledge) whereas HMO enrollees choose to enroll, and (2) HMOs bear all insurance risk while ACOs split the risk of loss or savings with another insurer (in Medicare’s case, risk is shared with the Medicare program).  Both of these differences are being eroded. Many ACOs are saying they should be allowed to enroll people so they can restrict enrollee use of out-of-ACO providers, and some influential ACO proponents are proposing that ACOs be paid premiums so they can absorb total losses and keep total profits.
One other important similarity between ACOs and HMOs: ACOs have failed to cut Medicare’s costs, just as the CBO predicted. 
Defect 2: Absence of Hospital Budgets
I mentioned above that an American single-payer system could reduce total spending by 10 to 15 percent just by eliminating excess administrative costs. A large portion of that savings would come in the form of reduced administrative costs for hospitals (the rest comes from reduced administrative costs in the insurer and physician sectors). Hospitals enjoy lower overhead costs in single-payer systems for two reasons. First, they are paid with annual budgets, not on a per-patient or per-procedure basis, which means they don’t have to keep track of very pill and x-ray for every patient. Second, for the covered services, they deal with only one payer, not hundreds, each with their own hoops to jump through.
Unlike Representative Jayapal’s bill, Senator Sanders’ bill does not authorize hospital budgets. There is a reason for that: It is not possible to set premiums for 1,000 or 2,000 ACOs, which consist of hospital-clinic chains with an insurance company or department plopped on top of it, and at the same time set budgets for each of the nation’s 5,500 hospitals. One has to choose one or the other: Premium payments for ACOs, or budgets for hospitals. Sanders chose ACOs. Jayapal chose hospital budgets.
But by sacrificing hospital budgets in order to make room for ACOs, Sanders guaranteed his bill cannot reduce hospital administrative costs much or at all. Research indicates US hospitals spend 25 percent of their revenues on administration, thanks to the complexity of our multiple-payer system, while hospitals in single-payer systems that use hospital budgets devote half as much to administrative costs.
We must take into account as well the additional administrative costs for doctors in Sanders’ proposed multiple-ACO system. Like hospitals, they will have to determine, for each patient, which ACO a patient belongs to and send the bill to the right one.
Cost Containment Must Accompany or Precede Universal Coverage
If you happen to believe that some fine day America will find the political will to insure everyone at the high price at which health care is sold today, then you should ignore what I have said here. You should feel free to endorse any legislation that proposes to raise taxes high enough, or levy compulsory premium payments high enough, to achieve universal coverage. No need to worry about whether a bill that purports to achieve universal coverage can reduce costs. No need to ask yourself why US per capita health care costs are double those of the rest of the industrialized world. The answer is: Primarily because of our high prices, and the excessive administrative costs generated by our multiple-payer system that drive prices up, and not “overuse” of medical services.
But if you think, as I do, that cost containment must accompany or precede universal coverage, then you must support legislation that includes evidence-based, cost containment provisions, such as Representative Jayapal’s bill. I do not believe our nation will find the political will to pay for universal coverage at today’s prices. Moreover, even if the political will were there, I don’t believe it is ethical to pay more to solve any social problem than we have to. Our society has numerous other demands on our resources, ranging from hunger to crumbling infrastructure to climate change. Paying more than necessary to insure all Americans means we will have fewer resources left with which to address other problems.
Representative Jayapal’s bill, HR 1384, meets the definition of a single-payer bill as originally outlined in PNHP’s 1989 article and as most experts define the term. It contains the four elements of a single-payer system: It relies on one payer (HHS, not multiple payers called ACOs) to pay hospitals and doctors directly; and it authorizes budgets for hospitals, fee schedules for doctors, and price ceilings on prescription drugs.
Senator Sanders’ bill contains two of those four elements – fee schedules for doctors and limits on drug prices. That’s a good start. He should add the other two. He should get rid of Section 611(b), the section that authorizes ACOs, and thereby ensure HHS is the single payer. And he should add a section authorizing HHS to negotiate budgets with each of the nation’s hospitals.
 Administrative costs in the US are generated both by the insurance sector (insurance companies, self-insured employers, and public programs like Medicare) and the provider sector (clinics, hospitals, nursing homes, etc.). Research indicating that the administrative costs generated by the insurer sector would fall under a single-payer system tends to focus on comparisons between the traditional fee-for-service Medicare program (which devotes 2 percent of its expenditures to overhead) and the private sector (insurance company overhead is on the order of 15 to 20 percent while the overhead of self-insured employers is 8 to 10 percent), as well as comparisons of the US with other countries. Research indicating that a single-payer system could cut administrative costs of providers tends to focus on comparisons of US clinics and hospitals with those of other countries, particularly Canada. Administrative costs are about 40 percent lower in Canadian hospitals and clinics than they are in the US.
 The former HR 676, Representative Jayapal’s new bill (HR 1384), and other single-payer legislation divides the medical sector into “institutional providers” and “individual providers,” and authorizes budgets for the former and uniform fee schedules for the latter. “Institutional providers” typically refers to hospitals and nursing homes.
 The section in Sanders’ bill that authorizes ACOs is Section 611(b). That section authorizes the Secretary of HHS to impose upon the non-elderly ACOs and other “payment reform” programs currently applied to the traditional FFS Medicare program. Section 611(b) reads: “Any payment reform activities or demonstrations planned or implemented with respect to such title XVIII [this is the title in the Social Security Act that created Medicare] as of the date of the enactment of this Act shall apply to benefits under this Act, including any reform activities or demonstrations planned or implemented under the provisions of, or amendments made by, the Medicare Access and CHIP Reauthorization Act of 2015 … and the Patient Protection and Affordable Care Act….” For a quick review of Medicare’s ACO programs and six other “payment reform activities” Section 611(b) would impose on the entire country, see a memo I wrote available here http://www.healthcareforallmn.org/wp-content/uploads/2018/08/op-critique-of-Section-611b-S-1804.pdf . Some of these programs are putting patients at risk https://www.nytimes.com/2019/01/18/health/medicare-hospitals-readmissions.html .
 It’s impossible to paint a precise picture of what S 1129 ’s multiple-ACO system would look like because Section 611(b) is so brief (see footnote 3). The big unknowns are the number of ACOs that would spring up, and whether Congress or HHS would change the rules that currently determine how people are assigned to ACOs.
Today there are over 1,000 ACOs in operation. Of these, about half hold contracts with CMS and the other half hold contracts with traditional insurance companies. We have very little information on how the ACOs with private-sector contracts assign insured people to ACOs, but what we have indicates they follow CMS’s method. CMS assigns Medicare beneficiaries to ACOs by a two-step process. First, beneficiaries are “attributed” to the primary care doctor they saw the most often over a look-back period of two or three years. Second, if the doctor to whom the beneficiary was attributed is part of an ACO, CMS’s computers sweep that person into that doctor’s ACO. Beneficiaries who had no primary care visits during the look-back period are not assigned to an ACO. By this algorithm, CMS currently assigns 12 million FFS Medicare beneficiaries, or about a third of the beneficiaries enrolled in the traditional FFS Medicare program, to Medicare ACOs. (For more details, see Table 8-2 p. 218, in the Medicare Payment Advisory Commission’s June 2018 report to Congress http://medpac.gov/docs/default-source/reports/jun18_ch8_medpacreport_sec.pdf?sfvrsn=0 )
This plurality-of-primary-care-visit method of assigning Americans to ACOs will sweep a much smaller portion of the non-elderly into ACOs because the non-elderly are far less likely to have made a visit to a primary care doctor in any given look-back period. In the event that S 1129 becomes law, would HHS alter its assignment algorithm to ensure that a much higher proportion of non-elderly Americans would be assigned to an ACO? It is more likely, in my view, that HHS would abandon the invisible, two-step assignment method entirely and allow, induce, or require Americans to enroll in an ACO. Obviously, if that happened, membership in ACOs would soar, and that in turn would induce more insurers, hospitals and clinics to band together to create ACOs.
Another unknown is how much ACOs would be paid. If they were overpaid relative to the FFS sector, as Medicare Advantage plans are overpaid vis a vis Medicare’s FFS providers, we would see more ACOs than we would if they were not overpaid.
 In Medicare’s case, the great majority of the 600 ACOs that currently have contracts with the Centers for Medicare and Medicaid Services (CMS) begin with exposure to upside risk only, that is, to the possibility of sharing savings with CMS, but must eventually accept two-sided risk, that is, both down- and upside risk.
I am sometimes asked how the ACOs in S 1129 can be called “insurance companies” or “risk-bearing entities” if they are paid fee-for-service. It’s true that CMS pays providers within ACOs by FFS throughout the course of the “performance year” in question, but the target for total spending by the ACO for that year is set as if the ACO were going to be paid just as an insurance company would be, that is, on a premium-per-enrollee basis. Before the performance year, CMS’s computers assign Medicare beneficiaries to an ACO (see footnote 4 above), then calculate an average per-assignee premium for the year (crudely adjusted for the health status of all the assignees), and then add up all the premium payments. That total equals the target against which the ACO’s profits or losses will be determined at the end of the year. If the ACO’s total spending exceeds the target at the end of the year, it owes CMS roughly half of the excess. Conversely, if the ACO’s total spending comes under the target at the end of the year, CMS splits the savings with the ACO. The fact that actual payment during the year is done by FFS doesn’t change the fact that risk is shifted from CMS to the ACOs. It’s the end-of-year adjustment based on the target level of annual spending that shifts risk.
This convoluted method of paying ACOs may change. Many ACO proponents, including Elliott Fisher (the “father of the ACO) have argued that ACOs should bear all insurance risk as they develop more expertise at managing risk. That would mean they would be regulated by the insurance departments of the states they operate in, they would have to set aside reserves, and they would presumably be allowed to impose financial penalties on enrollees who seek care outside their networks.
 The table below summarizes the performance of all four ACO programs set up by CMS over the last 14 years. The table indicates ACOs have saved almost no money for Medicare.
Impact of CMS’s ACO Programs on Medicare Spending*
Physician Group Practice demonstration (10 ACOs):
2005-2010: 0.3 % **
Pioneer demonstration (32 ACOs in year one, 8 by year five):
Medicare Shared Savings Program: (480 ACOs in 2017):
Next Generation (45 ACOs in 2017):
*A minus sign means the program’s net effect was to raise Medicare’s costs. “Net effect” is calculated by subtracting from gross savings the bonuses (or shared savings payments) CMS pays out to ACOs, in the case of the MSSP program. In the case of the other three programs, “net effect” is calculated by subtracting from gross savings the net of CMS’s payments to ACOs and ACOs’ payments to CMS.
**The ACOs in the PGP demo would have lost money if the PGPs had not upcoded at almost twice the rate of the groups that served as their controls.
Sources: For the PGP demo, see p. 64 of Evaluation of the Medicare Physician Group Practice Demonstration https://downloads.cms.gov/files/cmmi/medicare-demonstration/PhysicianGroupPracticeFinalReport.pdf .
For the Pioneer, MSSP and Next Generation programs, see Tables 8-6, 8-3, and 8-7 respectively of Chapter 8 http://medpac.gov/docs/default-source/reports/jun18_ch8_medpacreport_sec.pdf?sfvrsn=0 of MedPAC’s June 2018 report to Congress. The 2017 figure for the MSSP program is from slide 9 of a presentation to MedPAC at their January 2019 meeting http://medpac.gov/docs/default-source/default-document-library/aco_jan_2019_final.pdf?sfvrsn=0