Lessons From Massachusetts’ Failed Healthcare Cost Experiment

Massachusetts passed a massive medical cost control bill in 2012, a “Hail Mary” effort to make health-care more affordable in the nation’s most expensive medical market. The problems of the Massachusetts’ law offer invaluable lessons for the nation’s health-care struggles.

Driven in part by a Boston Globe investigation that exposed the likely collusion of the Partners Healthcare hospital system (including several Harvard Medical School teaching hospitals) with Blue Cross/Blue Shield, the largest healthcare insurer in the state, the law marked the biggest health reform since Romneycare in 2006. While most agree Massachusetts needed cost controls, there’s no evidence that the 2012 law has accomplished its goal—and these same failed policies have been folded into the national Affordable Care Act.

Romneycare, Massachusetts’ universal health-care law, already lacked effective rules to control the rapid growth of state medical costs. That, paired with the Boston Globe’s exposure of the likely collusion between Partners Healthcare, the largest hospital system in New England, and Blue Cross Blue Shield to raise health care payments to hospitals and doctors by as much as 75 percent, led to passage of a hefty, 349-page cost-control law in 2012. The legislation included a dizzying number of committees, an uncoordinated “cost containment” process, and dubious quality-of-care policies, like “pay-for-performance,” a program that pays doctors bonuses for meeting certain quality standards, like measuring blood pressure. These incentives might make sense in economic theory, but have failed repeatedly in well controlled studies. They’ve even created perverse enticements, such as some doctors avoiding sicker patients. The cost control law also encouraged widespread use of expensive electronic health records, even though there’s no evidence that they save any money.

Additional policies in the law are “accountable care organizations” (ACOs) and “alternative quality contracts,” programs that pay doctors for staying under an arbitrary budget and penalize them for going over it. But the best studies show, again, that these policies actually increase costs if you count the expense of implementing them. In fact, most of the hospitals in the federal government’s “Pioneer ACO” program have dropped out, including Dartmouth Hitchcock, the famous medical center, associated with Dartmouth College, that initially coined the term “ACO.”

Given the MA cost control law’s outcomes so far, we should be very skeptical about its projected savings. Suffice to say, the law is too complex, unwieldy, contrary to evidence and, worse, doesn’t have the teeth (e.g., regulating price) to enforce cost reductions among the most powerful medical providers. Nancy Turnbull, one of the architects of Massachusetts’ health reform and a member of the board of its insurance exchange, told the New York Times in 2012 that the law “was not nearly what we need to deal with the market power and the unjust price differences that result.”

So how has the law’s repeated disappointments been described to the public? Recently, a front-page, above the fold story in the Globe suggested the law resulted in a “slower rise in health spending” in 2015—which amounted to a decrease of 0.3 percent from the previous year. Researchers often doubt such tiny effect sizes and for good reason. They usually aren’t real. But even worse, the Globe’s “analysis” relies on just two data points, both from years after the 2012 law passed. Had the Globe’s story begun with data from before the law’s enactment, (see graph below) it would have shown that per-capita health-care spending growth actually increased by more than two percentage points since the policy began. This month, the Globe again wrongly reported that new data in 2016 shows the state has helped reduce health-care spending. The data from the last few years contrasts with the previous decade, when health-care cost growth had been declining steadily. Thus it’s unlikely that the tiny changes in costs from 2014-2016 had anything to do with the new cost control effort. And the average annual cost growth over the last three years is still about twice the level at the time the law passed.

This figure, with more complete data, shows that between 2002 and 2011, the year before the law, there was a much greater decline in Massachusetts healthcare cost growth—from 9 percent annual per capita growth to under 4 percent. The tiny changes in the growth rate from 2014 to 2016 (years after the law) is not evidence that the 2012 state law reduced costs, because cost growth in those years was still twice as high as it had been in 2012 and 2013.  In addition, the reporting did not even mention the start of Obamacare regulations less than two years before the MA cost control law, so it is impossible to know whether these policies caused the increase in health care costs observed in the graph. This untrustworthy analysis could fool Massachusetts policymakers and citizens into thinking that its policies were working, or convince other states to follow Massachusetts’ model.

The Massachusetts cost control law has not fulfilled its promise to “bend the cost curve.” If anything costs have gotten higher. The commission set up by the law “may encourage, cajole, and, if needed, shame [providers] into doing their part to control costs. “But this is insufficient to fundamentally change the behavior of the systems with the most market power. Charges for a single aspirin pill in a hospital can be higher than $25, but only pennies at the local drug store. It is already clear that the health care cost crisis is threatening the viability of essential government services from education to defense. To make matters worse, the state’s largest newspaper acts as the law’s cheerleader without even considering the history of health care costs that clearly shows the law’s shortcomings.

So what should be done now? In the near term, the legislature should consider much stronger cost controls that do not rely on the voluntary cooperation of hospitals or on market-based incentives and penalties. The state (and nation) must abandon ineffective, wasteful “alternative payment arrangements” that save miniscule dollars while costing delivery systems billions. Massachusetts and the nation need effective, negotiated price controls that fairly compensate hospitals but do not allow a few elite institutions to receive excessive reimbursement. But the current Republican health plan would retain all the failed aspects of Obamacare, while dropping the parts of the law that gave millions of poor Americans access to health-care.

There are better solutions: Our political system thus far has rejected government-run, single-payer systems, like the British National Health Service and Canada, but there are plenty of private, market-based models in Europe that provide better care to their citizens at half the cost. They rely on straightforward price controls, negotiated between the private and public sectors—not more failed market incentives. Germany’s system, for example, relies on hundreds of competing health insurers. The key difference is the use of medical fee schedules (price controls) that are negotiated between provider associations and insurance programs that prevent the kind of price fixing uncovered by the Globe. Switzerland has similar arrangements. The companies compete on quality of service and efficiency. Germany and Switzerland have managed to avoid most of the costly deductibles and copayments that keep health-care out of reach for ordinary Americans: no prohibitively expensive $10,000 deductibles (which is more than most Americans have in savings), as allowed under the ACA. The US could choose the best aspects of these successful plans that fit our culture and our health care environment.

Ten years ago, Massachusetts built the model for the nation’s health reform law, which then expanded health care to more than 22 million Americans and offered needed protections to America’s population, e.g., removal of pre-existing condition restrictions, free mammograms, and the ability to insure adult children. But we remain the only developed country in the world that fails to provide care to all its citizens. And we allow costs to grow at unsustainable rates. Rather than wait for the implosion of our medical system, Massachusetts must again lead by example in the nation’s polarized health reform debate—this time, by establishing affordable health-care as a right for all its residents.

Stephen Soumerai is Professor of Population Medicine and teaches research methods at Harvard Medical School.

Professor Ross Koppel teaches research methods and statistics in the Sociology department at the University of Pennsylvania and is a Senior Fellow at the Leonard Davis Institute (Wharton) of Healthcare Economics.

Marina Bolotnikova is an associate editor at Harvard Magazine.

16 replies »

  1. Steve, the journalists are confused because they are thinking in terms of the actual English language, while health care people are thinking in jargon.

    The phrase “pay for performance” makes doctors uncomfortable, because it implicitly means, “pay for (clinical) performance that is measured.” The alternative is either fee-for-service (pay for volume, unless there’s obvious fraud/malpractice) or, I suppose, employed doctors. Who haven’t been paid for clinical performance either in the kind of formal way that P4P measures. If one goes back to “Buy Right” in the 1980s, before we made up new jargon, that was the point: buy on the basis of some sort of clinical measure or just be content with volume.

    But clear distinctions make any industry built on jargon uncomfortable (see: Pentagon; Silicon Valley), and so we in health care start to define P4P to mean a certain TYPE of pay for clinical performance approach. But, in fact, the journalists are conceptually correct: either you have third-party payers/health systems adjusting pay based on clinical performance targets or you don’t. That’s a sea change: more accountability, less autonomy. The opponents understand this well, the AAFP folks who believe that performance should be measured by each individual patient in each individual interaction, with a perfect market at work, without government or insurers interference.

    The point you and colleagues justifiably make is that the third parties are not really as good at doing this as they think they are or as the journalists sometimes think they are. They can make things worse, not better. Nonetheless, since incentives matter, the choice is clear: either third parties continue to refine incentives for clinical performance (and make up new jargon so everyone feels happy), or we go back to trusting that the individual doctor-patient relationship in FFS will produce the best of all care in the best of all possible worlds.

  2. The article below deals with paying extra compensation for improved quality metrics, either at the hospital or physician level. The contrary research does not support its use and there are potential negative effects as well.

    ACOs and the BCBS quality contract are pay for performance as well. In some cases, the performance metric is costs, and the unit is the group of docs or hospital/doctor organizations.

    You are right that abandoning FFS and capitation are other issues, possibly in need of more evidence.

    I don’t like the common refrain that if you think PfP doesn’t work, “you must be for FFS.” Seriously, some journalists have said this to me. They think there are only two opposing alternatives. They are different interventions with different possible effects.

  3. At this point of the discussion it may be helpful to have a little more clarity on what we mean by “pay for performance.” I generally take this phrase to mean there is a combination of FFS payment (the base) and an at-risk payment (under 50% of the total) tied to some combination of cost and quality targets (weighted more to cost than to quality). Is that what we all mean? If so, the skepticism prevalent here about the impact of P4P on controlling costs makes sense.

    This is, naturally, not the same thing as saying that incentives don’t matter, or that being paid on a capitated or salaried basis would not help move the American system towards a population health approach to care. The U.K. base payment for GPs is capitation, if I understand correctly, and that makes an enormous impact on how doctors practice and effectively serves as a budget to limit their expenses. The Quality and Outcomes Framework added on top of it may not make much difference in improving quality, but this does not at all impugn the idea of moving away from FFS in order to control costs and practice from a population health perspective. You’d need other evidence to do that.


  4. David: Thanks for the comment. The UK Quality and Outcomes Framework hasn’t done any better than here, across the pond. Unintended consequences are universal! Incentives tend to reward past behavior without any new effect. We have published a very easy-to-understand, accessible explanation of the research and results in the below, brief article. Interestingly, using both US and UK examples, it shows that the stronger the research design the worse the results of pay for performance. The worst research with little control on bias, the better it looks! This is part of the growing problem of the non-reproducibility of science.

    You also did a good Podcast (link?) on this, and both Vox News and THCB published op-eds summarizing this work in op-eds.


  5. EMR’s are still template driven, which by their nature leads to garbage in, garbage out. Over $30 billion has been spent to date with little to show for it. Is medicine safer? No. Is it cheaper to deliver care as a result of EMR’s? No. Is the quality of care improved? No.

    Until medical technology and EMR’s get to the point that personalized care can be delivered and genetic information downloaded into an EMR and easily deciphered, they are and have been a waste of money and the time invested.

    Personally, I’m waiting for the first med-mal case that calls out EMRs. MD’s won’t recall the details of an office visit and will be forced to rely on a template.

  6. Watching the PBS series on The Vietnam War, it occurred to me that there is a parallel between that war (or the war in Afghanistan) and our decades old campaign against rising health spending.

    Our son, a retired Captain of the Marine Corps, tell me that it take 30 years or more to put down an insurgency. In the meantime, there are many small battles with reported victories, as we had in Vietnam. The media, certainly the right-wing media, used these vignettes to argue that there was light at the end of the tunnel. There was much dishonest or anecdotal reporting during the Vietnam war.

    In the health-care war, one side of the front defines health spending as cost. The other side defines it as income. The latter are smart and wily, knowing the combat terrain better than anyone else and always ready to invent new tactics to protect their incomes. From time to time the payers, who define health spending as costs, may clobber the health-care insurgence in small encounters. The AQC is an example. But for decades the payers have lost the war.

    A freestanding physician may charge $200 for a patient visit, out of which he or she will cover all facility costs. A hospital then “buys” that physicians, continues to charge $200 per visit, and adds a huge facility charge. And the payers pay, undisturbed by the fact that prices for identical services vary hugely among providers even in small areas. Drug companies can charge pretty much whatever they wish for on-patent drugs, with helpless payers footing the bill. And so it goes.

    So I am sympathetic to Stephen’s et al.’s perspective. I have attended decades of health care conferences at which the most wondrous cost containing tricks were presented, allegedly yielding huge savings. They never scaled up, even if the story about them were true. And health care cost kept rising steadily.

    We do not have the guts to introduce an all-payer system, such as Germany or Switzerland. And so we whine and pay.

  7. Let me follow up on the point David made. AQCs, for a commercial population, is difficult to compare to a Medicare ACO both because of the population and the limits on incentives, etc., that can be offered to the population. Commercial “ACOs” are kind of a joke: remember when anything could be a PPO? Now it’s a commercial ACO. So, one that basis alone, one has to be careful. By all means, feel free to yell, “j’accuse” at the AQC. Just line it up in front of a methodologically separate wall before shooting it down. 🙂

  8. Dear Stephen, Mike, et al.
    Having followed/studied closely both AQCs & ACOs, I’d agree the data/analysis is either/both corrupted and/or too limited to reach any confident conclusions.
    Per Mike’s point it’s difficult to compare AQCs with ACOs. Last I looked, or talked with Dr. Song, the mean age of covered benes. varies by 30+ years.
    Per Kip Sullivan’s related argument that ACOs are doomed to fail, I continue to wonder how these two statements can be simultaneously true: FFS is gamed to optimize/maximize reimbursement; and, ACO financial incentives inherently don’t work?
    Stephen, since the NHS has been at this far longer than us, what do we know about their QOF (Quality and Outcomes Framework)? Hope New Zealand was enjoyed!

  9. Hi Mike: Both Ross and I teach research design and the Alternative Quality Contract has been the the subject of some of the worst and most biased research we have ever encountered. 75% of health policy research never gets referenced in Cochrane reviews based on basic criteria.There are multiple scientific problems with the Blue Cross studies, followed by Boston Globe cheerleading. One of the worst and most well documented is that clinicians VOLUNTEERED for the study group while the “controls” said “forget it.” I wonder if this makes the naysayers different to begin with?? Of course! We recently wrote this Wash Post piece, in part devoted to this nonsense- the most recent data I think. The effects were tiny and untrustworthy; the news reporters even alleged better outcomes among the disadvantaged. Nope. Didn’t happen, not even according to the authors. This is just one more example of hope and hype. See the section on the contract.


    Of course ACOs and Alt Contracters are different but there are incredible similarities in research flaws, like the selection of the best compared to the avoidance of the worst. In fact many CMS demonstrations of incentive policies are designed that way.There is no reason why they couldn’t have done pilot RCTs to prove efficacy before they spread them nationwide, like the failed ACOs. We have embraced incentive policies in a huge way without the data to support them. Sometimes they do harm, like the avoidance of sick patients that might damage your scores. Best, S

  10. I agree that the current system has failed to tame our run-away healthcare spending nor has it significantly improve the quality of care provided to our citizens. However, the MA experiment should not be condemned for trying and it has had some successes (eg increase % patient with insurance has reduce mortality)

    While many will argue that this is prove the Government mandates are counter productive, IMHO, there needs to be a recognition that cowboy capitalism did not work in the healthcare arena. The healthcare insurance came into existence as a means to provide less expensive/higher quality healthcare to the then prevailing modus operandi of the cowboy capitalism healthcare industry. To revert to that system, via the removal of all state and federal regulations on the healthcare industry, would only ensure that we again learn that cowboy capitalism does not work in the healthcare industry and, in the process would hurt many, many people who would have little or no access to the healthcare industry.

    My only gripe about the article is in this sentence.

    “The cost control law also encouraged widespread use of expensive electronic health records, even though there’s no evidence that they save any money.”

    While those who have read my postings understand that I have great reservations about the current crop of electronic health records, I must protest that the references which are linked in the above sentence, are dated from 2012 and 2005, and thus are too old to be considered relevant to today’s discussion. The health information technology field is moving very quickly and any reference in this realm to articles that are more than 2-3 years old is problematic and may not be an accurate reflection our our current situation.

  11. One of the advantages of writing for a blog is you don’t go through peer review; one of the disadvantages is you don’t go through peer review. Since the authors are highly respected academics, I would respectfully suggest they correct their assertion that lumps the results of the alternative quality contracts in use in Massachusetts in with national research on ACOs. Unless you’re interested in polemics, they’re simply not the same thing in operational detail. (Disclosure: I worked with BCBSMA when they were set up many a year ago.)

    So, Song et al. in Health Affairs (July, 2012): “The ‘alternative quality contract,’ based on a global budget, lowered medical spending and improved quality.” In March, 2017, in the journal Healthcare, Khullar and Safran write about medical spending growth levels “less than half of the Massachusetts State Benchmark.” A nuanced view of the challenges and successes, taking into account the fee-for-service context, is in provided by Rob Mechanic in 2016 in the Journal of Health Politics, Policy and Law.

    The authors’ larger point, of course, is about the failure of the Massachusetts “cost control” law. As Mechanic notes (in suitable academic prose), “risk-based payment models” tied to global budgets haven’t exactly taken over. The issue of market power by providers needs to be confronted directly, but expecting Massachusetts politicians to take on the Partners health care system may be a very optimistic assumption.

  12. Amen! We’ve given complex and fragmented incentives enough of a chance to show that they don’t work to control health care costs. MA provides an ideal data set to prove this.

    We’ve danced around the real issues of market power for too long, all because it’s politically difficult. It’s time to try something straightforward and look at pricing directly. Even within price control policies, there are a diverse set of ideological options that a lot of people in health policy know, but we simply don’t hear enough about.

    On the rightward side, there could be a mandate that every provider set one price for a given service, and make it public. The price can be up to them. This would enhance competition, consumer transparency, and lower barriers to entry for new insurers (you still have to deal with consolidation and monopolies, of course).

    On the leftward, there could be an imposed fee schedule, like mandating Medicare rates for all providers. This has many of the virtues of single payer, without the massive disruptions to people’s current arrangements, and could easily be done on a state level.

    In the middle, there are all sorts of mutually negotiated arrangements, like setting fee schedules under a target state or national budget. See Maryland.

    I want to see much, much more written about this. It’s the only way the conversation will change.

  13. RE your comment: “private insurance companies have been doing it [colluding with hospitals] for years, may be without covert collusion.In the past, it was called contract negotiation. The providers would propose a 4% fee hike. The insurer said no. Then the provider responded with “we won’t accept anything less.” Then the insurer offered a 3.9% increase, and the provider said ok but only for one year”

    It happens across the land with small hospitals too. Many health insurance markets have a single dominant player (often BC/BS affiliate) and they negotiate prices and seldom are able or willing to negotiate hard (Jim Purcell has outlined the reasons in THCB articles). The dominant insurer sees itself as a funding mechanism for the health system and often fights new entrants who have lower prices (using certificate of need laws, lobbying power)…and they tolerate hospital mish mash of cost accounting/cross subsidization…hiding inefficiency. I saw it close up as a board member of a community hospital and (at an earlier time) of a dominant non profit health insurer.

  14. All of the world’s systems fail when you consider mental health, drug and alcohol rehabilitation, long term care, dentistry, and precision medicine-prescribed interventions especially in oncology and immunology and rheumatology, but eventually entering all branches of the health care sector.

    We all have a wicked problem and we are going to need a sort of revolution in all our approaches. Maybe we do need an implosion?

    To pay for all our needs may be an unsolvable financial dilemma. There may be no way to be egalitarian unless we ratchet back to banal care for all.

  15. I would re-brand price fixing…AS…institutional codependency. The major “complex healthcare” medical centers and the major private insurance companies have been doing it for years, may be without covert collusion.
    In the past, it was called contract negotiation. The providers would propose a 4% fee hike. The insurer said no. Then the provider responded with “we won’t accept anything less.” Then the insurer offered a 3.9% increase, and the provider said ok but only for one year, as opposed to the usual 3 year renewal. The insurer was then able to offer coverage at the most prestigious medical center in town….especially for the executive suite and their families.
    With or without a single payer, health spending will not be stabilized without direct provider risk-sharing: Primary Healthcare should be capitated with mandatory referral authorization (aka “purchase orders”), the specialists should also be paid fee-for-service, adjusted by a with-hold from a stop-loss protected risk pool, and the hospitals in a third risk pool also with stop-loss protection. Having worked in a gate-keeper HMO plan for 10 years as a fee-for-service primary physician, it will reduce the cost of healthcare by 30% with better outcomes. The improvement would require a higher level of “medical TRIAGE” by the primary physicians, phased in over a period of 3-5 years, AND apply to all payers and providers including Medicare, the VA, the community health centers, and the native American Health Service. Individual citizens could “buy out” the restrictions with a premium and higher co-pays. The medicaid and Federal-subsidized citizens would have a lower co-pay, pharmacy lock-in and possibly a smaller referral network. The simultaneous humanitarian processes to reduce the social adversities that increasingly lead to Unstable HEALTH would require a, community by community, promoted and nationally sanctioned strategy. Long-term improvement of the cost and quality problems associated with our nation’s healthcare will not be solved without both the immediacy of provider risk-sharing and a substantial improvement of each community’s COMMON GOOD as driven by the community’s level of Social Capital.
    It is time to recognize that the current “levers” of healthcare reform are producing no measurable benefit for either cost or quality and are unlikely to change with “just trying harder.” We can no longer ignore that health spending in 2016 for our national economy was at least $3,000.00 more per citizen within our economy than for any other developed nation’s healthcare AND our citizens experience broadly assessed inferior outcomes. (see “Mirror, Mirror…” report at http://www.commonwealthfund.org). The USA is the only developed nation with a worsening maternal mortality ratio among the developed nations of the world, for the last 25 years. It represents as least 500 citizens who die annually with a pregnancy just because they lived in the wrong nation during their pregnancy.

    see http://www.nationalhealthusa.net/home/rationale/

    The excess health spending in 2016 represent nearly $1 Trillion, the equivalent of paying for 8 or 9 Iraqi/Afghanistan wars in 2005, SIMULTANEOUSLY. The excess cannot be easily figured out. There are many factors that will not be solved without a gatekeeper HMO reimbursement process. Look at the long term progression of healthcare spending, the only time that health spending has not increased faster that economic growth was from 1994-2000, the HMO years. ( see Health Spending reports at http://www.altarum.org ).