President Obama has announced that he is in favor of paying for doctors and hospitals based on "value." This would be in contrast to the usual way of paying for physician services, on the basis of fee-for-service. The President says by rewarding "value" the government can avoid paying for "quantity," discouraging doctors and hospitals from doing things that we don't need. The President is about to step off a big cliff, one which will repeat the unfortunate history of the "sustainable" growth rate or SGR.
Here are some of the President's remarks on signing H.R. 2, the Medicare Access and CHIP Reauthorization Act, April 16, which repealed the Sustainable Growth Rate, our last big government experiment in changing the way physicians are paid:
For the last 13 years, we have been confronted with what's called a "doc fix." Basically, the way the law was written, to deal with Medicaid payments to doctors, there was always the danger each year that suddenly, arbitrarily, doctors' payments would get cut off.
The conflation of Medicare (and its "sustainable growth rate") and Medicaid (not involved in the SGR) is not uncommon among, for example, speechwriters and junior staffers. Does that matter? It depends on whether you want the leader of the free world making payment policies for physicians based on only the most general understanding of the issues.
"The danger each year that suddenly, arbitrarily, doctors' payments would get cut off..."
Where did this annual danger come from? An alien force? A foreign power? You might think so, right? No, it came from the Clinton administration, distracted, and a dopey theory that, if the volume of Medicare payments to doctors increased (more Medicare beneficiaries, more patients, more services), the unit payment should go down.
Of course, we don't pay our grocers this way -- you sell more milk this year, you are paid more. Nor is this theory found anywhere else in a free society. The "sustainable" part of the sustainable growth rate was really only meant to be "sustainable" for the government, not for the patients or their doctors.
The practice of medicine not yet being bonded indenture, some of the doctors periodically threatened to bolt, and their patients, a demographic that votes, noticed. So haste was made:
"Because we wanted to make sure doctors' payments didn't cut off, I'm signing it now rather than wait for getting members of Congress down here."
Well, the "fix" only took 18 years, 1997-2015.
In any event, the quick fix also heralded the beginning of the next problem. Said the President:
Because what it [the other parts of the "fix" legislation] starts doing is encouraging payments based on quality and not the number of tests that are provided or the number of procedures that are applied, but whether or not people actually start feeling better.
So we enter the era of value-based payment.
What is value-based payment? It is the latest evidence-free slogan, building on the (now discredited) "pay-for-performance" slogan and the "value-based purchasing" slogan. If successful in medical services, we might pay the grocer based on our nutritional status, not on the volume of milk sold to us, and pay the mechanic based on where the car has taken us.
Pay-for-performance was the latest thing 10 years ago. All the authorities were announcing that we should only "pay-for-performance," not for individual professional or institutional services.
Value-based purchasing was another government invention, beginning in 2005.
Both of these depend on "incentives" (additional payment) and "disincentives" (payment penalties, or, as CMS has decided, "adjustments.") We use these tools notwithstanding the difficulty of relying on the measurements behind the incentives and adjustments. Are the measures accurate? Do they mean anything to patients? Don't we know that (see Eduardo Porter on Goodhart's Law, in the New York Times) if you pay people according to a particular measure, you will get a lot of what you're measuring, but won't know what its real "value" might be. Value to whom? Based on what?
American health care has suffered from this "policy by slogan" for much of its modern (post-Medicare) era. Successive waves of regulation (health planning and certificate of need; rate and budget regulation; now incentives and disincentives associated with provider behavior) have been introduced and, for the most part, failed, at least in their stated efforts to constrain concentration of provider power, the resulting increase in health care prices, and the impact of those prices on state (through Medicaid) and federal (through Medicare) budgets.
Also, health policy has, as a result of programs associated with these slogans or initiatives, become increasingly complex. A reasonable hypothesis is that these trends have gone hand in hand: that complexity (at the institutional, third party, state and federal levels) has increased the price of health care, that the larger, more powerful and more expensive institutions have been most adept at managing that complexity, and that, as a result, we have available fewer community-based providers and, in some parts of the country, almost no private practice physicians, much less ones willing or able to grapple with splitting the red tape lengthwise.
So while these slogans may be "content-free," they are not consequence free. The slogans attract the expectations, desires and perhaps the business plans of onlookers. Since the impact of the slogan will be on the expenditure of nearly one trillion dollars per year in public funds, precision is warranted.
Even informed commentators may have their work confused with headlines having nothing to do with "pay for performance" per se, but rather conflated and mixed in with other (also largely evidence-free) slogans, including the impact of "accountable care" organizations, "delivery system reform," and the predecessors of value-based payment, pay-for-performance and value-based purchasing.
(See for clarity Uwe Reinhardt on metrics set centrally, vs. the professionalism of educators and health professionals, "The idea that everyone's professionalism and everyone's good will has to be bought with tips is bizarre.")
It is beyond the scope of these comments -- but should, at the same time, be considered -- that the latest iteration of "health policy by slogan" stems from the enormous influence of behavioral economics on modern American society, and especially on social policy. When we encounter the misadventures of, for example, Chinese bureaucrats , in their exercise of incentives, bringing about the very result they hope to avoid, we may dismiss their results, assuming ours will be better.
However, we don't need to look far to see that our national experiments with incentives and disincentives have had a mixed history. For example, a "metric" given to the Veterans Administration hospital in Phoenix, Arizona was to cut down the wait time for individual Veterans to see physicians. Employees who were pressured to achieve this metric, who knew that no additional physician time was available, found "work-arounds" that, inadvertently, made the situation facing the Veterans worse. General Shinseki, notwithstanding his distinguished career, was the sacrifice at the time, but blame should have been shared by behavioral economists whose theories lay behind the appointment "metric."
A parallel controversy may be seen in the current debate over educational testing. In fact, the very idea of value-based payment (pay for the patient experience or outcome, not for the professional service) is remarkably similar to proposals that would link payment of teachers to the results of tests given to children (pay for the outcome for the child, not the service delivered by the professional).
At least two major issues link these proposals (value-based payment and teacher compensation based on testing). The first is the frustration of policy-makers with relatively poor performance of some students in some schools. This frustration does not take into account the impact of economics, the role of parents, the activities of peers, other (non-school) educational and recreational opportunities, etc., and, instead, judges the teacher to have failed if the child does not progress.
Likewise, the literature of value-based payment in health services vilifies payment for "quantity" of service rather than (in the mind of the payer) "quality," notwithstanding that all other parts of American society reward increased productivity (quantity of output, in a given time period).
A second theme, more darkly, may lay behind political enthusiasm for value-based payment in education and in health services. That theme is the indisposition of some public officials to fulfill the promises (to professionals in education and in health care) made by their predecessors. The cost of paying for professionals, in other words, grows, in contrast to static or declining enthusiasm for payment.
In The Prince, Machiavelli counsels:
What physicians say about consumptive diseases is also true of this matter, namely, that at the beginning of the illness, it is easy to treat but difficult to diagnose, but, if it has not been diagnosed and treated at an early stage, as time passes it becomes easy to diagnose but difficult to treat ... [The Romans] because they perceived troubles when they were merely brewing, were always able to overcome them.
The time, in other words, to stop an unfortunate new (and predictably expensive and complex) initiative in the theory of paying physicians may be now, especially when physicians have fresh in mind the 18 year nonsense associated with a "sustainable growth rate."
The Measurement of Quality
Proposals to link reimbursement to "value" presume a consensus on these questions:
(1) How do we measure quality?
(2) Who does the measurement?
(3) What do we do when we discover that our measurements are flawed?
There is, to the contrary, no apparent consensus on these questions. Rather, the measurement of "quality" and of "value" is an evolving area.
For example, we have proceeded for some years under the assumption that the readmission of a patient to a hospital is "bad," that is, to be avoided. Penalties have been in place since 2012 for hospitals seen as readmitting an excessive number of Medicare patients in defined diagnostic categories within a 30-day period.
Moreover, there has been an assumption that hospitals with higher volume will have "higher quality" and will therefore have lower readmission rates. There is no support for any of this in the literature. To the contrary, hospitals with the highest volume of patient admissions had the highest readmission rates in one recent comprehensive study, whereas those with the lowest volumes had the lowest readmission rates.
What would the explanation be for that? The authors indicate that "This finding suggests that smaller medical centers may provide higher quality transitional care than larger centers." There is no evidence for that, either. Alternative hypotheses abound: that smaller hospitals and smaller medical centers have more short stay admissions, fewer complex cases. Or, in the alternative, that smaller hospitals may be found in suburban and rural areas (the academic medical centers having consolidated in the larger urban areas), with tighter social structure, family and community support, etc.
The problem is obvious: If we penalize hospitals for failing to provide "value" we risk disrupting necessary services, a disruption based on highly contested and controversial evidence.
History of Value-Based Payment
The first brand of value-based payment in health services was "pay for performance," begun in the U.S. in 2003. As frequently observed , it was "one of those slogans that seemed to upset no one. To most people it's a no-brainer that we should pay for quality and not quantity." This author went on to note the disappointing results from pay for performance: "Sometimes it's because providers don't change the way they practice medicine; sometimes it's because even when they do, outcomes don't really improve."
And, of course, it is also possible that the changes that are made (albeit for short-terms, with small impact) have nothing to do with the interests of individual patients.
Another study indicated the short-term nature of any results found in this quest. The Hospital Quality Incentive Demonstration in England in 2008 found that short-term relative reductions in mortality for conditions linked to financial incentives in hospitals participating in pay for performance programs were not maintained.
Another English program, begun in 2004, the Quality and Outcomes Framework (the world's largest pay for performance experiment), had fully a quarter of the income of family practitioners linked to performance. Early into the program, researchers focused on chronic disease management (in this case hypertension), and noted that "Pay for performance had no discernible effects on processes of care or on hypertension related clinical outcomes. Generous financial incentives, as designed in the UK pay for performance policy, may not be sufficient to improve quality of care and outcomes for hypertension and other common chronic conditions."
The largest hospital-based pay for performance program in the U.S. showed no evidence of decreasing 30-day mortality. The authors concluded that "Expectations of improved outcomes for programs modeled after [the] Premier HQID [Hospital Quality Incentive Demonstration] should therefore remain modest."
So while it is clear that financial incentives change some physician behavior, whatever change takes place is only arguably related to positive outcome for the patient, and doesn't last, in any event. The cost of such experiments, aside from the cost of the program and its appurtenances, included a "loss of autonomy and of professionalism and becoming less skilled in dealing with certain conditions," and of course the offsets, where family practitioners who "welcomed the initial pay increase... [then] began to resent the program as successive governments clawed back the initial large increases with a succession of below inflation-raises."
Value-based purchasing, the second generation of pay for performance, has also proven to be problematic.
First, the "values" which were "purchased" changed over time. From its beginning in 2005, to its current incentive and penalty phases, the value based purchasing program has had a changing constellation of variables measured, and now rewarded or penalized.
Second, the "values" proved highly susceptible to influence that would not necessarily have been related to "quality" metrics. For example, hospitality and "guest relations" programs to raise the scores on surveys given to Medicare beneficiaries no doubt improved the patient experience, but may have only arguably been related to any aspect of the "quality" of the medical care.
Third, safety net hospitals fared poorly. Safety net hospitals were held accountable for outcomes which may have been (but probably were not) within the influence of those hospitals.
Just as the pay for performance experiments were shown to have little impact, or were of no lasting value, non-payment (for "mistakes") has also been a "non-starter." In October of 2008 the Centers for Medicare and Medicaid Services stopped paying for hospital-acquired conditions that were deemed (by them) "preventable." Four years later, a research effort reported that
"We found no evidence that the 2008 CMS policy to reduce payments for central catheter-associated bloodstream infections and catheter-associated urinary tract infections had any measurable effect on infection rates in U.S. hospitals."
None of this is to argue that we shouldn't strive to eliminate unnecessary infections, or other shortcomings that have brought alarm and concern to the question of patient safety in U.S. hospitals.
To the contrary, it illustrates that focus on financial punishment (non-payment) has, at least so far, had little impact.
Nor is this to pretend that the question is "settled" for the long run. Rather, the purpose of these comments at this time is to raise doubts that any wholesale press toward "value-based payment" - - with its inevitable incentives and disincentives -- would be based on evidence, or indeed, anything other than a capacity to influence the rules.
CMS should, for the moment, eschew and avoid any embrace of value-based payment methodologies. This is unlikely to happen, unless directed by Congress.
CMS should retain independent academic researchers, not otherwise related to CMS, to conduct studies of any proposed physician reimbursement change, using small pilot programs, double-blind and randomized intervention, with transparent reporting. This is also unlikely to happen. The integration of research (funded by one or another of the federal agencies), announcement of results (not vetted, often opaque), and new regulation and reimbursement schemes (most without Congressional input or oversight) is nearly seamless.
CMS should clarify and define its terms, and demonstrate the evidence behind its use of and argument for those terms. Nearly the entirety of the current CMS proposal for value-based payment is replete with vaguely defined terms, subject to highly variable interpretation.
CMS should refrain from conflating fraudulent services with those that are not "valuable." If in fact it is CMS' desire is to ensure that it is not paying for "quantity" of services that are inappropriate, the answer is to bolster its defenses against fraud, that is, the provision of services which are not medically necessary. Evidence of services not medically necessary should be pursued with fully as much vigor as any of these "reform" proposals. If the services are medically necessary, on the other hand, why should CMS pay less to the professionals and institutions that provide those services?