Monday, November 21, 2011

Bungling toward bundling: Not so fast, please

If it were not such a serious distraction, the tendency of policy analysts to focus so greatly on “payment reform” -- capitation, global payments, and bundling -- as an answer to increasing health care costs would be all right.  But it is not all right.  It is not all right because the underlying premise behind payment reform is that over-treatment of patients is at the heart of rising health care costs.  While over-treatment exists, it is hard to conclude that it is the area for greatest possible progress in controlling costs.  Accordingly, if we focus on this “remedy,” we will fail to direct attention to other parts of the cost equation, parts that could be more important in bending the cost curve.

What are the determinants of high health care costs in America?

One part of the problem is demographic, a large increase in the number of elderly, who end up needing hospitalization or treatment for chronic illnesses; the arrival of the baby-boomer cohort at the age of hospitalization; and the pending arrival of a sedentary next generation who will suffer obesity and the sequelae of that malnourishment.

A second part of the problem is the lack of effective primary care.  Many people have not had access to primary care.  Those who do have discovered that primary care doctors are often forced into a triage role:  They spend the proverbial “18 minutes” with each patient, an inadequate amount of time, leading to excessive referrals to higher paid specialists.

A third part of the problem is application of “the rule of rescue” in American medicine, the tendency to spend larges amounts of money in high level tertiary and quaternary treatment, well beyond any rationale estimate of the value of a human life.

A fourth problem is that the unit costs of what we provide are high.  Starting with doctors, who need to recover the high cost of medical education in their salaries, to medical equipment and supplies and drugs, to the physical facilities in which care is offered.  Some of these high costs are simply the result of being a high-income country; some are structural in that they reflect regulatory requirements; and some are market-power driven.

A fifth problem is defensive medicine, the tendency by doctors to order unnecessary tests or conduct unnecessary procedures because they fear being sued for malpractice if things go awry.  (Note:  This, not the judgments awarded in courts, is the real cost of our malpractice system.)

A sixth problem is the degree of harm caused by doctors and hospitals.  The number of hospital-acquired infections, for example, is excessive, leading to further morbidity in the hospital setting, along with the associate costs of treating patients for secondary diseases that could have been avoided.

A seventh problem is likely the inefficiency of a multi-payer system, compared to the simpler administrative system that could exist with a single payer.  I say “likely” because there are other inefficiencies often associated with single payer systems, not the least of which is the rationing that results and the emergence of a parallel, private system of insurance and care for those who can afford it.

Finally, an eighth problem is the incentive given in a fee-for-service payment system to over-treat patients, in that a doctor’s income is based to some extent on the number of steps taken, not the results of treatment.

If we were being rational and rigorous about policy prescriptions, we would rank order these causes and determine the costs and benefits of policies that might offset them.  For example, we cannot change demographic patterns, but it could make sense to introduce public health programs to promote exercise and proper nourishment.  We could change the compensation system for primary care doctors so they could spend more time with patients.  We could subsidize physician education so they wouldn’t have to earn so much to pay off loans.  We could reform malpractice laws to reduce defensive medicine. And we could certainly engage in full-scale process improvement training of doctors and implementation of those techniques in hospitals to reduce the extra medical costs associated with harming patients.  (Those of us who have done the latter have demonstrated conclusively the cost savings, not to mention the mortality and morbidity benefits.)

But, our public policy leaders have not done this.  Instead, they assert that pricing-based over-treatment is the key problem, and they offer capitated rate plans and bundled payments as the solution.  If you look closely, you will find that most of those proposals come from payers, either insurance companies who have a corporate desire to shift risk to providers or government officials who are trying to reduce appropriations.  Or from economists, who have a tendency to simplify market behavior and blame everything on pricing regimes.  As I have said, when you have a hammer, everything looks like a nail.

We shouldn’t dismiss a change in the payment system just because it might benefit the insurers or the government, but we also shouldn’t adopt it just for that reason -- or because it fits into economists’ idealized models.  Instead, we should determine how big a portion of the over-treatment problem comes from the payment system versus other causes.  And then we should rigorously review the experience of such regimes and evaluate their costs and benefits.  We should also determine how practical it is to implement a new pricing regime.  For example, let’s look at the business and clinical relationships of the primary, secondary, and tertiary care doctors who will have to jointly share risk.  What would the internalized system of transfer payments look like, and how would it be decided?  It is often the case that these analyses are lacking.

A case in point is offered by Ezekiel Emanuel in the last of series of op-eds he has published in the New York Times.  This one is called, “Saving by the bundle.”  He offers bundled payments for chronic diseases as a partial solution to Medicare cost increases, and the article makes some good points.  But is there rigorous support for what he proposes?  He notes:

For two decades Medicare has been experimenting with bundled payments.  Since 2009, Medicare has been using the Acute Care Episode bundled payment program to cover 37 cardiovascular and orthopedic procedures.  While there has not been a definitive evaluation, preliminary data suggest savings of up to 10 percent and improved quality of care.  Unfortunately, the program does not cover rehabilitation and other post-discharge services.  Worse, it is voluntary and only a few hospitals are participating.

As a public policy recommendation, this would get a “D” in my classroom. Why?  In reverse order, we have a self-selection bias in our sample; an incomplete assessment of all treatment-related costs and benefits; preliminary data not yet subject to peer review; and 20 years of experiences that has not created momentum for change.

The Times editorial staff is also on the bandwagon in an editorial entitled, “Fixing Medicare.”  Endorsing full capitation, they say:

The solution, most experts agree, is to have Medicare pay doctors and other health care providers fixed sums to manage a patient’s care and then let doctors decide which services are truly necessary.

The editors recognize one possible downside of this strategy, but they brush it off in one sentence, offering no consideration of the regulatory and oversight costs involved:

Close monitoring would be needed to ensure that doctors don’t deny medically important services to improve their bottom lines.

(They fail to mention another impediment:  To manage care in this fashion, the patient must be seen by a closed network, whose providers share in the risk pool.  That is all right, but a major issue facing Congress will be whether it wants to take provider choice away from the general body of American elders.)

I think it can be demonstrated that capitation and bundling can work well in certain settings.  For example, treatment of “dual-eligible” people (those on both Medicare and Medicaid) seems to work better and be less expensive when care is managed under a capitated contract arrangement.  But, even there, we have to ask, “Better than what?”  Better than a completely disorganized system of care for poor elderly people who are shunted from provider to provider, often without a primary care doctor to advocate for them.  Still worth doing, for sure, but let’s not extrapolate from this extreme case to the general population.

Being one of those aforementioned economists, I have no aversion to an assumption that price patterns matter.  But when you make policy, you don’t just proceed on that basis.  After all, most of our economy runs on a fee-for-service basis, and we don’t suggest that government intervention is necessary to change that to create more efficiency and higher quality.  With some exceptions, buyers and sellers in those markets prefer a fee-for-service approach, Yes, health care is different.  But that is no excuse to ignore the full range of diseconomies in health care, figure out which ones matter the most, and go from there.  If we do this wrong, we will find out that “payment reform” is not reform at all.


Barry Carol said...

Paul –

While I support both bundled payments for surgical procedures and capitation where appropriate and feasible, I think many of the other potential cost savings strategies including tort reform, more widespread use of living wills and advance directives at the end of life, paying PCP’s more, administrative simplification and fraud mitigation could easily be pursued simultaneously on parallel tracks.

I’m more skeptical about free medical school tuition significantly reducing the income doctors feel that they need to earn because the opportunity cost of what they could alternatively earn in the business or finance world is much higher in the U.S. than it is in Europe or Canada. While they may be perfectly willing to work for half of what they could earn in business or finance, I don’t think they would be willing to work for 10% or 20% or 25% as much.

Uwe Reinhardt has written extensively about the large differences in the cost per service, test, procedure or drug in the U.S. vs. other countries. Drug prices are lower elsewhere because of price controls and very low marginal production costs. However, imaging costs and, especially, hospital based care are much more expensive in the U.S. I’ve never seen a good comparison of U.S. hospital costs vs. similar hospitals (AMC, community hospital, etc.) in Europe or Canada. I would be especially interested in seeing the difference in number of employees per licensed bed broken down by type – nurses, techs, transporters, food service, laundry, IT, management, administrative, etc. Perhaps you or one of your readers knows of and can share such a comparison.

Anonymous said...

This like 90% of all health care analysis is written by a socialist hospital administrator. Your 8th reason is a good start but the rest just express your limited view of economics and dont get to the real underlying factors contributing to ever escalating health care costs.

Paul Levy said...

Please say more.

Unknown said...

Paul: Re: drivers of escalation in health care costs as per your recent blog post. Your list is excellent. Save for demographics, all the issues mentioned result in delivery system inefficiency, and the sum of those inefficiencies are large.

However, you omitted mention of the most important underlying cost driver over the last fifty plus years: technological expansion. Without addressing bio-medical R&D, we will not be addressing the most significant long term cause of healthcare cost inflation. This is not a reference to rationing of available technology; the real issue is intelligently controlling the development of new technologies.

The sooner we get a grip on technological expansion, the sooner we will tame the beast that is nibbling away at our economy. It will be a long discussion, but we need to stop avoiding to subject. It's the elephant in the room.

John Bakke

Paul Levy said...

Thanks, I meant to refer to that indirectly when I talked about unit costs, but you have made it more explicit.

Anonymous said...

As a current medical student, I want to tell you how much I learn from and enjoy reading your analysis. Though I follow the health policy literature closely and involved with health policy groups on my campus, your blog- more frequently than any other source that I read- really challenges my thinking on these issues. Thanks a lot and keep it up!

Paul Levy said...

Thanks, you are kind to say so. In truth, nobody knows what will work in this field, so we are all feeling our way. You'll get a chance, too, in just a few years!

Thomas Pane said...

I agree with every one of these factors. I would add that there is too little direct cost-sharing from health consumers. The recent dip in health spending growth during the recession was attributed to a drop in consumer demand, driven by a combination of factors including job losses. That was a small, albeit temporary, 'bend in the cost curve'.

Instead of using health insurance as a so-called prepaid discount model for all care (routine and emergency), perhaps it is better to transform it to true insurance that covers the big problems, and smaller services are paid for directly. This gets providers competing on service and price, as well as quality.

Insurers are offering more high-deductible plans and health savings accounts, which get consumers more interested in how they spend each health-care dollar.

We can't unwind the current system overnight, but a little tweak on both the supply and demand sides can help.

Mitchell Clionsky said...

I truly enjoyed your comments and believe that you are very accurate in your assessment. I believe that your view of PCP's spending 18 minutes with their patients is more than generous. I think it is really 8-10 in most cases because the reimbursements keep shrinking and demand remains high, particularly in Massachusetts where universal insurance has occurred. I also think that there should be great concern for pitting the doctor's potential earnings in an ACO against the patient's needs for care. The reward scheme will cause doctors to underdiagnose and undertreat and for patients to not even realize that they have a potentially treatable condition. In many chronic illnesses, the under-identification of such conditions will never reach the level of a malpractice-actionable condition but will eventually take years off patients' lives and put extra dollars into the pockets of doctors who control the care. No effective controls have been built into the system as we rush to global payments, much as we bailed out the banks and Wall Street without considering that those same institutions might decide to bonus their executives later the same year.

We need to be spending a lot more on prevention, earlier intervention into chronic diseases, and strategies to improve adherence to medication regimens and a lot less on high tech interventions, such as multiple organ or face transplants.

David Morf said...

Excellent post. I'm a regional economist and functional systems analyst for scaleable processes, now working with others on Wicked Problem topics (multi-participant scenarios, especially healthcare). We focus on Canada and the US. We see US healthcare as likely to benefit by harnessing employers and regional development agencies to support region-wide ACOs that would agree to fund and leverage systemic region-wide primary and therapeutic care linked to public health and aligned with population preventive practices and with remedial health care using technology enabled healthcare analytics, team-based frameworks, patient navigators, care coordinators, coaches, and hand-offs (i.e., not a traditional craft-based medical model for PCPs, dentists, specialists, nurses, hospitals, pharmacists). The goal is a self-funding virtuous circle to boost population health, workforce health, and regional economic competitiveness. I'd welcome your thoughts on my most recent post at Other germane posts from us are on the site. Best regards, David Morf

Anonymous said...

Great stuff but isn't it really two different posts? The first, a list of issues that drives up overall U.S. healthcare costs, -- of which, as you say, payment mechanims is one issue and assuredly not the highest ranking -- is one post. Then your article bridges into a New York Times series about Medicare costs.

I am not suggesting you can split the two subjects but linking them this way hides a major point: the Patient Protection and Affordable Care Act is proposing more ACO/capitation for those under 65 on one hand while moving Medicare subscribers (most of whom are over 65) away from the ACO/capitation approach on the other. I think the same (treating the same heatlhcare issue differently for those under 65 vs. those over 65) for some of the other issues you identify.

-- Dennis Byron

Jonathan said...

I am not sure all of your 8 points are really driving costs. I would not put malpractice law suits and defensive medicine high on the list (not that it doesn't drive up costs, but it is a small percentage. I would put greed on the list, however, both for docs and for hospitals.

I like your point about primary care, although I am not sure that the primary care professionals I would choose would be traffic cops sending patients to other specialists. My view of primary care is that the practitioner would manage most of the health problems of the family, only sending the patient to the specialist when regular treatment was not working, and special expertise was needed. Just because one has a cardiac problem doesn't mean you have to go to the cardiologist. The primary care doc may wish to discuss the case with a cardiologist, however.

Most of the major problems of the elderly that became very expensive to treat, had their prodrome years earlier and could have been prevented by adequate primary care under a reasonable prevention strategy.

The biggest of cost savings, however, are not in the primary care or specialty care area; it is in the managing of hospital costs. Since between 35% and 45% of all costs are generated in the hospital sector, I would stop the DRG process and the fee for services to hospital process. I would simply look at two numbers: number of bed days per year, and total budget for the hospital per year. Than I would divide the former into the later and come up with an average per diem rate to be paid from a central account to the hospital. I would allow for a small overhead margin (say 5%), and than pay the hospital 5% less for the number of bed-days expected in the next year (I would use the average B-Ds over the past 5 years). Most hospitals have about the same mix of patients every year so the overall costs would not change.

A hospital wanting to make some money on this process would conduct the quality management process to reduce costs and improve quality. Since the monthly income of the hospital is based on the costs at the beginning of a two year cycle, reducing the bed-day costs through efficiency, better staffing, lowered nosocomial problems, and more rational use of procedures, will increase the hospital's bottom line. A five or ten percent savings each year removes between 3 to 5% from the aggregate medical cost of the system, or $100 billion to $250 billion annually.

Another $500 billion could be reduced by eliminating the private health insurance companies with their 20 to 25% overhead. And the last $500 billion would be saved by squeezing the drug industry, reducing the amount of chronic disease through good primary care, and managing the frequency of physician ordering of procedures, not punitively, but from an educational perspective.