Tuesday, August 6, 2013

Compassionate, Efficient Health Reform

The following was drafted quite a few months ago, and had its genesis in a list of recommendations for improving the health care system that David Dranove solicited from a number of academics for an issue of Health Management, Policy and Innovation. I've dawdled in finishing and polishing it up, but seeing the stimulating reform proposal posted today by Jay Bhattacharya, Amitabh Chandra, Mike Chernew, Dana Goldman, Anupam Jena, Darius Lakdawalla, Anup Malani and Tom Philipson motivated me to return and finish it; so here it is finally.

Compassionate, Efficient, Health Reform


One can hardly say that there’s been too little discussion of health reform recently. However, much of the discussion is focused on the ACA and its details. That’s fine, but we’ve gotten very far away from thinking about overarching principles that we think should guide the design of a health system, and what that implies for what it would look like.[1] What follows are some thoughts on what such a health reform might look like. They are informed by my read of the research evidence, and my observations of the U.S. health care system over a long period of time, but should be understood as representing only my personal opinions.

This is not intended as a criticism of the ACA. While the ACA certainly isn’t perfect, in my opinion we’re better off as a country with it than without it. However, there will be modifications to the ACA and other changes to the health system as we move forward, so having a framework to structure our thinking will be useful as we consider these inevitable changes.

Guiding Principles

What I propose below is guided by the following. First, economic efficiency is a goal. This simply means avoiding waste, i.e, trying to generate the maximum benefits net of costs. The second goal is that no American is exposed to excessive risk to their health or finances due to medical expenses. Last, the overarching design principle is to create basic ground rules for the system and then let the system run, avoiding heavy handed regulation or micro management. The key objective of these ground rules is to give participants the right incentives insofar as possible, while achieving insurance objectives. With that in mind, compassionate, efficient health reform would do the following.

Health Insurance Reform
      First, eliminate the tax exclusion of employer sponsored health insurance. The exclusion of employer sponsored health insurance from income taxation distorts the demand for insurance. This leads to people with employer sponsored health insurance holding excessive coverage, which drives up medical spending and thus insurance premiums. Ironically, not taxing health insurance ends up making both health care and health insurance less affordable. Eliminating the tax exclusion of employer sponsored health insurance will eliminate a major distortion in health insurance, health care, and labor markets. It can generate substantial tax revenues (it’s estimated that the value of the state and federal income tax exclusion for 2009 was $260 billion[2]), while potentially allowing for lower income tax rates. It’s also worth pointing out that the subsidy is biggest for those who face the highest marginal tax rates, i.e., it’s regressive.

·         Second, automatically enroll every U.S. citizen in a standard, basic health insurance plan. Everyone will be enrolled – there will be universal coverage. Individuals will be randomly assigned to insurance companies, who will be required to cover them. Individuals can opt out of this initial assignment into a different plan, so long as that plan offers at least the standard, basic coverage (it can be more generous, but not less).

·         Third, the plan will provide protection against medical expenses that are catastrophic for the individual or household, given their finances. The function of the plan is to provide insurance against large expenses associated with treating episodes of ill health. It will therefore have a high deductible, and fairly high coinsurance or co-pays, but will have a stop-loss to prevent financial ruin. Preventive care that’s been shown to be effective can be “carved out” and have lower cost-sharing. Cost-sharing features will be on a sliding scale according to income, so individuals only face risk that they can reasonably bear. Low-income individuals will have lower deductibles, coinsurance or co-pays and stop-losses than will high-income individuals. In addition, premiums will be subsidized on a sliding scale according to income, so insurance is affordable for everyone. Insurer premiums will be risk-adjusted, and there will be a high-risk pool. No denials of coverage or coverage rescissions will be allowed. Under this plan ultimately Medicare and Medicaid will be phased out so that everyone will obtain coverage as indicated above.

           The government subsidies for insurance coverage above will be entirely financed via a dedicated consumption (sales or value added) tax, e.g., a la Fuchs and Shoven[3], with as few loopholes as possible. All government funding must only be from this source – no other sources of revenues may be applied. This way the cost and financing of government spending on health care will be as clear and transparent as possible. All other funds will be privately financed.

Supply Side Reform

The main goal of reforms here is enable competition, and to eliminate barriers to entry to providing health services. Lack of competition leads to poor service, poor quality, and high prices, and impedes innovation (especially organizational innovation).[4]
      First, strongly enforce the antitrust laws in health care. There has been a great deal of consolidation in health care markets in recent years, especially in hospital and insurance markets, but also in physician markets and between the different kinds of market participants (e.g., insurers-hospitals, hospitals-physicians, etc.). Consolidation has resulted if few, if any documented benefits, and has harmed competition and led to increased prices, reduced quality, and impeded the emergence of new, innovative forms of health care delivery. Antitrust enforcement can help solve problems in specific markets. It can also have a deterrent effect on those considering anticompetitive actions.
      Second, ease barriers on new forms of health care organizations entering the market, such as retail clinics, freestanding surgery centers, specialty hospitals, telemedicine, etc. In contrast with much of the rest of the U.S. economy, the health care industry has been rigid and unresponsive. New organizational forms that are responsive to patients’ needs are long overdue.
           Third, free up entry into the medical profession. Twice as many people apply to medical school as get accepted, and this has been true for many years. Quite a few more applicants can be accepted without diminishing the quality of medical students. Therefore, artificial barriers to creating new medical schools or expanding the number of slots in existing medical schools need to be eliminated. There has been some recent progress in this domain.[5]
      Fourth, free up entry into specialties. Specialty societies have a great deal of influence on residency training. This creates crazy distortions such as dermatology being the hardest specialty to enter, while primary care specialties have excess training capacity. Artificial barriers to entry into residency training programs should be eliminated.
      Fifth, reduce or eliminate public subsidies to medical education. These only add to the crazy quilt of distortions in this area. With twice as many applicants as accepted students, there is clearly excess demand for medical education. Public subsidies are not only unnecessary, they overwhelmingly go to children from upper middle class or upper class families. Certainly medical training should receive no more in subsidies than training in science or engineering.
      Sixth, allow non-physician medical personnel, such as nurses, nurse practitioners, psychologists, pharmacists, etc. much greater freedom to treat patients independent of physicians. Nurse practitioners and pharmacists (for example) are highly trained medical professionals who can do more than they are currently allowed due to restrictions on scope of practice in many states. Not only can these practitioners substitute for physicians in some cases, they can complement them and thereby enhance productivity.
      Seventh, regulate health insurers nationally, rather than on a state-by-state basis. Insurers currently must operate separate risk pools in every state in which they operate and are regulated differently in every state. This is clearly inefficient. Insurers should be allowed to pool risk nationally and should face one set of nationally agreed upon rules and regulations. This will require a national regulatory body to replace state regulatory agencies.


These ground rules are intended to provide a general framework for the health care system. They are deliberately intended to be general, not specific, in particular so there are incentives for innovative and efficient new arrangements and so such arrangements can spontaneously emerge. While I believe these are sensible changes that would move our health care system in the right direction, there are and will be alternative proposals that are worthy of consideration as paths towards a more efficient, compassionate health care system.

[1] There have been some excellent discussions at a high level, some of which overlap with what I propose here. E.g., Antos, J. Pauly, M.V. and G. Wilensky. (2012) “Bending the Cost Curve through Market-Based Incentives.” New England Journal of Medicine. 367(10): 954-958. http://www.nejm.org/doi/full/10.1056/NEJMsb1207996; Emanuel, E. et al. (2012) “A Systemic Approach to Containing Health Care Spending.” New England Journal of Medicine. 367(10): 949-954. http://www.nejm.org/doi/full/10.1056/NEJMsb120590; Kotlikoff, L. “The Healthcare Fix: Universal Insurance for All Americans.” (2007) Cambridge, MA: MIT Press; Goldhill, D. “The Health Benefits that Cut Your Pay.” New York Times,February 16, 2013. http://www.nytimes.com/2013/02/17/opinion/sunday/the-health-benefits-that-cut-your-pay.html; Christensen, C., Flier, J. and Vijayaraghavan, V. “The Coming Failure of Accountable Care.” The Wall Street Journal, February 18, 2013; Bhattacharya, J., Chandra, A., Chernew, M., Goldman, D., Jena, A., Lakdawalla, D., Malani, A., and T. Philipson (2013). “Best of Both Worlds: Uniting Universal Coverage and Personal Choice in Health Care,” Washington, DC: American Enterprise Institute, http://www.aei.org/files/2013/08/02/-best-of-both-worlds-uniting-universal-coverage-and-personal-choice-in-health-care_105214167938.pdf.
[2] Gruber, J. (2010) The Tax Exclusion for Employer-Sponsored Health Insurance,” National Bureau of Economic Research, Working Paper 15766, http://www.nber.org/papers/w15766.
[4] E.g., Gaynor, M. and Town, R.J. (2012). “The Impact of Hospital Consolidation – Update,” The Synthesis Project, Policy Brief No. 9, Princeton, NJ: The Robert Wood Johnson Foundation, http://www.rwjf.org/content/dam/farm/reports/issue_briefs/2012/rwjf73261; Richman, B.D., Mitchell, W. and Schulman, K.A. (2013). “Organizational Innovation in Health Care,” Health Management, Policy and Innovation, 1(3): 36-44, http://www.hmpi.org/pdf/HMPI%20-%20Richman,%20Mitchell,%20Schulman,%20Organizational%20Innovation%20in%20Healthcare.pdf

Sunday, March 31, 2013

Pittsburgh's Health Care Market -- It's a Pretty Ugly Mess

The Pittsburgh Post-Gazette has an article on the UPMC (hospital system) and Highmark (health insurer) situation in which I and a number of others express opinions. Bottom line -- bilateral monopoly is ugly, and the losers are consumers.

Thursday, March 7, 2013

Can Health Care Transparency Make a Difference?

There's been a lot of discussion of transparency in health care recently, e.g., a USA Today op-ed and a counterpoint by Paul Ginsburg. The appeal of transparency is obvious. As movingly documented by Steven Brill in Time, prices are high and often differ quite substantially, even across close by providers. However, we don't know the prices for the health care that we consume, and it's extremely difficult to find out what these things cost (e.g., this recent study in JAMA).

While the appeal of transparency is obvious, it's important to realize that buying health care is not like buying milk at the grocery store. A key factor is health insurance. Health insurance is very important -- people need to be insured against the catastrophic expenses that can occur with serious illness. Thus people with high health care expenses won't be exposed to most of those expenses (and shouldn't) and therefore will have no reason to respond to information about health care prices.

Further, the distribution of health care expenses among the population is very uneven. Fortunately severe illness is relatively rare. What that means is that most of us will have quite low health care expenses in a given year, while a very small number of people will unfortunately be very sick and have very high health care expenses.  The figure below illustrates this. It shows the distribution of health care spending in the US population for 2009 (source: data brief by the National Institute of Health Care Management). While this is a few years old, the distribution has remained essentially unchanged over a long period of time. The first thing to notice is that the vast majority of spending is accounted for by a small fraction of the population. The top 5 percent of spenders account for about half of all health care spending, and the top 50 percent account for almost 97 percent of spending. 

What this means is that most of the spending on health care in the US is accounted for by people with very high expenses. Further, many (but not all) of these people have expenses so high that they are well beyond the cost sharing features of any health insurance plan, even a high deductible plan that features extensive consumer cost sharing.

Using the query tool for the Medical Expenditure Panel Survey, I found the expenditure levels associated with various percentiles of the 2010 spending distribution. Individuals in the 90th percentile of the spending distribution, who account for almost two-thirds of all health spending, have spending levels of $9,512.76 and higher. These people are almost certainly going to be beyond the cost sharing limits of their health insurance plans. As a consequence, two-thirds of health care spending is unlikely to be responsive to transparency efforts.

Since these people are beyond the cost sharing features of their plans, they have no incentive to pay attention to the costs of care, either in total or cost differences across providers. In addition, people who are spending this much money on health care are unfortunately likely to be quite sick. The demand for care of people who are very ill tends not to be very responsive to prices.

Does this mean that there's no point to transparency efforts? No. Forty nine (.999...) percent of the spending is accounted for by people who spend (just under) $814.86 or less. For these people, cost sharing can make a difference -- large, but conventional, deductibles (e.g., $1,000, $500) would be relevant to many of these people. However, for transparency efforts to work, price information has to be presented to people in a way they can understand and use -- the total cost to them of obtaining care, as opposed to the individual components. Further, there have to be alternative providers. Many health care markets have become extremely consolidated. If there are no good alternatives, then transparency efforts will be of little practical use.

In addition, transparency with regard to the quality of care is both important and valuable to all, regardless of spending level. Providing accurate and informative measures of the quality of care is challenging, but a lot of progress has been made and there are ongoing efforts in this area.

Last, there's some concern that by making prices public transparency efforts will facilitate collusion by providers. If so, this could unintentionally make things worse by reducing rather than increasing competition and leading to higher, rather than lower, prices. There's evidence that this has happened in other industries (e.g., cement), however I don't think it's terribly likely in health care (although not impossible). Even with transparency efforts, health insurers will play a major role as buyers of health care. Insurers' ability to direct large numbers of patients means that providers have large incentives to make deals to obtain those patients. Creating and maintaining collusion becomes difficult under these circumstances.

In sum, trying to achieve greater transparency in health care is a worthwhile effort, but it has to involve usable measures of both price and quality. Further, it's unrealistic to expect consumer shopping alone, and hence transparency efforts, to drive the health care market. Transparency is but one ingredient --- much more is required.

Wednesday, February 27, 2013

Are Price Controls the Answer? Netherlands Edition

This post is co-authored with Misja Mikkers, who is Director of Strategy and Legal Affairs at the Netherlands Healthcare Authority and is affiliated with Tilburg University, the Netherlands and Copenhagen Business School, Denmark.
In a previous post one of us (Gaynor) examined some evidence on whether price controls are effective in slowing the rate of growth of health care spending, and how they compare with competition in private markets. In this post we examine some evidence from the Netherlands that may bear on the matter. In the previous post a data point from the Netherlands was shown as part of an international comparison, but it’s worthwhile to examine the Netherlands experience a bit further. The Netherlands is particularly interesting because they have employed rate setting in health care and subsequently deregulated much of their health sector to allow prices to be market determined. The question of whether the end of rate setting and the introduction of competition raised or lowered health care costs and prices in the Netherlands is hotly debated.

The article by Steven Brill in Time also led also to a lot of discussion in the Netherlands. Rob Wijnberg (former editor of the important Dutch Newspaper NRC-next) tweeted, “brilliant article about the reasons why competition in health care doesn’t work" (https://twitter.com/robwijnberg, Feb 26, translation M. Mikkers). 

Health reform in the Netherlands has been gradual and has had a number of different elements. Initially competition in health insurance was introduced (with an individual mandate), while maintaining rate setting for providers. A partial and gradual deregulation of provider prices followed. For hospitals some services have been deregulated (this is called the B Segment) and some services remain under price controls (the A Segment). The proportion of services in the deregulated B Segment has increased over time. In 2005 8% of hospital services were in the B-segment. This percentage then increased to 20% and 30% in 2008 and 2009 respectively. In 2012 virtually all elective care (70% of hospital services) was in the B segment.

The figure below (source: Market Scan Hospital Market 2011, Netherlands Healthcare Authority [in Dutch]) shows the percentage change in hospital prices over time in the price controlled A Segment and deregulated B Segment. As can be seen, growth in the deregulated segment, where prices are market determined, is substantially lower from about 2006-2007 onwards. In fact, from about 2008-2009 to 2010-2011 prices were falling in the deregulated segment while they were still growing in the price controlled segment. This doesn’t necessarily mean that competition controls prices and rate setting does not (lots of other things could be going on), but neither do we see what we’d expect if rate setting was doing a superior job of controlling prices.
The next figure (below, source: CBS [Dutch Central Bureau of Statistics] ) shows the growth rate in hospital spending in the Netherlands in the period before rate deregulation (2001-2005) and after deregulation began (2006-2011). The average annual growth rates between the two periods are virtually indistinguishable. Again, this isn’t scientific proof that rate setting doesn’t control costs (either in general or in the Netherlands), but there’s no slam dunk for rate setting in the patterns that we observe. We may be able to see more interesting patterns in the future. By 2014 and 2015 virtually all prices for elective care will have been deregulated and all parties in the market will be fully exposed to the consequences of their (price) negotiations.  

In sum, the Netherlands is a good place to look for the effects of rate setting versus markets on prices and spending, since they have employed both. A quick look at some descriptive statistics doesn’t yield any slam dunks for rate setting. If anything, shifting to markets may have substantially reduced price growth. However, careful study will be required in order to draw firmer conclusions. Last, while we believe that there are lessons for the US from experiences in other countries (and vice versa), we do have to be cautious in making strong inferences across very different health care systems and societies.