Something Not Unlike Research

Kenneth Arrow and the Overdemandingness Argument for Discounting Future Health Benefits

post by Paul Kelleher

There are a number of different arguments in the economics literature for assigning less value to future benefits than to present ones (this includes health benefits). This practice is called discounting, and standard economnic analyses of programs that produce benefits in the future will discount those benefits at some positive rate. For example, a cost-benefit analysis using a discount rate of 3.0% will say that 1.3 expected lives saved in 10 years are equivalent in value to 1 life saved today. 

One common argument in defense of discounting is what I shall call the Overdemandingness Argument. This argument takes the form of a reductio ad absurdum. Here is an example of it in the literature (I give other examples in a paper on discounting out just today):

There has been considerable debate about the value of [the discount rate] since Ramsey. He argues that it is ethically indefensible to use a positive value, as this would discount future generations’ well-being relative to the present one. However, Arrow shows that weighting all generations’ welfare equally results in very high rates of savings being required of the current (or even of every) generation. He demonstrates that, under reasonable parameter values, the current generation could be required to save approximately two-thirds of its income! To avoid this result, a positive pure rate of time preference should be employed. (Moore et al. 2004, p. 804)

As stated here, the Overdemandingness Argument is woefully incomplete. For it relies on a crucial, unstated premise, namely Premise 3 in the following more complete version of the argument:

1. Assume a discount rate of 0%. (This means a benefit provided today will be not be valued more than an exactly similar benefit provided at any point in the future.)
2. So long as the current generation’s consumption is not perilously close to zero, a dollar productively invested today on behalf of future generations will enable them to derive more “raw” (i.e. undiscounted) benefit than that dollar could provide if consumed by the current generation.
3. We have a moral obligation to maximize total net benefits across time.
4. Therefore, the current generation has an obligation to sacrifice its own consumption for investment in behalf of future generations.
5. But Conclusion 4 is absurd, so there must be something wrong with the argument that leads to it.
6. The only plausible weakness is Premise 1.
7. Therefore, Premise 1 is false, and a positive discount rate should be used to convey that future benefits are less valuable than present ones.

Premise 3--the premise that Moore et al. left unstated--could be rewritten thus: we have a moral obligation to pursue efficiency singlemindedly. This view is called consequentialism in moral philosophy, and utilitarianism is its most well-known variant.  If the Overdemandingness Argument does not contain a premise about consequentialism, it is incomplete. Once that assumption is added, the argument is complete, and premises 1-3 do logically entail the ostensibly problematic Conclusion 4. But in order for the Overdemandingness Argument to be a successful argument in favor of discounting, Premise 1 must be the only possible problematic premise. And it is precisely because they leave Premise 3 unstated that the economists I quoted above view Premise 1 as entirely responsible for the problematic Conclusion 4. However, someone who rejects that conclusion may now choose to reject consequentialism (Premise 3) instead. The Overdemandingness Argument's case against Premise 1 is therefore significantly weakened.

Despite being cited favorably by the economists quoted above, Kenneth Arrow saw very clearly that the assumption of consequentialism is doing essential work in Overdemandingness Argument. As he writes in "Discounting, Morality, and Gaming" (pdf),

But with zero time preference [i.e. no discounting] and a long horizon, the savings become inordinately high, possibly approaching one as the horizon goes to infinity. A reconciliation must be based on the notion that individuals are not morally required to subscribe fully to morality at any cost to themselves. (p. 13, emphasis added)

The substance of this observation is correct, but Arrow’s wording is not innocuous. He assumes here that morality is consequentialist, and thus that any deviation from consequentialism is a deviation from morality. However, many who reject consequentialism claim that morality itself permits deviations from consequentialism’s dictates. Indeed, Arrow at one point suggests precisely that:

I find this to be an incredible and unacceptable strain on the present generation...I therefore conclude that the strong ethical requirement that all generations be treated alike, itself reasonable, contradicts a very strong intuition that it is not morally acceptable to demand excessively high savings rates of any one generation, or even of every generation. (p. 16, italics added)

If it is indeed a substantive question whether morality is consequentialist or not, then this will be a question about which many economists, qua economist, wish to remain neutral. Often, this will be because economists are not formally trained to evaluate answers to this philosophical question. But even if they, like Arrow, are willing and able to make philosophical headway on the question, it is doubtful that economists, qua economist, ever need to do this. To see why, consider the following remark by Arrow and colleagues in their chapter on discounting in the 1995 report of the Intergovernmental Panel on Climate Change (pdf):

Economists are in general agreement that cost-benefit analysis, including discounting, is useful in examining policies with long or complex time paths, or policies whose effects extend across generations. At the same time, cost-benefit analysis, and the techniques that go with it, including discounting, focus on economic efficiency, and therefore have limitations as a guide to policy. (p. 130, emphasis added)

If economists are to focus on efficiency only, then it is not clear why the technique of discounting should be a part of the economist’s repertoire in the first place. After all, one can focus exclusively on efficiency (e.g. investigate its nature, make recommendations for its measurement, rank policies with respect to it, etc.) without recommending the singleminded pursuit of it. But, as we have seen, it is only the singleminded pursuit of efficiency that could generate the worry that current generations will be crushed under the weight of a duty to maximize future well-being. Since economists, qua economist, do not recommend the singleminded pursuit of efficiency, I conclude that they have no “overdemandingness”-related reason to discount the raw results of their efficiency-focused inquiry.

Of course, this does not show that discounting the value of future benefits is wrong. It just shows that one common reason for doing so is unsound.

A Clue to What Weinstein and Stason Were Thinking

post by Paul Kelleher

This is a follow-up to my recent post analyzing one argument for uniform time-discounting in cost-effectiveness analysis.

As I explained, Weinstein & Stason's (W&S) argument relies on a comparison of five hypothetical health interventions. 

(Key: $10’ = $10,000, $70’ = $70,000; “LY” means “life-year saved/generated”)

Their most controversial premise is that programs A1 and A2 are "equivalent." This is controversial because W&S also assume that programs A and A1 are equivalent. However, A and A1 can be equivalent only if $10,000 today is worth $70,000 in 40 years. But then if $10,000 today is equivalent to $70,000 in 40 years, we should conclude that program A1 is superior to A2, since A1 gets us one life-year for $60,000 less than A2 (we can just put $10,000 in the bank now and wait until the account grows to contain $70,000 in 40 years). This is why it looks absurd to say, as W&S do, that A1 and A2 are equivalent.

Now, I was trying to be charitable in my first post. So instead of claiming that W&S were simply wrong to claim that A1 and A2 are equivalent, I introduced a notion of equivalence on which A1 is indeed equivalent to A2. I called that static-equivalence, and it was meant to convey that both A1 and A2 involve buying 1 LY for $70,000 at the time the LY can actually be purchased. For A1, that is in 40 years. For A2, it is now. I then went on to note that programs A and A1 are not equivalent in that same sense. Rather, A and A1 are dynamically-equivalent, since their equivalence relies on the fact that productive forces can transform $10,000 into $70,000 over time. Finally, since W&S wanted to create a "chain of equivalences" running from program A through to program A3, I claimed that this chain cannot even get started, since there are two completely different notions of equivalence at work here.

I still stand behind this analysis, and I still think W&S's argument is plainly invalid. But I was always unsatisfied that I couldn't explain why W&S (or anyone!) would assert that A1 and A2 are equivalent. I now I think I may have the answer. (The explanation I'm about to give is also suggested by Erik Nord in the paper of his linked to in my first post, but I didn't quite see this until my second or third reading of it).

To understand where W&S may be coming from, it helps to return to the Washington Panel Report chapter on discounting. The "W" of W&S (Weinstein) was a co-author of that chapter. After restating W&S's argument, the Panel writes (my emphasis): "Closely related to [W&S's] consistency argument is the paradox of Keeler and Cretin...The following simple example...summarizes the essence of the argument:"

Assume that an investment of $100 today would result in saving 10 lives (or 1 life per $10 of investment). If the $100 were invested at a 10% rate of return, in 1 year it would be worth $110; and with this $110, it would be possible to save 11 lives. If the original $100 were invested for 2 years at 10%, it would be worth $121, and 12 lives could be saved. If the social decision maker is attempting to maximize the health output obtainable from the original $100, and if the value of future lives saved is not discounted [at the same 10% rate], then the cost-effectiveness of the investment would be...postponed "until next year" on a perpetual basis.

The Panel's example highlights two potential problems. The first is that decision makers may be tempted to hold onto allocated health funds indefinitely, since it always seems a good idea to save the money for one more year. The second problem arises even if we avoid the first by establishing an end-date by which the health funds must be spent. Even with an end-date, decision makers will still prioritize those who need help at end of the imposed time-window, since more lives can be saved by waiting as long as possible (assuming, as is being done here, that the nominal cost of saving a life does not increase over time). Thus the Washington Panel concludes:

Hence, setting [the discount rate for costs equal to the discount rate for health] creates what economists term horizontal equity among potential beneficiaries.

In other words, uniform discounting removes the basis for systematically prioritizing future beneficiaries over present ones. If the amount of good one can do declines at the same rate as the real costs of saving those lives, then there is no incentive to wait to save lives. 

I think this may provide the best clue to what W&S must've been thinking. Given that the Washington Panel Report was co-written by Weinstein, and given that the Report says that W&S's logic is "closely related" to the logic used to avoid the "postponement problem," some sense can now be made of W&S's claim that A1 and A2 are equivalent: W&S may have wanted to rule out the perceived intertemporal inequity that exists when there is systematic reason to postpone health expenditures. To say that A1 is superior to A2 is to invite "horizontal inequity."

I was aware of the so-called "postponement problem" when I wrote my previous post. But I was not aware that W&S's argument might be influenced by it. If this new explanation of their argument is correct, I can give the response to W&S that I'm inclined to give to the postponement problem itself: Since cost-effectiveness analysis is intended to be a framework for evaluating efficiency only, economists should not build equity-related considerations into their calculations of efficiency. If one is interested solely in quantifying efficiency (as cost-effectiveness analysis is intended to be), why not just admit that more good can be done by waiting until one's health budget is as large as possible? If we think it's wrong to wait that long, this reveals that we have concerns other than efficiency. Big surprise.

At the very least, the Washington Panel Report can be commended for being open about letting equity-based considerations shape its ostensibly efficiency-based economic framework. 

A Critique of Weinstein & Stason’s Consistency Argument for Time-Discounting (wonky)

post by Paul Kelleher

UPDATE: I have a follow-up post here that offers a hypothesis as to why W&S accept what most others believe is a patently false premise (i.e. that programs A1 and A2 are equivalent). 

In 1996, an expert panel appointed by the U.S. Public Health Service released a report on improving the methods of cost-effectiveness analysis in health policy. That report, known in the biz as the Washington Panel Report, included a chapter on discounting the health benefits produced by medical interventions. Discounting is a super tricky issue that raises questions of ethics and economics, which is why I like it. It is also a super important issue, as I hope will be made plain by this post.

The Washington Panel recommended (1) that both the future costs of health promoting policies and their future benefits should be discounted, and (2) that costs and benefits should be discounted at the same rate. What does this mean? Put simply, to discount is to treat as less valuable. So consider costs first. The main reason for treating a dollar in the future as less valuable than a dollar today is that today’s dollar can be invested productively to yield more than one dollar in the future (even after adjusting for inflation). This means a cost of 100 (inflation-adjusted) dollars borne in 20 years can be covered by an investment of less than $100 today. Hence, future costs should be discounted relative to present costs.

So that’s discounting costs. What does it mean to discount health benefits that lie in the future?

Understanding DALYs

post by Paul Kelleher

Cost-effectiveness analysis (CEA) compares, you guessed it, costs with benefits. Costs are typically conveyed in monetary terms. Benefits are usually conveyed in terms of either QALYs or DALYs. QALYs are quality-adjusted life years. DALYs are disability-adjusted life years. QALYs have been around since the 1970s (as I understand it), whereas DALYs are a more recent innovation. In 1994, Christopher J. L. Murray, then of Harvard University, introduced DALYs in his paper "Quantifying the Burden of Disease." Murray then went on to the World Health Organization (WHO), where DALYs were used in the controversial attempt by WHO to rank the equity and effectiveness of the world's national health systems. Both QALYs and DALYs are used widely today as the units of benefit-measurement in cost-effectiveness analyses.

This week I've spent a lot of time reading and thinking about CEA, and I've encountered quite a bit of confusion in the literature on how DALYs should be understood. For example, I was looking at a paper entitled "The Economics of HIV/AIDS in Low-Income Countries: The Case for Prevention" which reports CEA results in terms of the cost per "loss of DALY averted." This implies that, whatever DALYs are, you should want more of them rather than fewer (i.e. you want to avert their loss). This understanding of DALYs is made explicit in "Using Cost-Effectiveness Analysis for Setting Health Priorities," a report by the Disease Control Priorities Project (which appears affiliated with the World Bank and WHO): 

In cost-effectiveness analysis, the DALY represents the number of years of disability-free life that would be gained from a particular health intervention—yielding a cost per DALY where cost data are available or can be inferred. Gaining a DALY through a health intervention reduces the burden of disease; it is the same as averting the loss of a DALY. 

Compare this understanding of DALYs with that stated in Disease Control Priorities in Developing Countries by Dean T. Jamison, who worked on WHO's Global Burden of Disease Project and continues this work at the at the University of Washington's Institute for Health Metrics and Evaluation (which Chris Murray now directs):

Disability-adjusted life years (DALYs) are a variant of QALYs in that they measure the losses from disability or premature death; a CEA will determine which interventions maximize QALYs or minimize DALYs.

So which understanding of DALYs is correct? Do we want to maximize DALYs (i.e. avert their loss) or do we want to minimize them (i.e. praise their loss)? The answer is that Jamison is right: DALYs are to be minimized (and their loss praised). But how could those others, who work with DALYs all the time, get it so wrong? Let me try to explain.

Both QALYs and DALYs are designed to integrate considerations of longevity and quality-of-life. For example, I generate 1 QALY for every year I live in full health. The number of QALYs one generates is calculated by multiplying the length of time (measured in years) by a "utility weight" which places one's health status on a scale of 0 to 1 where 0 is as bad as death and 1 is full health. So if I live 2 years in a health state that has a utility weighting of 0.8 (perhaps I have mild lower back pain), then I am said to generate 1.6 QALYs (2 x 0.8) over that time span. Given this understanding, one obviously wants more QALYs rather than fewer, since more QALYs correlates with more years lived in good health.

QALYs are therefore used to measure healthy years lived. DALYs, by contrast, are used to measure healthy years lost. That is, DALYs measure the gap between the number of healthy years one in fact lives and some normative ideal of how many healthy years equates to full health over a lifetime. The normative ideal of longevity is set using the world's longest actual life-span, which is set by the Japanese. The normative ideal of health-related quality-of-life is a trickier subject that I want to pass over here. The important thing is that it is stipulated by DALY-users that full health for each person is a certain number of healthy years of life. So when we suffer ill health and disability, that is regarded as stripping us of those healthy years. DALYs are used to measure the healthy years lost due to ill health.  

The confusion about DALYs enters when people treat DALYs as the healthy years that are lost to ill health. These folks then assume that successful health interventions work by clawing back those lost healthy years (or by "averting their loss" in the first place). Strictly speaking, however, DALYs don't get clawed back when health improves---they disappear or are obliterated. Thus, when one is in poor health, DALYs come into existence and displace one's stock of healthy life years (as set by the normative standard); but when one's health improves, one's newly regained healthy years of life displace or obliterate the DALYs that poor health created. Hence Jamison's dictate: minimize DALYs. One minimizes DALYs either by preventing the poor health that creates them in the first place, or by eliminating and substituting them with regained healthy years of life.

I hope that helps.

 

The epigenetic lottery

post by Bill Gardner

Tyler Cowen discusses the (possibly diminished) US social mobility, and asks

How much of [social] immobility is due to “inherited talent plus diminishing role for random circumstance”?  Is not this cause of immobility very different — both practically and morally — from such factors as discrimination, bad schools, occupational licensing, etc.?  What are you supposed to get when you combine genetics with meritocracy?  I do not know how much of current American (or other) immobility is due to this factor, but I find it discomforting that complaints about mobility are so infrequently accompanied by an analysis of this topic.

So, the idea is that nature roles the dice and gives each child a set of genetic endowments. Families will also invest their resources in cultivating the children's talents. Then, as we become adults, 'society' (meaning the schools, job markets, etc.) places each of us in the social hierarchies of wealth, income, and power (the figure to the right is from Catherine Mulbrandon [@visualecon], and shows the 2005 income distribution among the bottom 98% -- adding the next 2% takes you way off the page). If societies are increasingly meritocratic, those outcomes will mirror the individual distribution of genetic talent and family resources to an increasing degree. So it could be that reduced mobility and increasing social inequality of outcomes reflects increasingly accurate meritocratic selection processes. And many would argue that such processes are not inherently unjust, so maybe we should just get over being upset about the worsening inequality of outcomes. If you add the ideas that genetic endowments are exogenous, and that trying to compensate for the inequality of family resources is either wrong or futile, then you have even stronger justifications for current inequality.

I don't buy these conclusions, for lots of reasons, but this post is not the place to provide the discussion that Cowen is calling for.

I want to raise just one objection here: The story about the genetic lottery is too simple. Our biology is less exogenous than most theorists believe. Michael Rutter explains that genes do not give us talents:

...although the DNA is what a person inherits, the effects of that DNA are dependent on a chain reaction that ends up with the expression of the products of the DNA in the structure and function of individual cells that is how environments alter genes (emphasis added).

Referring to this account of the interaction of nature and nurture, James Heckman writes

The field of epigenetics surveyed in Rutter (2006) demonstrates how genetic expression is strongly influenced by environmental influences and that environmental effects on gene expression can be inherited (emphasis added).

He then argues that there is evidence

that high quality early childhood interventions foster abilities and that inequality can be attacked at its source.  

and we are therefore not absolved from responsibility for the inequitable distribution of adult talent visible in infancy and early childhood.

The problem I'm pointing to in the meritocratic argument is that there isn't really a genetic lottery  in the way we imagine: and to the degree there is, the dice are loaded in the womb. If Heckman is right, there is much more we can do about inequality than most of us are willing to acknowledge.

However, if Heckman is right, it is also the case that much if not most of what we can do is equalize the epigenetic and child developmental lotteries that will govern inequality among future generations. And we have to decide whether we are motivated to struggle for equality in that future, a promised land that we will not see.

'Health Shock': How childhood chronic illness affects children's schooling and adult employment

Pediatrics is sometimes treated as the least important branch of medicine. This kind of makes sense, because children are mostly healthy. Focusing just on health outcomes, however, is terribly short-sighted. As Jason Fletcher and Michael Richards show, chronic diseases that begin in childhood have powerful effects not only on children's adult health, but also on their non-medical development and achievements. 

Fletcher and Richards looked at the association between the 'health shock' of  getting diabetes in childhood and achievements in schooling and work for more than 15,000 children. They found

effects in several measures of educational attainment, including a high school dropout rate that was six percentage points higher than among young adults without the disease. We also found lower employment and wages: A person with diabetes can conservatively expect to lose more than $160,000 over his or her working life, compared to a peer without the disease. 

This is an observational study, so whether diabetes causes these effects is a matter of inference and interpretation. Fletcher and Richards used a data set that included detailed information on the children's family backgrounds, the health of parents, and other diseases that children had; and their estimates include statistical controls for such factors. We cannot necessarily infer that an intervention that prevented diabetes in a child would increase the years of schooling they would achieve, or how much they would earn as adults. But based on these data, it would be worth a shot.

Thinking, Fast and Slow about Informed Patient Choice

The following is an invited post from Thom Walsh PhD, MSPT, OCS, Diplomat MDT (@thomwalsh on twitter).  Thom very recently completed his doctorate in health policy at The Dartmouth Institute for Health Policy and Clinical Practice.  He is currently on faculty as a curriculum specialist in The Dartmouth Center's Health Care Delivery Science program.

Nobel-Prize winning economist Daniel Kahneman has made it his life’s work to scrutinize the many ways in which human beings fail to behave rationally.  His most recent book, “Thinking, Fast and Slow” summarizes the 40 years of research conducted by him and his colleague Amos Tversky on the topic.

Kahneman categorizes human thinking into two systems: 1) fast and 2) slow.  The fast system (System-1) represents rapid decision making based on assumptions, prior personal experience, and neurologic short cuts that have proven advantageous to survival.  System I “prepares us to perceive the world around us, recognize objects, orient attention, avoid losses, and fear spiders,” using subconscious actions and reactions that are unavoidable.  One cannot unlearn the simple mathematical formula 2+2=4 any more than they can stop themselves from flinching when surprised by the sudden bark from an angry dog.  To work at such instantaneous speed, System-I thinking relies on biases and heuristics.  Kahneman explains the “hindsight bias,” the “optimistic bias,” the “availability heuristic” and the “fast and frugal heuristic,” along with many others.  These multiple biases and heuristics detail the seemingly infinite ways in which humans divert from the utility maximization path laid out by econometrics and utility theory.  See here for a basic overview of decision theory in econometrics

By contrast, the slow system (System-2) is lazy and laborious.  Activating System-2 thinking requires us to pay attention and allows us to follow complex directions, compare choices, reflect on the past and project into the future.  Activities that rely on System-2 thinking require self-control and sustained effort.  In short, engaging System-2 is oftentimes depleting and unpleasant. 

Difficulties arise when we are presented with a situation that intuitively engages System-1 thinking, but actually requires System-2 thinking.  For example, if I told you the price of my son’s new ball and bat was $1.10, and the bat cost $1.00 more than the ball, your brain would likely switch into System-1 thinking by immediately concluding that the ball cost 10¢.  And, you would be wrong.  If the ball were 10¢, the bat would be $1.10 and our total would be $1.20.  By engaging System-2, we find the actual cost of the ball is 5¢.

Kahneman’s brilliant studies also demonstrate that humans are risk-averse when there is hope for a gain, yet risk-seeking when faced with a decision between certain losses.  He goes on to explain that we tend to rely heavily on System-I thinking when faced with uncertainty, trade-offs, or long-term horizons.

Uncertainty, novel situations, trade-offs, hopes, certain and long- term horizons are common features of many health care decisions.

The surprising truth is that medical providers seldom have solid evidence of superior treatment effectiveness.  More frequently, the clinical context is a choice between treatments with outcomes that differ depending on the option chosen without clear evidence to tell the patient or provider which is the superior choice.  Trade-offs exist with probabilities of good and bad outcomes over various time horizons.

For example, a patient suffering from 6 weeks of uncomplicated low back and leg pain from a herniated disc could chose surgical or non-surgical treatment.  Surgical care holds the promise of relieving the leg pain more quickly than non-surgical care but at the added risk of the surgical procedure itself.  Non-surgical care holds the promise of relief over a longer period of time and without the surgical risks.  Existing evidence suggests the long-term outcomes are equivalent.  (See here, here, & here for supporting evidence.) For this reason the patient’s preferences should determine the treatment pathway.  More commonly, however, it is the provider’s opinion that determines the patient’s treatment destiny.

Following this thought process further leads to the uncomfortable realization of a seldom discussed medical error, operating on the wrong patient.  I do not mean operating on patient A instead of patient B.  By “wrong patient,” I mean one who would have, had she been fully informed of the risks and benefits of all treatment options, chosen non-surgical care.  Evidence of the over-use of surgical interventions in this manner is detailed in Dartmouth Professor Jack Wennberg’s latest book, “Tracking Medicine.”

Patient preference-sensitive clinical situations are best addressed by aligning the fully informed patient’s preferences with the care she receives.  Doing so requires System-2 thinking by both the clinician and the patient.  This is not easy and there are legitimate reasons for both the clinician and the patient to be unhappy.

The situation is difficult for clinicians because System-2 thinking is laborious, but not reimbursed.  There are no evaluations or procedures to order.  Further, medical school does not provide the skills required to elicit a patient’s preferences, fears, and risk-tolerance, let alone their numeracy.  This final aspect is especially important because medical decisions frequently involve very small numbers that are difficult to place in context.

Patient preference-sensitive clinical situations are difficult for patients because they may not want to be so involved.  Patients engage the health care system at their most vulnerable times.  They reveal secrets to us their spouses may not know.  They disrobe in front of us and we peer inside of them with our technology or surgical technique.  Despite all this, patients reveal their anxieties and hopes to us, after the briefest of interactions, in the hopes of receiving information, wisdom, and care that can prevent or ease pain and dysfunction.

In return, they should receive our best information and judgment.  Past attempts to involve the patient in sharing the medical decision have, in my opinion, come up short.  They have relied on econometric utility-maximizing theory, weighting each of the probabilities in a decision tree and then mathematically folding branches back to reveal the most rational choices with the highest probability of producing the most utility for the patient.  This process has always struck me as an unusual because, in over 15 years of clinical work, I have never met a truly rational patient.  Nor for that matter, have I encountered a fully rational and unbiased clinician.

The work done by Kahneman and Tversky debunks utility theory and replaces it with “prospect theory.”  Kahneman describe this well in his book.  It is incumbent on us to find ways to design the delivery of care in ways that incorporate the insights gained from prospect theory and employs methods that lead to fully informed patient choice.  I think you’ll agree and look forward to your comments.

 

So you want to think more about rationing...

post by Paul Kelleher

In light of the conversation we've been having about the American College of Physicians' Ethics Manual's statements on resource allocation, I thought I'd point to some short, web-accessible sources that can help advance our thinking about these issues. It should go without saying that this is a non-exhaustive list, and that I intend for these to provide useful food for thought, not knockdown arguments or silver bullets.

1. Peter Ubel's contribution here (begins on page 11).

2. Peter Singer's NYT Magazine article, "Why We Must Ration Health Care," here. 

3. Frances Kamm's Letter in response to Singer's article, here.

(if you hit NYT's paywall when trying to access Singer and Kamm, let me know in the comments and I can help you out)

4. Norman Daniels's contribution here, especially from page 50 onward.

5. Marthe Gold's podcast here (the first is longer [14min] and worth your time). It is based on this 2005 paper in Health Affairs.

6. Jonathan Wolff's remarks here, especially from about 16:45 onward.

These last two focus on the ways in which the UK's National Institute for Health and Clinical Excellence (NICE) sets health care limits and priorities. 

Locating Values in Health Care Policy Decision Making

post by Paul Kelleher

A few days before Christmas, the New York Times's Economix blog published a characteristically excellent post by Uwe Reinhardt on the recently released Wyden-Ryan proposal (pdf). As many of you know, Reinhardt is a master when it comes to explaining complex health policy concepts and principles in clear, accessible prose for non-experts. In this particular post, Reinhardt pellucidly lays out "the major economic functions any health system must perform" (see his first numbered list for these functions). He then remarks:

Most of the debate over health policy in this country has been over two questions.

First, to what extent should healthier or wealthier members of society be asked to subsidize the health care received by their poorer or sicker fellow Americans? Second, influenced by the answer to the first question, who should perform the functions listed above: government, private nonprofit entities or for-profit entities?

The first of these questions is quite obviously a central values question for health policy, and I've written a few posts addressing it here on the blog. As Reinhardt notes, the answer to the second question must be influenced by the answer to the first, but the second question is also clearly more empirical and technocratic in nature than the first.

Reinhardt says that the Wyden-Ryan plan, which is a form of "managed competition," answers the first values question thus:

Let there be a limit to the extent to which healthier or wealthier Americans are forced to subsidize the health care received by sicker or poorer Americans – that is, to the practice of social solidarity.

According to Reinhardt, Wyden-Ryan incorporates this "truncated" conception of social solidarity by limiting the size of the voucher that individuals would receive to purchase health insurance plans within a regulated health insurance marketplace:

The major point of contention in the political debate over managed competition...is over the nature of the “defined contribution,” which defines the extent to which the concept achieves social solidarity in health care. The crucial distinction here is between “vouchers” and “premium support” (as explained in Section B of this Brookings Institution document). [emphasis added]

I think Reinhardt makes a slight mistake here, for I do not think it's true that the answer to this question--i.e. how generous the vouchers should be--"defines the extent to which [a form of managed competition] achieves social solidarity in health care."

To illustrate, imagine that you and I have both been given a very very generous subsidy by the government to purchase insurance on the open market. Let's assume it's $30,000. In most cases, this will be enough to purchase gold-plated health care for a single person. And this may well be true in your case, since you're relatively healthy. But now suppose I have a super expensive pre-existing condition. Depending on how expensive my condition is to treat, there may be no insurance company that would offer me coverage, even if I had $30,000 per year to spend on premiums. If we nevertheless think there is something wrong with that situation, we may choose to place regulations on the marketplace that forbid insurance companies from denying me coverage and that ensure I have affordable access to needed medical care.

Now, these regulation cannot exist in a vaccuum: if insurance companies are required to cover me, then other regulations will have to be put into place to ensure that the market doesn't become disfunctional. Such additional regulations are precisely the regulations the Affordable Care Act imposes for precisely these reasons, as Reinhardt masterfully explains elsewhere.* Yet since these regulations have potentially burdensome implications for others--e.g. they may involve an individual mandate (or its equivalent) and/or make insurance premiums more expensive for others--there is every reason to think that they too help define the extent to which health policy embodies social solidarity. Put differently, if you think it violates some freedom or prerogative of yours to be mandated to have health coverage or to pay a premium higher than is justified by your own actuarial health risks, then you may object to the marketplace regulations embraced by the ACA and Wyden-Ryan. And this may influence--or be influenced by--your understanding of the demands of reasonable social solidarity. 

This isn't a big disagreement with Reinhardt, obviously. And I hope he'd agree with my minor correction to his useful post. Still, I do think it's super important to be able to locate the areas of health policy that must be responsive relevant social values. Reinhardt himself proves why: since health care policy has so many moving parts, it may be possible to adjust several dials to achieve the same degree of social solidarity while lessening the burden placed on others. But we won't get it right if we think that the size of the voucher is the only choice that must be responsive to our understanding of social solidarity. 

*Really, if you are going to give someone you love just one thing to read on the ACA, give them the Reinhardt piece linked to here.

A worse than useless survey

post by Paul Kelleher

Robin Hanson points us to a forthcoming paper in Economics Letters by Olof Johansson-Stenman that purports to "provide survey-based support for the consequentialist assumption associated with welfare economics." Consequentialism is the normative view that an act or policy is right if and only if it maximizes the good (i.e. the net balance of good over bad). As such, it treats an act's rightness or wrongness as solely a function of outcomes. Modern welfare economics is consequentialist in that it aims to be a science of efficiency focused on welfare-maximization. But then again, plenty of welfare economists insist that what they study is but one sphere of moral investigation, and they would not agree that an interest in efficiency entails a single-minded interest in it.

Johansson-Stenman sent his survey to 2,450 Swedes over the age of 18 in the Spring of 2004. Forty-five percent responded (n=985). Here is the main question asked, together with the respondents' answers (click the image to enlarge it):

According to Johansson-Stenman, "the result is quite consistent with the consequentialist ethics underlying conventional economic welfare theory, since almost two-thirds chose this alternative." But this is an illegitimate interpretation of his results. As noted, consquentialism is the view that rightness and wrongness depends only and exclusively on the outcome. The survey, however, only asks people what they think the badness of an outcome "depends primarily on." So for all we know, the 67% who gave what the author characterizes as "the" consequentialist answer may actually agree with those welfare economists who are not consequentialists at all because they believe that other moral considerations (e.g. individual rights and liberties) matter too.   

The survey also makes the fatal mistake of asking its main question in terms of "badness" rather than "rightness" or "wrongness." To see why, consider this passage from a paper (pdf) by Frances Kamm, philosophy's leading non-consequentialist:

I shall distinguish between goodness, fairness and justice. To make one distinction clearer, consider the following case: A doctor must decide whether to stop a big pain in person A or a small pain in person B. She thinks, correctly, that she will do more good if she helps A. But she also remembers that yesterday B suffered a much bigger pain than A will suffer and no one helped B (A suffered nothing in the past). So she thinks it would be unfair to let B suffer again, even though she will do less good if she helps him. If it is overall right to do this, this means she does the morally better thing in helping him and the state of affairs in which B is helped rather than A is morally better than one in which A is helped. But this is not because it produces more good.

Here Kamm claims that sometimes considerations of fairness can conflict with considerations of goodness and badness. But note that someone using Kamm's conceptual scheme might well have given the survey answer that Johansson-Stenman associates with consequentialism. For the survey asks respondents to indicate which considerations they associate with "badness," and the simple (but not simplistic) Kammian answer is that welfare-related considerations are associated with badness, while fairness or equity considerations must be called in alongside welfare considerations to determine which action is "overall right" or "morally better." So it is quite possible that many survey respondents took themselves to be answering a question about only one part of morality, viz. badness, while also holding beliefs about other parts not asked about.

I conclude, therefore, that this survey is useless in helping to answer the question of whether most people are consequentialists. Indeed, since its publication has already led one influential writer (Hanson) to claim that "Most are consequentialist," I fear that it's actually worse than useless.