Ground Control to Major Reform

This article is part of a series on Health Policy. See also:

  1. Ground Control to Major Reform
  2. Hospital Salaries Could Cut Care Costs
  3. The Appropriate Practice Scope of Chiropractic May Be a Political Question, Not a Scientific One

The United Network for Organ Sharing (UNOS) runs the nationwide waiting list for donor kidneys. 100,000 Americans currently sit on this list; unfortunately, 50% will die before a kidney arrives, as wait times can exceed 10 years.  Further, since the number of recipients is growing faster than the number of donors, the wait time, and consequently the mortality rate, can only be expected to increase. (1)

But there’s hope! Recent updates to the UNOS donor kidney allocation policy might drastically reduce wait times.

Larry Swilling took to the streets for fear that his wife was losing the race with the waiting list. He found a donor just a month ago.

UNOS policy revisions

Since its establishment 25 years ago, the UNOS waiting list has worked on a first-come, first-served basis.  Of course, factors like blood type and organ size are considered. (4) Besides these, though, priority is granted solely based on the order in which recipients sign up for the list. While this policy underscores fairness, it comes with great drawbacks.  Age of the donor and recipient are not considered, which means that a young, healthy kidney from a 30-year-old donor might easily end up in a recipient with only a few years left to live. Thus many life-years are wasted. Worse, the reverse scenario might occur.  Chances are that a 30-year-old recipient of an old, worn-out kidney will end up back on the list again, further increasing wait times. The first-come, first-served system is fair, but it sacrifices efficiency. (2)

Just a few months ago, the UNOS approved sweeping changes to its allocation policy. One of these changes was that the 20% of kidneys expected to last the longest would be reserved for those 20% of recipients expected to live the longest following the transplant. These changes will be implemented sometime in 2014. (3)

I’m not satisfied, though.  The 30-year-old might not receive a 70-year-old kidney, but the 40-year-old might. Experts agree. “You haven’t given an ethical justification for why you’re going to create this top 20 percent,” says Lainie Friedman Ross, a bioethicist at the University of Chicago. (2) The problem is lessened, but only slightly, and now a different age group bears the worst of the burden.  Why stop at the top 20%? Why not reserve the 20%-40% healthiest organs for the 20-40% youngest recipients, and so on?

Dynamic allocation policy programming

Harvard Business School professor Nikolaos Trichakis studies methods for efficient air traffic control.  In a purely-fair system, the first plane to land on the runway is the first to take off. However, once again, fairness precludes efficiency. One plane, slotted near the front of the line, sends its passengers toward a four-hour layover.  Meanwhile, another plane sits at the back of the line as its passengers miss a tight connection. A simple swap would have allowed a successful trip for all parties.


Upon hearing of the proposed UNOS update, Trichakis wasn’t satisfied either. So he developed a matching algorithm that includes “life-years” as one of its components, as well as criteria for physiological compatibility and fairness. Using decades of past medical data, he input the same restrictions we use today, and then allowed the program to compute optimal parameters for maximization of life-years, using algorithms similar to those in Ben’s stock performance optimizer. Trichakis’s system provided an 8% increase in life-years, even with standard fairness restrictions! This would undoubtedly exceed the effect of UNOS’s 20% system. And by relaxing fairness constraints, projected life-years increased by as much as 30%. (5) Trichakis is currently working with doctors to learn the specifics of organ allocation and distribution, so as to further improve and hopefully implement his procedure. (6)

Trichakis’s 8% is a no-brainer, since fairness isn’t sacrificed. What about his 30%? Here, more interesting questions arise.


Let’s talk a little more about the fairness constraint. For one, fairness prioritizes those who have been waiting longer.  Also, the constraint prioritizes certain physiological traits: each recipient possesses certain traits that might facilitate or obstruct donor matching. For example, the O blood type is the universal donor, so type-O patients can only receive transplants from like patients. Meanwhile, AB is the universal recipient.  So, a type-AB patient will have a much easier time finding matches. Patients’ natural immune response varies as well. A patient with a sensitive immune system will be more likely to reject a kidney, and thus can only accept kidneys containing very low levels of potentially-dangerous antigens. The fairness constraint gives priority to type-O and immunosensitive patients so that no patient is denied access to donor organs because of inherent physiology. (5)

We could increase life years from 8% to a huge 30% if we simply told type-O and immunosensitive patients “tough.” Some patients would die, but many would live much longer! In the end, isn’t “life-years” what matters? However, such a decision would be unconscionable. Why? Well, it seems wrong to penalize A for conditions outside his control, even if that penalty might extend the life of B. Hold on a second, though: isn’t B now penalized instead? A’s remaining years rises from 0 to 5, while B’s drops from 30 to 20. And it’s outside of B’s control that A exists in the first place!

At the same time, it seems more heinous to directly end A’s life than to indirectly shorten B’s. Deontological ethics seem to be at play here. One could even argue that A needs those 5 years more than B needs his 10. Even 1 year—even just a day—would allow for final goodbyes and reflections.  “Isn’t one moment of bliss sufficient for the whole of a man’s life?

The 8 vs 30 debate can rage on, but for now, let’s at least choose 8! UNOS’s policy amendments are a great step in the right direction, but as long as more can be done, it should. We may refuse to yield to slackened fairness constraints, but let’s at least recognize the power of bioinformatics.


  1. Bioethics and Transplantation: Who Has a Kidney?
  2. Who’s Next In Line For A Kidney Transplant? The Answer Is Changing
  3. Public comment sought on proposed revisions to deceased donor kidney allocation policy
  4. Arthur Caplan’s piece on Organ Transplantation from From Birth to Death and Bench to Clinic: The Hastings Center Bioethics Briefing Book for Journalists, Policymakers, and Campaigns
  5. Fairness, Efficiency and Flexibility in Organ Allocation for Kidney Transplantation by Nikolaus Trichakis
  6. What Air Traffic Can Teach Us About Kidney Transplants featuring mention of Trichakis



4 comments on “Ground Control to Major Reform

  1. Ben says:

    “One could even argue that A needs those 5 years more than B needs his 10.”

    This is consequentialist reasoning. “Let’s choose a good,” the reasoning suggests — in this case, enjoyment of life-years — “and maximize it.” Perhaps more surprisingly, even much more abstract reasoning — for example, “Let’s choose a good, say fairness or justice, and maximize it” — would be consequentialist reasoning all the same.

    We reach into deontological territory when we choose a duty — for example fairness, justice, or, utterly paradoxically, the maximization of utility — and adhere to this duty at all times, even when we know that our adherence will trigger a net decrease in adherence elsewhere.

    Interestingly, it could be argued that as the UNOS is the sole body handling kidney allocation, the two branches of ethical theory are likely to coincide! Let me explain. There’s no “elsewhere” to speak of; there’s no other allocating body onto which decreased adherence to some duty might be impinged. We’re now choosing between (a) choose a good and maximize it and (b) choose a duty and adhere to it. But the crucial distinguishing factor — the question of whether, in the calculation of morality, to consider the ramifications of a moral deed on the likelihood of other parties to commit a similar moral deed — is absent.

    This line of reasoning could be challenged by arguing that the mere existence of other people in general — as opposed to the existence, in particular, of another allocating body — is enough to produce a distinction between the two ethical theories. After all, deontological adherence to a duty could trigger, in other people (though perhaps not upon other kidney allocation organizations), a reduced adherence. And the paths have now split.

    Responses to this could suggest that the values to which a kidney allocation body might adhere are reserved, in particular, to decisions made by kidney allocation bodies, and that the presence of individual actors in general is not sufficient to trigger a divergence between the ethical theories. For example, the deontological moral code which advocated “Follow fairness constraints at all times when making kidney allocation decisions” is likely, in the absence of other allocating bodies, to coincide with the consequentialist moral code “Maximize the adherence to fairness constraints whenever making kidney allocation decisions.”

    • Josh says:

      So it appears that our argument is that, on the greatest scale, UNOS’s policy of maximizing life-years while simultaneously following fairness constraints likely follows both consequentialist and deontological guidelines, given the absence of other organ allocating bodies against which we might assess the effect of UNOS’s policies.

      The more interesting matter is the ethical doctrine within the UNOS, across different patients. Here, it might be tempting to say that the UNOS follows deontological ethics in regards to life extension. Maximize A’s life-years, even if we know that this will cause a greater reduction in B’s life-years. I don’t think this stands, though, because we’re not just maximizing life-years arbitrarily; we’re maximizing life-years in accordance with fairness constraints, which, as you mentioned, corresponds more closely to a maximization of enjoyment of life-years rather than a maximization of the years themselves. And this policy of wholesale maximization is consequentialist.

      This argument shows cracks too, though, once we realize that the premise by which we decided that compliance with fairness constraints promotes maximum enjoyment of life-years might not be valid in the first place. “A needs those 5 years more than B needs his 10,” we decided, “and therefore allowing A the transplant maximizes enjoyment of life-years. However, won’t the time come someday where B, too, is reaching the end of his days? Won’t B be aching for just 5 more years, much like A was 20 years ago? Alas, those extra years will never come, because B’s kidney went to A long ago. The fact is that both A and B stand to suffer equally from the prospect of being denied a transplant; this suffering just occurs at different times. Thus following this fairness constraint doesn’t maximize enjoyment of life-years after all.

      What does it maximize? If the answer is indeed fairness, we need a definition of fairness that’s independent both of maximization of life-years and of enjoyment of those years. Here’s what I think is really going on.

      An out-of-control freight train is speeding towards a tight turn, and will almost certainly derail. Opposite the thin retaining wall is a crowded restaurant. In the event of a derailment, as many as a dozen diners could die. As the railyard operator, you recognize the impending catastrophe. Your only chance at averting it is to throw a switch, which would divert the train to a normally-closed section of rail. Unfortunately, one of the workmen is sitting on that section, taking lunch. Do you throw the switch anyway, killing one to save twelve?

      My guess is that many of you would say yes. Now, let’s adjust the question. Your switch is broken too, and now, your only chance at averting the crisis is to push an unsuspecting workman into the path of the oncoming train. This would cause it to jolt and derail before reaching the crowded restaurant. Are you still willing to kill one to save twelve?

      To sit here and discuss the benefits of disregarding the fairness constraint resembles the former case. But to actually deny a patient a transplant due to his sensitized immune system much more resembles the latter. The reality is that allowing A the transplant probably doesn’t maximize some calculable ideal; it just feels fair. And to deny A the transplant would provoke a visceral reaction of distress, much like the feeling of throwing a man onto the tracks. Ultimately, it’s that feeling we’re trying to avoid. We side-step the visceral revulsion of directly turning A away, now, even if we might feel that same revulsion 20 years from now when we realize we’ve indirectly damaged B’s prospects. And that decision, my friends, is deontological.

  2. Richard says:

    Interesting piece, which raises some important bioethical issues. I agree about the usefulness of Trichakis’s work, as you’ve explained it, at least. At the end, Josh says, “At the same time, it seems more heinous to directly end A’s life than to indirectly shorten B’s.” This is true, especially if in lengthening A’s life we are saving A’s life, whereas for B we would merely be prolonging his life beyond a given minimum. Who knows, maybe being younger, B will be more resilient and eventually will end up in a position like A when he has no years remaining and gets a kidney in time, earning him five of his lost ten years back. That looks like an even distribution of life years. But it probably wouldn’t be as idealized as that.

    Regarding ethical theories in the commentary, Ben comments and eventually says: “the crucial distinguishing factor — the question of whether, in the calculation of morality, to consider the ramifications of a moral deed on the likelihood of other parties to commit a similar moral deed — is absent.” This might be true, but it need not mean a convergence of consequentialism and deontological ethics – a duty is something which must be held across different times and scenarios, we can imagine one and the same company at different times as constituting a ‘plurality’ of duty-following scenarios. The consequentialist may shirk duties across scenarios in response to the environment, but the deontologist will not. A net decrease in adherence to UNOS-specific duties could be achieved despite there only being one UNOS-like group if UNOS at one time moralizes itself into a position where its own resources become less and less (as they’ve been giving away kidneys by fairness and allowing the demand to increase as a side effect) and they feel forced to renege at a later time. So I’m not sure if the point that “the deontological moral code which advocated “Follow fairness constraints at all times when making kidney allocation decisions” is likely, in the absence of other allocating bodies, to coincide with the consequentialist moral code “Maximize the adherence to fairness constraints whenever making kidney allocation decisions.” really follows. The deontological maxim is to act in a way you can will to be universal (for all individuals across all times), which means that UNOS alone may, over time, become a threat to their original duty to be fair if circumstance prevails upon their desire to ultimately save as many lives as possible. If such a conflict should arise, the consequentialist principle “Maximize the adherence to fairness constraints whenever making kidney allocation decisions” would still hold but only because ‘maximise’ means ‘make as great as possible’ where the expression ‘as possible’ would await interpretation based on the untoward consequences which could arise in a particular given scenario. However, the deontologist principle would be scrapped.

    I agree that there are arguments against distinguishing consequentialist and deontologist principles over certain domains, but I think more needs to be said before we say that it happens in the case Ben mentions.

    Having read Josh’s subsequent comment, I see he touches on the points I make in certain respects.

    • Josh says:

      Richard, I agree with your points. When asking whether universalizing a certain action leads to increased or decreased achievement of some outcome overall, we should ask this question with respect to time as well, and not just space, so to speak. In other words, the fact that there’s only one governing body (UNOS) shouldn’t preclude us from asking questions about whether or not its procedures conform to deontological ethics. Some action by the UNOS might promote fairness, or enjoyment of life years, today, but lead to a net negative value of that outcome by the time the future arrives.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s