Balancing Equity and Efficiency in the Market for Kidneys

Dr. Lainie Ross is the Associate Director of the MacLean Center for Medical Ethics at the University of Chicago. She is also the Carolyn and Matthew Bucksbaum Professor of Clinical Ethics, the Co-Director of the Institute for Translational Medicine, and a practicing clinician and professor in the Departments of Pediatrics, Medicine, and Surgery. She serves on the Department of Health and Human Services’ Secretary Advisory Committee on Human Research Protection (SACHRP) and as the chair of the American Academy of Pediatrics Section on Bioethics Executive Committee. She previously served on the ethics committee for the United Network for Organ Sharing (UNOS). Dr. Ross has both her MD and PhD in Philosophy. Some of her work in the field of medical ethics aims to ensure the equity of the kidney allocation system.

In your U.S. News & World Report op-ed, “New Kidney Allocation Proposal is Ethically Unacceptable,” you assert that the Kidney Committee’s 20/80 proposal is unethical. Could you tell our readers why you’ve taken this stance?

Since that article was published, the Kidney Transplantation Committee’s proposal has been modified but is, nonetheless, still unethical. The original proposal aimed to classify the top 20 percent of deceased donor kidneys as the best kidneys, and match these kidneys to the top 20 percent of candidates. They were going to use an age-matching program for the other 80 percent of kidneys and candidates.

Their newer proposal still includes the component that the top 20 percent of kidneys will go to the top 20 percent of candidates. However, in this newer proposal, the other 80 percent of patients will receive kidneys based on a revised “point system.”

The first ethical problem regards their measurements. How do you distinguish the top 20 percent of candidates and the top 20 percent of kidneys? A 19.5 percent kidney looks a lot like a 20.5 percent kidney. In fact, a 19.5 percent kidney looks a lot like a 30 percent kidney.

A kidney’s “ranking” is determined by KDPI (kidney donor profile index). The accuracy of the KDPI in distinguishing the top 20 percent of kidneys from the next 20 percent of kidneys has a c-statistic of approximately .68, which means that the KDPI classifies kidneys correctly only 68 percent of the time. Additionally, candidates are ranked by estimated post-transplant survival (EPTS), a measure that also has a weak c-statistic. It also only includes one health condition (diabetes) when long-term survival is more closely correlated with cardiac disease. It only includes diabetes because it is easier to measure. So, one reason this method of allocation is unethical is that the measurements are too arbitrary.

On the upside, their new proposal modifies the current point system, which is based on how long a person has been on the kidney waiting list. The current system heightens health disparities, as minorities are oftentimes not immediately put on the waiting list after beginning dialysis, whereas wealthier, more health-sophisticated individuals manage to get on the waiting list before they even need dialysis. So, for these health-sophisticated individuals, they’re accruing points even if they’re not on dialysis.

The Kidney Committee’s new proposal allocates points based on both dialysis time and the amount of time a patient has had a Glomerular filtration rate (GFR) score less than 30. Individuals who are not on the waiting list, but who have been receiving dialysis, will retrospectively receive points, which will undoubtedly reduce disparities. However, I don’t believe that factoring in time with a GFR score below 30 is ethical. Patients do not need dialysis until their GFR is below 15, but the Kidney Committee decided to include a “GFR less than 30” measure because they don’t want to stop preemptive transplants—the transplant of somebody before they go on dialysis. This is because if you’ve never had dialysis, or you’ve been on dialysis less that a year, your transplant results are better than if you’ve been on dialysis for many years. However, this heightens disparities, because whites and people of higher socioeconomic status are more likely to be preemptively transplanted. I wish they utilized pure dialysis time as their measure, because pure dialysis time would be the fairest.

In your article, you mention that a change in the allocation system must benefit the worst off, such as individuals who have had type-1 diabetes since their childhood. How does the 20/80 system affect individuals with co-morbidities like type-1 diabetes? 

In the Kidney Committee’s proposal, diabetics will never be in the top 20 percent. When I sat on the UNOS committee, one model to revise kidney allocation used a measurement called LYFT: Life years following transplant. This model would have been great for diabetics. Diabetics do really poorly on dialysis—they actually have the highest mortality rate per year of patients on dialysis. The LYFT proposal would allocate kidneys by subtracting the number of years a patient could live on dialysis from the number of years a patient would be expected to live with a transplant. A diabetic, who does really poorly on dialysis, will have a higher LYFT. Black patients, on the other hand, tend to do really well on dialysis and, as a result, have lower LYFT scores. Thus, their chances of receiving a kidney are harmed by the LYFT system. I was against LYFT because I thought it unfairly discriminated against minorities. Nevertheless, it would have helped the type-1 diabetics.

Now, their new proposal will exclude type-1 diabetics from the top 20 percent—so I have asked them point blank: What is your ethical theory about type-1 diabetics? Should they be treated like everyone else? Should they be given special privileges? In one proposed system they would have received priority and in another, they are disadvantaged—how can we justify this? UNOS is required under NOTA—the National Organ Transplant Act—to consider both equity and efficiency. However, the new proposal tends to focus on efficiency exclusively. In fact, in their report, they fail to even mention equity.

How would you change the kidney allocation process?

I think you have to start with equity. Our demand for kidneys is so much greater than our supply—about 700,000 people are on dialysis, yet only 100,000 are on the deceased donor waiting list. Some of the 600,000 on dialysis could benefit from a transplant but have not yet gone through the work-up and may not even know the advantages of a transplant. We shouldn’t ignore the 600,000. Although some may not be eligible for transplantation, who gets on the waiting list is not random: men, whites, and those of higher socioeconomic status are more likely to be on the waiting list than the dialysis numbers would anticipate. But for the 100,000 on the waiting list, they all want a fair chance at getting the optimal renal replacement therapy, which is transplant. However, 70-year-olds don’t need a kidney that’s going to last for 30 years.

So in my proposal, the Equal Opportunity supplemented by Fair Innings (EOFI) model, we offer a two-step multi-principled approach to allocation. In the first step, we ensure that regardless of age, all waiting list candidates have an equal chance of getting a kidney, which is based on the equity principle of equal opportunity. In the second step, we use age as a proxy for quality of both candidates and kidneys on the basis of the equity principle fair innings, which says that younger people need kidneys that will last longer so that they can achieve a full lifespan (some would call this equity principle “prudential life span equity”). We would allocate younger kidneys to younger people and older kidneys to older people, and we would ignore all co-morbidities like diabetes.

This means that, although younger people don’t necessarily have a longer life expectancy and some younger kidneys are not better than older kidneys, younger individuals will still get the younger kidneys. We don’t deny the inaccuracies of our measurements in determining longevity, but rather, we acknowledge that there is variation in kidneys and candidates, and we accept that there is some randomness. This means that part of the allocation system is a “lottery”—where most younger kidneys are better than older kidneys. And lotteries are another equity principle that we endorse. In our minimalistic modeling of EOFI, we show it is more equitable and more efficient than the current system, although not as efficient as the 20/80 model.

Switching gears, what do you feel is the most pressing ethical issue in medicine today?  

Disparities and access issues. The Affordable Care Act (ACA) will address this, at least by improving the number of people who have health insurance. So, from a medical ethics perspective, that’s fabulous.

There’s another pressing and interesting issue: learning health systems. We’re moving in a direction where every hospital will be expected to have electronic medical records. It will be easy to conduct quality improvement research with all of this data without patients even knowing that their data are being used. All of us are going to be patients and research subjects at the same time, all the time. Why is this ethical? Because while it is your data in the clinical context, once it is de-identified (stripped of your identity), it is data about a young female with problem X and then we aggregate your data with other young females with problem X to see which treatments worked and what are best practices.

When patients hear this, some become angry that they are research subjects without their permission. Hospitals, doctors, schools, and the government all need to educate the public that it is to everyone’s benefit to allow and encourage the use of de-identified health data for advancing medical knowledge. When our health data are de-identified, no one knows if it is you or your classmate who has problem X, but only a young female with problem X. And the data are aggregated so that the information is presented about groups of people.

In contrast, patients are right that they need to be informed when research is being done to them—whether it is the use of a new medication or new device. When we do such research, we clearly need to involve the patients who are then active participants in the research and we need to get their informed consent, to educate them about their right to refuse to participate or even to withdraw.

I serve on the Department of Health and Human Service’s Secretary Advisory Committee on Human Research Protection, so I take the issues of human subjects protections very seriously. But the public must be educated that health care is evolving into a learning system—we take care of patients and use the data to determine if what we did was best practice. And we do this by aggregating de-identified data pulled from medical records.

In your opinion, what is the impact of the ACA on medical ethics?

I am not an expert on the ACA, but I appreciate that the ACA incentivizes the use of medical records and is leading us towards the learning health system. As I mentioned, the ACA will also improve access for at least 10 million people, which is a step in the right direction. However, it is only a step, as many millions will still be without access to affordable health care.

Access is the first step; and hopefully we will all use our access to obtain good preventive medicine and primary care. If people have good primary care, we could prevent the person with diabetes from developing kidney failure, we could prevent the person with high blood pressure from developing kidney failure, and so on. I think access is the most important step from a medical ethics perspective, and that’s why I’m excited by the ACA.

Feature Photo: cc/(NomadicEntrepreneur)