A proposal out of Harvard and MIT to rethink how kidney transplants are allocated could result in a fairer system giving patients longer lives.
The new empirical model, which is intensely data driven, would provide a flexible framework to policymakers responsible for deciding which potential recipients get organs as they become available—decisions that must be based on various priority and fairness criteria.
“The new system will explicitly give points based on the likelihood of survivability.”
The method—the work of Nikolaos Trichakis of Harvard Business School and Dimitris Bertsimas and Vivek F. Farias, both of MIT's Sloan School—can help policy designers create the most equitable point system based on their chosen constraints and criteria. They detail the proposed model in a new paper, Fairness, Efficiency and Flexibility in Organ Allocation for Kidney Transplantation.
In early simulations, the model suggests that life-year expectancies for the program can be increased by up to 8 percent, depending on variables plugged into the process.
As with the "Moneyball" metrics movement in baseball, the goal of the model is to use the best available data to make the best possible decision in each individual case. But here, the ultimate objective is adding years to patients' lives, not just points to the team's winning percentage.
"Because of the organ shortage, you want a system that is transparent and perceived as fair by the candidates," says Trichakis, an assistant professor in the Technology and Operations Management Unit. "If you're one of them and you're not getting a kidney, you should be able to understand why and what the criteria are."
A Long Line
About 500,000 people in the United States suffer with terminal end-stage renal disease. The only treatments are maintenance dialysis and kidney transplantation. The former requires visits to a treatment center for at least 12 hours a week, while a transplant—from either a living family member or a matching deceased donor—can have the recipient soon resuming regular life activities. In mid-November, the active waiting list for a kidney stood at 72,845. Once an organ is available, there can be thousands of compatible recipients queuing up.
In the time since the US Congress passed the National Organ Transplant Act in 1984, organ allocation has been handled by the Organ Procurement and Transplantation Network (OPTN). This national registry and waiting list is managed by the private nonprofit United Network for Organ Sharing (UNOS), which has the unenviable task of making priority and allocation decisions for each new organ that becomes available.
Currently, this is done under a point system that takes into account a number of factors including the potential recipient's proximity to the available organ, blood type, life expectancy after a transplant, and various fairness criteria such as time waiting on the list.
Those decisions can be complicated by cold realities. Organs typically need to be transplanted within 36 to 48 hours, otherwise they begin to deteriorate, so recipients who live close to the source of the donated organ often are logistically preferable. Another difficulty: waiting lists can be much longer in some areas of the nation than in others.
But one particular concern has policymakers rethinking the current kidney-allocation process, a problem ironically created by improved medical treatments for renal disease. "Because patients are living longer, more and more of them are accumulating points just for being on the list a long time, which dwarf the points accumulated for a matching profile," says Trichakis. The end result: the allocation process "no longer meets the needs of the patients."
So, despite the best efforts of those involved, the current transplant point system is in need of fine-tuning. OPTN is developing a new policy and has put out a formal request for information to help guide that decision-making process, to which Trichakis, Bertsimas, and Farias are contributing.
Their key idea is to propose a new way of thinking about the problem. So far, OPTN has been conceiving the new process as an allocation system to provide good matches so patients survive longer, but with fairness. By contrast, rather than designing a policy and then looking at what the outcomes are, Trichakis and his coauthors allow policymakers to start with desired outcomes and then work backward using historical and other data to create a policy.
The group's proposed method determines how many life years can be added by looking at the expected survivability for every case. "The new system will explicitly give points based on the likelihood of survivability," says Trichakis. Fairness is measured by looking at the distribution of recipients across different groups by age, race, diagnosis, and blood type.
"This gives policymakers the ability to focus on what's important and relevant," he adds. "Right now, they focus more on what the policy should be and then the outcomes."
Putting It To The Test
In order to test their methodology, the research team performed a series of case studies on policies with different criteria for matching. In the first instance, using the criteria and constraints of the current dominant policy, their method produced a program with life-year gains of 3 percent over the existing one. The results were significantly better—life-year gains of 8 percent—in a second study for policies that could be based on all score components considered by policymakers, while continuing to meet the fairness properties of the dominant proposal.
The model also allows policy designers to consider fairness factors not used before in the allocation process. "For instance, what is the impact of reducing the percentage of transplants to patients on dialysis for greater than 15 years by 1 percent?" their paper puts forward. "In the case of some constraints, relaxations of fairness constraints can result in life-year gains on the order of 30 percent. As such, we believe this is a valuable tool in the policy-design process."
The authors are preparing a new version that is specifically aimed at the medical community. The ultimate goal is to bring this work to the attention of the United Network for Organ Sharing and the Organ Procurement and Transplantation Network, allowing it to be considered alongside other proposals.
"The first step is for UNOS and OPTN policymakers to make a proposal, which would go to Health and Human Services for approval," Trichakis says. "The Office of Civil Rights is involved too. Everyone is still in the loop of thinking about how the system should look: how many points for dialysis time, waiting time, things like that.
"I wouldn't say there's going to be a proposal in the next month or so," he continues. "There are different proposals on the table. But we will start talking with policymakers about what we can offer."
Interestingly, Trichakis is not the only HBS professor playing a significant role in development of organ donor programs. Alvin E. Roth, cofounder of the New England Program for Kidney Exchange, is a pioneer in creating a system that allows living donors to exchange with others. So far, the program has been responsible for over 70 successful exchanges.
Kidney exchange programs are organized on the regional level, but the OPTN recently started a pilot program to test the concept on a national scale, where Roth's work is likely to prove equally influential. "Al's work has already influenced local policies and is perhaps about to go national," Trichakis says.