Potential Inefficiency of a Proposed Efficiency Model for Kidney Allocation

Unintended Consequences of the New System 

The stated goal of developing a new algorithm is to “achieve the greatest total number of life years from transplant among all the recipients, given the available organ pool. Allocation of each kidney to the candidate with the greatest LYFT for that organ is a way to maximize the total number of life years gained.”1 However, is this the case with an algorithm that accounts for only deceased donor organs? We suggest that the current UNOS algorithms that maximize LYFTs are inadequate because they simulate gains in LYFTs for only a nonrandom sample of kidneys transplanted. They may be less efficient or even inefficient in the overall distribution scheme because they assume no effect on living kidney donations, the 40% of donor kidneys that offer the best LYFTs. One reason for excluding living kidney donors from the KARS proposal is that historically, UNOS did not collect data about living donor transplantations, which makes it difficult for modelers to predict what will happen to the 40% of kidneys that come from living donors (Fig 1A).

View Large Image Download to PowerPoint

Figure 1. (A) Total transplants, (B) transplants to minors 17 years or younger, (C) transplants to adults aged 18 to 34 years, (D) transplants to adults aged 35 to 49 years, (E) transplants to adults aged 50 to 64 years, and (F) transplants to adults 65 years and older, all from living and deceased kidney donors. All data available from United Network for Organ Sharing (http://www.optn.org/latestData/rptData.asp).

However, recent data from the Scientific Registry for Transplant Recipients show a correlation between allocating optimal deceased donor kidneys with short waiting times to a specific recipient population and changes in the use of living donor organs in that population. In November 2003, the OPTN/UNOS Board of Directors approved a policy that would give priority to minors (<18 years) to receive kidneys from deceased donors younger than 35 years. The new policy was implemented in September 2005. Figure 1 shows the impact of these policies using data from UNOS itself.6 The percentage of total kidney transplantations in children using living donor organs had been slowly decreasing for several years after earlier attempts to prioritize deceased donor allocation to children by giving them extra points within the distribution algorithm. This trend is seen for 2004 to 2005, in which living donor kidneys to minors decreased as a percentage of the total (50% to 47%) despite a numeric increase in both deceased donor and living donor transplants. However, a much more dramatic decrease in 2006 occurs after implementation of the new policy’s more stringent prioritization of minors so that they can receive optimal deceased donor kidneys without a protracted wait. A 26% decrease in living donor kidney transplantations in children was observed between 2005 and 2006, although the total number of transplants in this recipient group remained essentially unchanged (Fig 1B).6 During the same time span, numbers of living donor transplants for all other age groups were static (Fig 1C to F). Although we have only the temporal correlation and lack definitive proof that the decrease in living donors to minors is a response to the specific policy change, there are no other ready explanations for this observation. If a new allocation system is instituted that gives young adults (18 to 35 or 18 to 49 years) priority for standard-criteria donor kidneys, it could lead to a similar sharp decrease in living donation to these adults, similar to the observed change in living donation to pediatric recipients.7 If these young adults were no longer to seek living donors, it could lead to a decrease in overall living donor transplantations. In addition, if standard-criteria donor kidneys were shifted from older to younger adults, there could be increased pressure for living donation to older recipients, for whom the remaining alternative would be to wait for a less desirable expanded-criteria donor kidney. Both these changes could lead to a decrease in LYFTs.

Is Maximization of LYFTs the Proper Goal of an Allocation Policy? 

Whether these unintended adverse consequences would occur if the new kidney allocation protocol were implemented is an empirical question. However, the fundamental moral question of whether maximization of LYFTs ought to be the principal characteristic of an allocation policy also was raised.5 It is not surprising that algorithms that give priority to younger kidney transplant recipients result in the greatest gain in LYFTs because younger individuals in general look forward to longer anticipated net lifetime survival than older ones. However, patients with end-stage renal disease do not support schemes that discriminate on the basis of recipient age.8 The general public also expressed support for allocations that focus on equity (fair distribution) over efficiency (utility) for kidneys.9 The decision to maximize utility as defined by life-years gained despite its inequitable impact on older individuals contrasts with current deliberate exclusion of other utility criteria (race/ethnicity and geographic variation) that could significantly impact on LYFTs.10, 11, 12, 13 The proponents of the currently modeled algorithms need to provide a moral justification for why they focus on only some utility criteria and not others. Otherwise, their modifications appear arbitrary and it is unclear whether their balance of equity and utility is morally justifiable.

Conclusion 

As UNOS considers changes to the deceased donor allocation process, it is imperative to remember that organ allocation is a balance between criteria of equity and efficiency from 2 discrete sources: the deceased kidney organ pool and living organ donors. An allocation algorithm that modifies the distribution of one source of organs may unwittingly influence the other. A new distribution algorithm that gives primacy to efficiency based on life-years gained for deceased donor kidney transplants may be self-defeating because it may have the unintended consequence of reducing benefits from living kidney donation within the overall kidney allocation system.