Journal Logo


Potential Impact of Risk and Loss Aversion on the Process of Accepting Kidneys for Transplantation

Heilman, Raymond L. MD1; Green, Ellen P. PhD2; Reddy, Kunam S. MBBS3; Moss, Adyr MD3; Kaplan, Bruce MD1

Author Information
doi: 10.1097/TP.0000000000001715
  • Free
  • Social Media Collection


Kidney transplantation significantly improves both life expectancy and quality of life for patients with end-stage renal disease. According to the 2014 Scientific Registry of Transplant Recipients (SRTR) annual report, there are still approximately 100 000 patients on United Network of Organ Sharing (UNOS) kidney waiting list; however, there are only 17 000 kidney transplants done each year.1 The demand for kidney transplant continues to exceed the supply. Although the number of patients on the waitlist is increasing the number of living donors has decreased.1 As a result of these trends, the median waiting time for deceased donor kidney transplant has increased to more than 7 years and death on the waitlist now exceeds 8%.1

There is significant variability between UNOS regions and individual transplant centers in the acceptance of kidneys from deceased donors.1 The overall kidney discard rate has actually increased from 18% to 19.8% during the initial 18 months after the adoption of the kidney allocation system in November of 2014 ( For certain groups of suboptimal kidneys, the discard rates are much higher. For example, the discard rate for a high Kidney Donor Profile Index (KDPI) (ie, KDPI > 85%) kidneys is 56%.1 Using the UNOS database, Garonzik-Wang et al2 showed that at the center level there is significant variability in the aggressiveness of transplanting suboptimal kidneys. This study found that center aggressiveness correlated with waitlist size, organ shortage, and waiting time. One possible explanation of these findings is that loss aversion may influence the aggressiveness of a center to transplant high KDPI organs and that this loss aversion may be different among different centers. This indicates that as a field a greater understanding of how loss aversion impacts organ acceptance decisions may be of great benefit. The aim of this review is to explore how loss aversion and risk aversion impact the acceptance rate of kidneys from deceased donors. We are not trying to imply that these decisions are wrong per se or imply any moral judgment. Rather, we are merely pointing out that “hard wired” psychological principles may influence the process of accepting kidneys from deceased donors.

Clearly, there is a need to increase the utilization of deceased donor kidneys. Recently, the Organ Procurement and Transplant Network reported the results of a Deceased Donor Potential Study.3 Future donor estimates were derived by using filtering methodologies applied to the National Center for Health Statistics and the Agency for Healthcare Research and Quality databases to estimate the potential donor pool in the United States. The Organ Procurement and Transplant Network report estimates that the potential donor pool could be as high as 38 000 donors per year. This suggests an underutilization of potential donors across all demographic groups, the 2 largest subgroups being aged 50 to 70 year and donations after circulatory death. Hence, these data would suggest that there are opportunities to improve the utilization of kidneys from deceased donors, particularly in the more suboptimal donors. This begs the question, why are we not taking advantage of the full potential of the deceased donor pool?

The process of accepting diseased donor organs is complex. Multiple issues including the quality of the organ, the appropriateness of the recipient, the anticipated cold ischemia time, and the center resources and experience all influence the decision. One potential influence is the perceived risk of poor outcome of the transplant which is driven by both recipient and donor characteristics, many of which are quantifiable. Another is the risk of regulatory sanction. To control for the risk of transplanting kidneys from suboptimal donors the regulatory bodies, particularly SRTR, adjust for multiple donor and recipient characteristics that that are known to influence the transplant outcome. However, there is a perception in transplant community that the SRTR risk adjustment may not adequately adjust for all significant variables. For example, the SRTR does not adjust for preimplantation biopsy findings. The complexity of the process is magnified by the time constraints placed on the surgeon to make a decision to accept an organ. As a result, the process of accepting an organ has an element of risk and is highly influenced by the experience of the surgeon and their willingness to accept risk.

Behavioral economic theory suggests that people are more influenced by the risk of low probability but large loss than high probability but lower expected potential gain. Kahneman and Tversky4 demonstrated that due to a heavier weighting given to large losses than to aggregate total gain, humans will often make choices with lower expected value rather than take the risk of loss when the outcome can lead to a large loss, such as flagging by Center for Medicare Services.

Prospect Theory and Framing

As pointed out in previous discussion in this journal, classic economic theory predicts that an agent will make choices to maximize their potential value subject to budget constraints.5,6 That is, individuals choose the option that provides them with the greatest expected value (eg, always accept kidneys when expected benefits are greater than the expected costs). There are a number of factors that influence the decision process. The expected benefits of accepting the organ includes the improved health for the recipient as well as the expected graft survival time which is correlated with the organ quality and the compatibility with the recipient. When the surgeon declines an organ, it is implied that the perceived costs (such as decreased quality of life or decreased longevity) outweighed the benefits. In other words, the surgeon determined that recipient would be better off without transplanting the organ in question. However, this idealized concept is not necessarily true.

Prospect theory demonstrates that individuals give more weight to factors framed as potential losses (ie, costs) than to potential gains (ie, benefits).4 Hence, a transplant surgeon may overweigh the losses associated with accepting a kidney and hence, reject the organ even though the benefits are strictly greater than the costs. This behavior is termed loss aversion and has been demonstrated in a series of laboratory experiments conducted by the forefathers of prospect theory, Nobel laureates Kahneman and Tversky.4 If this behavior does exist in transplantation, surgeons may be discarding organs that would benefit the recipient; hence, perhaps partially explaining the underutilization of the donor pool.

In addition to loss aversion another behavior that may influence decisions is that of risk aversion where individuals choose a certain outcome over an outcome with less certainty (Figure 1). For example when confronted with the choice of taking a 5% chance of losing $10 000 or a 100% chance of losing US $501. However, 5% × −$10 000 = −$500 which is greater than US $−501. Risk aversion is the fear of larger loss which results in accepting the unfavorable settlement of the 100% chance to lose $501.

With risk aversion utility curve is concave. An individual has the choice between a guaranteed payoff and one without. In the guaranteed outcome, the individual receives US $50 (U(CE)). In the uncertain outcome, a coin is flipped to determine if the individual receives US $100 or nothing. The expected outcome for both scenarios is US $50. With risk aversion the individual will accept a certain payment of less than US $50 (eg, US $40, resulting in an RP of US $10, rather than taking the risk of 0. As a result, the utility curve with risk aversion becomes concave. CE, certainty equivalent; E(U(W)), expected value of the utility (expected utility) of the uncertain payment; E(W), expected value of the uncertain payment; U(CE), utility of the certainty equivalent; U(E(W)), utility of the expected value of the uncertain payment; U(W0), utility of the minimal payment; U(W1), utility of the maximal payment; W0, minimal payment; W1, maximal payment; RP, risk premium.

We describe 2 areas where loss aversion may be found in transplantation. One is the impact of using the KDPI as a tool to assess the quality of a kidney. The second is the potential impact of regulatory sanction on a transplant center's decision process to accept organs. In particular, we will focus on the impact of high KDPI on the discard rate of kidneys and the value of SRTR adjustment to control for risk of adverse outcomes. We also discuss briefly the potential impact of risk aversion on organ acceptance.

High KDPI kidneys

Rao et al7 proposed the Kidney Donor Risk Index (KDRI) for deceased donor kidneys to quantify graft failure risk and to support the complex decision of organ acceptance at the time of organ offer. The following donor characteristics are used to calculate the KDRI: age, height, weight, ethnicity, history of hypertension, history of diabetes, cause of death, serum creatinine, hepatitis C status, and donation after cardiac death status. The KDPI maps the KDRI to the cumulative percentage scale compared with a reference population of donors (all donors in the United States from whom a kidney was recovered during the prior calendar year) to rank order the quality of kidneys. In the new kidney allocation system introduced in December 2014, KDPI greater than 85% replaced the extended criteria donor (ECD) status, and these kidneys are allocated on a regional basis to increase utilization. Approximately 14% of all kidneys recovered from deceased donors have KDPI greater than 85%, and the discard rate for these kidneys is 56% compared with 17% for the kidneys with lower KDPI. It should be noted that the C-statistic for the KDRI (and thus the KDPI) is modest (0.63) and, from a statistical perspective, does not provide sufficient discrimination as a reliable metric for an individual decision to accept or turn down an organ. Thus, the KDPI confers a behavioral reference point where linearity of perceived risk is broken. By framing KDPI greater than 85% as needing consent, we implicitly create a reference point of high risk and thus allow the asymmetry of loss aversion to take place.

An analysis by Hirth et al8 might serve as a warning that the process of “negative framing” can have a detrimental effect on transplant rate. Using the SRTR database and logistic regression analysis of the transplant rates, the authors showed that even after risk adjustment, with the adoption of the ECD classification in 2006, there was a significant decrease in the transplant rate for donors older than 60 years. In a more recent analysis, Bae et al9 demonstrated a higher discard rate after the introduction of the KDPI. This study analyzed the discard rates pre and post KDPI for standard criteria donor (SCD) kidneys with KDPI greater than 85%. They found that the adjusted odds ratio was 1.42 (95% confidence interval, 1.07-1.89) for discard of high KDPI SCD kidneys after adoption of the KDPI label by UNet. The authors conclude that there is a harmful effect of labeling an SCD kidney with high KDPI resulting in an increased discard rate.

Moreover, analyses have demonstrated that there is a survival benefit with using high-KDPI kidneys. Massie et al10 used a time-dependent model to assess the mortality risk in the SRTR database associated with receiving a high KDPI kidney compared with a conservative approach of waiting for a low KDPI alternative. This analysis showed a survival benefit with accepting the higher KDPI kidney even in a model with KDPI greater than 90%. The time to survival benefit varied between 7.7 and 19.8 months dependent on the KDPI threshold between 70% and 90%. Benefits for high KDPI was greatest for recipients older than 50 years and at centers with median wait time of 33 months or longer. They concluded that KDPI greater than 85% alone should not be used as a reason to reject a deceased donor kidney. This behavior is at odds with maximizing value, thus suggesting application of the concept of loss aversion.

Given the long waiting time for kidney transplantation resulting in increased death on the waiting list, the evidence would suggest that loss aversion as opposed to maximizing value may be influencing at least a portion of the decisions regarding organ acceptance or turn downs.

Regulatory Oversight

The SRTR uses risk adjustment to analyze expected patient and graft survival after kidney transplantation. This risk adjustment model is used to create program-specific reports (PSRs) for each center’s outcomes at regular intervals. This risk adjustment includes several recipient and donor characteristics. However, some potentially important characteristics, such as donor biopsy findings, are not included in the model. The C statistic for the risk adjustment model is 0.66 which is very modest. Centers that have low performance because of inferior 1-year patient or graft survival can be flagged by the UNOS Membership and Professional Standards Committee to identify programs with lower than expected outcomes over a rolling 2.5-year cohort.

PSRs are also used by Medicare for center approval and by private insurers to determine which programs are included in their preferred provider networks. Flagging for lower than expected outcomes carries with it a potential economic loss to a center. This loss may be far greater than the gain of performing a greater number of transplants due to the asymmetry in the “cost” of loss versus the “cost” of gain. Thus, even a low chance of flagging due to its high-cost skews decisions toward loss aversion, even at the expense of expected total value.

In addition, PSRs are framed almost entirely by a center's chance of being flagged, and thus centers and physicians are “nudged” to think in terms of loss rather than gain (that is the structure of PSRs and quality improvement defaults to avoid flagging rather than to being excellent).11

Schold et al12 studied the influence of PSRs on the volume of kidney transplants over time. The primary finding was a significant association between outcome of PSRs and kidney transplant volume. The low performing programs saw a decrease in both standard and ECD transplants (decrease 4.7%) while the non-low performing programs saw an increase in ECD donor transplants (increase 3.9%) during the same period.

In another study, Schold et al13 also showed an association of candidate removal from the waiting list and center oversight as measured by the PSRs. This analysis showed that centers with low performance on the PSR had higher waitlist removal rates and a lower mortality risk in the patients removed from the waitlist, suggesting a relatively healthier cohort being removed from the waitlist in the low-performing centers.

The KDPI derives only an aggregate association with survival, with a very modest C statistic (0.63); therefore, the KDPI should not be the sole reason to discard a kidney except perhaps in patients with extremely low estimated posttransplant survival scores. Even if the C statistic were high, this would not justify discarding a kidney that might increase survival for an individual patient. Regulatory worries are a deterrent; however, given the low C statistic of KDPI and the overall low correlation coefficient of the risk model, regulatory concerns may be lessened due to reversion to the mean. Reversion to the mean is a mathematic principle used in many fields (such as finance), whereby extreme events from the mean will tend to come closer to the mean upon repeated events. In the case of expected outcomes, extremely high expected event rates will tend to revert closer to the population mean and extremely low event rates will be higher as they come closer to the population mean. Given this, high expected events should place you at no more (and likely less) risk than very low expected event rates. As mentioned, because the C-statistic for the expected outcome model is low, the risk adjustment model may not fully account for the risk taken and may potentiate risk averse behavior. This exacerbation of risk aversion may not produce optimal benefit to patients or transplant centers.

In a recent analysis of the SRTR database, Snyder et al14 analyzed the impact of transplanting a higher volume of high-KDPI kidneys on individual PSRs. Despite a clear relationship of increasing KDPI with worse patient and graft survival, there was no relationship of an individual program's utilization of high-KDPI kidneys and the program's PSR after risk adjustment. By showing that the risk adjustment accounts for the negative impact of transplanting high KDPI kidneys on graft outcome, this should reassure centers on the potential of using these organs for the appropriate recipient. However, the loss aversion resulting from labeling as high KDPI has a potent impact on the decision to accept these suboptimal organs.


The classical economic theory states that in an individual's decision process will optimize value within the bounds of budget constraints. However, modern psychology as well as experiments in behavioral economics challenges this concept, which have relevant implications in transplantation. The process of accepting kidneys from deceased donors is complex involving a multitude of factors including recipient factors, donor factors, and available center resources. The perceived high cost of taking a perceived high risk likely results in loss aversion and a higher discard rate. Hence, we suggest that the transplant community be aware of the potential impact of the psychological influences demonstrated by the concepts of behavioral economic theory so that those making decisions on organ acceptance acknowledge that these forces are at work.

If we accept the published evidence on the impact of labeling of kidneys as higher risk and on the behavior of programs when they are concerned about the risk of regulatory sanction, we see that programs become more conservative and are more likely to discard suboptimal kidneys. With the severe shortage of kidneys for transplantation it is essential that we optimize the utilization of these higher-risk kidneys to further increase life years gained and quality of life for patients with end-stage renal disease.

Another possible unintended consequence of loss aversion behavior that comes with flagging for low performance is not accepting certain group of patients to the waiting list who are higher risk. The intention may be to avoid taking recipients who may have risks that are not well adjusted in the current SRTR model, but other higher-risk patients whose risks are well adjusted currently in the SRTR model may also be denied the opportunity for transplant. Currently, we do not have data on the number of patients listed as a percentage of patients evaluated at a center (which should be tracked by SRTR in our opinion as a quality metric) but we suspect that this ratio goes down after flagging, detrimental to that group of high-risk recipients.

Published data confirm the adverse effect of labeling a kidney as high-KDPI resulting in a higher discard rates.9 However, because the C-statistic for the KDPI is modest at best and was not intended to provide sufficient discrimination as a reliable metric for an individual decision to accept or discard an organ, the higher discard rate indicates a negative framing effect. We would also urge regulatory bodies to be more aware that by framing wording around negative outcomes, they are in essence nudging stakeholders into risk adverse behavior. Our current regulatory system likely did not take these principals into account and thus unintended negative consequences inadvertently may have arisen.

The current efforts by US Health Resources and Services Administration funding the transplant Collaborative Improvement and Innovation Network project to increase the utilization of higher KDPI kidneys by disseminating the best practices and evaluating alternative means of monitoring are an acknowledgement of this need. This endeavor is one with which we agree and which should be applauded. We are as human beings all prone to these psychological principles, and to move our field forward, we must with great humility acknowledge our limitations and create an environment that does not accentuate our tendency to miscalculate value.


No title available.


1. Hart A, Smith JM, Skeans MA, et al. Kidney. Am J Transplant. 2016;16(Suppl 2):11–46.
2. Garonzik-Wang JM, James NT, Weatherspoon KC, et al. The aggressive phenotype: center-level patterns in the utilization of suboptimal kidneys. Am J Transplant. 2012;12:400–408.
3. Klassen DK, Edwards LB, Stewart DE, et al. The OPTN deceased donor potential study: implications for policy and practice. Am J Transplant. 2016;16:1707–1714.
4. Kahneman D, Tversky A. Prospect theory—analysis of decision under risk. Econometrica. 1979;47:263.
5. Schnier KE, Turgeon N, Kaplan B. A primer of neoclassical (traditional) and behavioral economic principles for organ transplantation: part 2. Transplantation. 2015;99:2247–2251.
6. Schnier KE, Turgeon NA, Kaplan B. A primer of neoclassical (traditional) and behavioral economic principles for organ transplantation: part 1. Transplantation. 2015;99:2024–2028.
7. Rao PS, Schaubel DE, Guidinger MK, et al. A comprehensive risk quantification score for deceased donor kidneys: the Kidney Donor Risk Index. Transplantation. 2009;88:231–236.
8. Hirth RA, Pan Q, Schaubel DE, et al. Efficient utilization of the expanded criteria donor (ECD) deceased donor kidney pool: an analysis of the effect of labeling. Am J Transplant. 2010;10:304–309.
9. Bae S, Massie AB, Luo X, et al. Changes in discard rate after the introduction of the Kidney Donor Profile Index (KDPI). Am J Transplant. 2016;16:2202–2207.
10. Massie AB, Luo X, Chow EK, et al. Survival benefit of primary deceased donor transplantation with high-KDPI kidneys. Am J Transplant. 2014;14:2310–2316.
11. Thaler RH, Sunstein C, Nudge R. Imrpoving Decisions About Health, Wealth, and Happiness. New York, NY: Penguin Books.
12. Schold JD, Buccini LD, Srinivas TR, et al. The association of center performance evaluations and kidney transplant volume in the United States. Am J Transplant. 2013;13:67–75.
13. Schold JD, Buccini LD, Poggio ED, et al. Association of candidate removals from the kidney transplant waiting list and center performance oversight. Am J Transplant. 2016;16:1276–1284.
14. Snyder JJ, Salkowski N, Wey A, et al. Effects of high-risk kidneys on scientific registry of transplant recipients program quality reports. Am J Transplant. 2016;16:2646–2653.
Copyright © 2017 Wolters Kluwer Health, Inc. All rights reserved.