Kidney donation is a reasonable choice for effective altruists and more should consider itpost by Tom_Ash · 2014-11-15T11:53:04.671Z · EA · GW · Legacy · 44 comments
Table of Contents Summary Benefits Direct Benefits of Kidney Donation Indirect Benefits Costs Long-Term Risk Immediate Financial Costs of Kidney Donation How to Evaluate the Donation Choice as an Effective Altruist Conclusion None 44 comments
This article was written by Thomas Kelly and Josh Morrison, who asked me to format and post it on the Forum. It makes a detailed case which I'd encourage people to read - it's made me seriously consider donating, since I hadn't realised the health costs were so low. People in the UK interested in donating can do so here; people in the US can do so here.
This piece makes the argument that kidney donation is a reasonable choice for many effective altruists. For some relevant background information, I donated my kidney through the National Kidney Registry in September 2014. I’m also the cofounder of Waitlist Zero, a nonprofit that seeks to end the kidney shortage by promoting support for living kidney donors and increasing living donation rates. This is also going to be more relevant to potential donors within the United States as I understand more about kidney donation in this country.
I encourage EAs who are considering kidney donation to talk with me or my co-founder Josh (also a kidney donor). We definitely understand it’s not a choice for everyone, and we aren’t interested in making a hard sell to become a donor. But we are interested in knowing what motivates people to consider donation, and we also have ideas on how to maximize the (significant) public awareness and political benefits of individual non-directed donations.
Table of Contents
a. Direct Benefits of Kidney Donation
i. Benefits of an Individual Transplant
ii. Kidney Chains
b. Indirect Benefits
i. Is Kidney Donation Contagious?
ii. Political Efficacy of Donor Support Movement
a. Surgical Discomfort, Inconvenience, and Risk of Complication
b. Long-Term Risk
i. No Decrease in Life Expectancy
ii. Increased Risk of Kidney Failure
iii. Minor Long-Term Income Effect
1. Financial Effect of Mortality Risk
2. Financial Effect of ESRD Risk
c. Immediate Financial Costs of Kidney Donation
4. How to Evaluate the Donation Choice as an Effective Altruist
a. How Does Donating Compare to Other EA Choices?
b. Now vs. Later
c. EA Identity and Movement Building
The direct benefits of kidney donation are somewhat difficult to calculate, but our best estimate is an increase of about fourteen quality-adjusted life-years saved with an optimistic estimate of twenty QALYs and a pessimistic one of six. Significantly higher indirect gains may be possible by influencing others to donate and by advocating for policies that would end the transplant shortage, though these gains are more uncertain.
Kidney donation is a significant but laparoscopic and broadly safe surgery with a hospital stay of a couple nights and a return to work within a few weeks. It does not decrease one’s life expectancy and causes long-term harms in less than 1% of cases. If one does not have paid medical leave, the direct financial costs of lost wages may be significant.
For a donor who doesn't forgo wages or pay significant travel expenses, the expected benefits to recipients are more than 100 times the expected costs to the donor (both calculated in QALYs).
The decision to donate is broadly consistent with EA principles. Different EAs will give different weights to the inconvenience, discomfort, and health effects of donation, and comparisons to effectiveness of monetary donations are challenging. That said, depending on subjective valuation of costs, many EAs may find kidney donation’s cost-effectiveness to be of a roughly similar order of magnitude to other EA interventions such as monetary donations to a GiveWell-recommended charity.
Direct Benefits of Kidney Donation
A kidney donor who gives to a stranger can cause multiple transplants. It is difficult to calculate the impact of each transplant, but it is clear they yield a significant number of life-years, with our best estimate as being about 14 quality-adjusted life-years.
Benefits of an Individual Transplant: A commonly used estimate from 2004 is that kidney recipients of a living donor can expect to receive an increase in 3.5 years of quality-adjusted life compared to remaining on dialysis. While there is some definite uncertainty in this figure, our estimate would be something closer to ~8 QALYs/transplant. The difference is due largely to technological improvement, lags in the data, sampling bias, and incongruity with deceased donation results.
Our Estimate: The half-life of a living donor transplant is 14.2 years. Patients are generally medically eligible for transplant when they have a survival prognosis of about 5 years on dialysis. If we were to take the experience of 60-64 year old patients as typical (since their median life expectancy on dialysis is 5.1 years), their average life-years gained per transplant though a transplant would be 8.9 (14.0-5.1). The quality adjustment used in the Matas paper discounts a transplant year of life at 84% and a dialysis year of life at 68%, though we think that estimated difference might be conservative, since dialysis is very burdensome. Using that discount, however, implies a QALY gain of 8.29 years (14.0*0.84-5.1*0.68).
Given the difference between our calculations and the Matas figure, we’ve tried to be conservative in our estimate. Nevertheless, we think there are good reasons for optimism.
1. Technological improvements have occurred in the past decade and are likely to be underrepresented in the data due to time lags in reporting. The Matas and Schnitzler 2004 figure relies on data from patients who were waitlisted or received transplants between 1995 and 1999, when 5.7% of kidneys were reported to fail within a year of transplant. Now 3.2% do. The 14.2-year graft-survival figure is itself incomplete because it refers to kidneys transplanted in the late 1990s, and survival rates have been increasing since then. For example, 2013 figures reported that the percentage of transplants that fail within five years of transplantation fell to 15.4% from 17.5% in the previously reported year.
2. Methodological Questions: The Matas figure only comes from recipients on Medicare, excluding healthier, wealthier recipients with private insurance who likely live longer. Note that the Matas figures also discount QALYs by 5% annually.
3. Kidney Exchange Chains: If you donate to start a chain (see below), your transplant may last longer: 3 year graft survival rates are about 1.5 percentage points higher in chains than in average living donor transplants, which makes sense since chains are able to deliver higher quality matches.
4. Deceased Donor Estimates: It is hard to reconcile estimates of deceased donor impact with QALY estimates as low as 3.5. Living donor kidneys last about 1.5-1.7 times as long as deceased donor kidneys. Estimates of deceased donor impact between 1999 and 2011 have given a range of 2-10 years with an unweighted average across studies of 4.31 years gained.
Kidney Chains: A non-directed kidney donor can usually facilitate more than one patient by facilitating chains of paired kidney exchange (exchanges between incompatible donor-recipient pairs). These chains of donation include a mean of 4.6 transplants, though the longest on record led to thirty-six transplants. That said, it is not correct to think of one non-directed donor as being the but-for cause of every transplant in the chain—an easy-to-match pair in the matching pool will likely become a part of some chain within a few months. Hard-to-match pairs are more difficult to calculate – some patients can only match with less than 1% of the population. Generally, chains alternate between easy-to-match and hard-to-match pairs. Donors with O blood type start significantly longer chains. Roughly guessing, we expect participating in a chain to cause 1.5-2 transplants that would not have otherwise occurred.
Combining our estimates, we'd guess that donating to start a chain leads to ~14 QALYs, though there is substantial uncertainty around the right figure to use with a range of 7-20 seeming realistic.
Is Kidney Donation Contagious? It’s possible that your living donation will persuade others to donate. If I hadn’t read stories about other people donating, I doubt I would have. Other non-directed donors I know have reported people they know personally deciding to donate in part due to hearing their stories.
Non-directed kidney donation has grown substantially over the last few years. However, a sizeable fraction of this increase may be attributed to a coordinated effort to encourage Orthodox Jews to become non-directed donors to other Orthodox Jews.
Kidney Donations to Strangers Per Year in US
In England, since 2009 when their transplant system began allowing non-directed donation and a small non-profit began publicly raising awareness, such donations increased from 0 to 10% of all living donation, compared to 3% in the U.S.
It is obviously difficult to estimate the probability distribution, but for an effective altruist who is reasonably open about their donation (e.g. perhaps talking to friends and colleagues about it privately or writing for a public venue), I think it's fairly likely that you would eventually convince someone else to donate.
Political Efficacy of Donor Support Movement: Non-directed kidney donors have a powerful story and unique moral authority for advocating political change for the transplant system. Policy measures exist that would end the transplant shortage and reduce government healthcare spending by providing support to donors (such as health insurance or an annuity) sufficient to clear the shortage between need and availability. Each year about 20K more people are added to the waitlist than receive a transplant, which means ending the shortage through living donation would save ~160K QALYs/yr. It is difficult to quantify one’s political impact, but for EAs who think this cause is worthwhile, donation should be more compelling.
Surgical Discomfort, Inconvenience, and Risk of Complication: Kidney donation is a laparoscopic surgery that typically necessitates a typical hospital say of two nights, prescription painkiller usage of about 5-8 days, and a return to deskwork within 2-4 weeks. For donors who do not have high blood pressure, the chance of death during surgery is about 1.3/10,000 – about the same as in childbirth. For donors taken as a whole, it's 3.1/10,000. Short-term complications affect about 2-5% of donors and typically involve things like wound infections.
The total QALY costs of this discomfort and risk are not large. The undiscounted QALY cost of a 3.1/10,000 risk to a person with 60 remaining years of perfect-quality life is 0.019 QALYs (roughly equivalent to a week of full value life). I haven't investigated the disability weights that should be applied, but assuming that a donor gets no value from their time in the hospital, half value from the following week, and ¾ value from the following 4 weeks before returning to full health, the total "QALY cost" of the operation, hospital stay, and recovery period is <2 weeks. Combining the risk of death and the hospitalization and recovery period, we get costs of ~0.06 QALYs for donating.
At face value, this may seem surprisingly low – donating a kidney is only as bad as giving up 3 weeks of life? Upon reflection, though, this makes sense – donors are able to return to work and most everyday activities within 4 weeks, and run only a very small risk of death.
Note that the above assumes incurring zero utility from donating, but that’s unlikely. Donors may have many positive experiences (e.g. reading, spending time with friends or family, developing interesting stories) during those 4 weeks and should expect a feeling of pride in the tangible and public accomplishment of donation even above what might be likely from a more abstractly-experienced financial gift.
Most of the costs that prevent people from donating in practice seem to be psychological ones (i.e. fear) rather than these literal incapacitation costs. Of course, these will vary across potential donors, and effective altruists may be particularly well-disposed to take them on.
No credible studies have demonstrated that kidney donation decreases life expectancy. However, it does increase the risk of kidney failure. The best data currently indicates that long-term medical harms affect less than 1% of donors.
No Decrease in Life Expectancy: Kidney donation has not been shown to measurably decrease long-term life expectancy. Donors live longer than the general population because they are healthier to start with. The remaining kidney grows after surgery to accommodate the function of the previous kidney, and chronic kidney disease tends to impact both kidneys equally. A study of World War II soldiers who suffered a traumatic injury that destroyed one kidney found them to have similar health outcomes to two-kidney’d soldier controls.
Increased Risk of Kidney Failure: The average American has a lifetime risk of kidney failure of 3.2%. Kidney donors must have very good kidney function pre-surgery, so their lifetime risk (if they do not donate) is about 0.14%. A 2014 study indicates that people who donate have a 0.9% lifetime chance of developing kidney failure.
The risk of kidney failure mounts over time and the risk of developing ESRD during one’s working years are much lower. At age 50, only 0.28% of living donors had experience kidney failure. As such the expected loss of income due to ESRD is very low. Kidney failure is typically the outcome of decades of degenerative kidney disease (in about half of cases this is caused by obesity and/or diabetes, so about half the risk is controllable by diet and exercise). Donating one’s kidney likely does not increase the chance of that chronic kidney disease; instead, donating reduces one’s total kidney function to about 70% of where it was before. Only donors with excellent kidney function are allowed to donate, so this does not typically have any health impacts. However, for those who develop chronic kidney disease in the future (a small number), donating reduces somewhat the buffer of excess function before chronic kidney disease leads to kidney failure.
There are several challenges to calculating long-term risk of living kidney donation. The first is that ESRD takes so long to develop. Given the medical advances in kidney donation, such as the switch to a laparoscopic procedure, it’s possible that long-term risks of donation have diminished for current donors but have not yet showed up in the data. That said, the long progression of kidney disease may also imply that studies will tend to underestimate the failure effect due to limitations in follow-up duration, and there may be significant uncertainty about health impact (trying to extrapolate to health effects thirty to fifty years in the future is very difficult). Finally, the significant majority of donors give to relatives making it possible that these related donors are actually at greater risk than unrelated donors (and because of transplant center selection criteria, non-directed donors are likely to be healthier to start with than the average donor).
While kidney failure for living donors remains tragic, living donors receive significant priority on the organ waitlist and are thus much more likely to receive a kidney transplant of their own.
Moreover, it seems appropriate to believe technological innovations like artificial organs are likely on a multi-decade timescale, significantly reducing the expected costs of risking ESRD in middle or old age. Based on this consideration (along with fact that living donors receive priority on the waitlist), I'm inclined to significantly discount the potential long-term negative health effects for donors who are young and healthy today.
A 1% risk of developing kidney failure in 30 years has a relatively limited QALY impact even if not discounted for technological improvement. If the average patient develops kidney failure at 65, and the average donor receives a transplant, they should expect to live about 11 years with a disability weight of 0.84, compared to about 15.5 for Americans without ESRD, yielding a total estimated QALY cost of 0.063 or about three weeks. Nearly any positive discount rate would reduce that cost (30 years in the future) substantially.
Minor Long-Term Income Effects: One argument that’s been made is that kidney donation could be on some level self-defeating from an EA standpoint because it reduces your long-term giving by an amount with greater impact than your donation. We disagree strongly with this position, and think the best estimate of long-term income loss is something less than ~$150.
· Financial Effect of Mortality Risk: If you donate $10K per year and have 40 years of remaining working life, the risk of death in surgery would reduce your donations by only an expected $120. Risks to income from potential kidney failure are unlikely because of evolving technology, but even if they occurred, they're likely to become an issue near retirement.
· Financial Effect of ESRD Risk: Long-term income effects of ESRD risk are likely negligible. Here’s a quick back-of-the-envelop calculation to demonstrate. Say there is a 1% risk of kidney failure, and 50% chances each of retirement, technological improvement, ability to receive a transplant, and defection from the EA cause (each conservative estimates). Say that if none of those conditions holds one would lose $250K in charitable donations thirty years from now. This means that the expected loss will be $156.25 (250,000*.01*.5*.5*.5*.5) in 2044, but this figure needs to be discounted to net present value to be compared to a choice today. If you invest a dollar today at 5% annual compounding interest, in thirty years, it would be worth $4.32. Thus, the long-term income loss expected from kidney donation would be equivalent to paying $36.17 (156.25/4.32) today.
Immediate Financial Costs of Kidney Donation
The costs of surgery and of any complications from surgery are covered by the recipient’s insurance (usually Medicare). However, financial costs you incur in donating (such as lost wages if you don’t have paid medical leave) are typically borne by the donor. In the United States, there is no federal law requiring paid leave for organ donors. However, federal employees receive paid leave, as do employees in many states and employees of companies with generous benefits in other states. The cost of donation for donors who do not receive paid leave is of course substantial. The other major cost of donation is travel, which is reimbursed by the National Living Donor Assistance Center for any donor whose recipient makes under 300% of the US poverty line, and is also under 300% of the poverty line. The National Kidney Registry also reimburses all costs for donors who donate to start chains.
There are meaningful costs that are not reimbursed in the United States. Homemakers who donate do not receive any money to pay for childcare, eldercare, or chores they cannot complete. In addition, there is no compensation for the pain or discomfort of the process.
For full-time employees with paid leave, the financial costs of living donation are greatly ameliorated.
How to Evaluate the Donation Choice as an Effective Altruist
The choice to donate a kidney to a stranger clearly has a powerful positive health effect to the recipients that is at least an order of magnitude larger than the cost to the donor. It fits effective altruist principles by being a gift that is significant to the donor and has a measurable and highly positive impact. This does not by any means imply that donating is the right decision for all (or even most) effective altruists since different people will value the costs of having surgery differently.
How Does Donating Compare to Other EA Choices? Donating a kidney is similar in direct impact to, but on the lower end of, other EA choices like giving money to a GiveWell recommended charity. Donating a kidney gets ~14 QALYs at a cost of something on the order of ~0.1 QALYs. If you can give a month of your life to extend someone else's by 14 years, that seems to be a good decision and one where a small sacrifice yields a significant gain for others. Believing that donating will encourage others to do so or that the returns to political advocacy are high could easily double these expected returns.
GiveWell estimates the cost-per life saved for their top charities as ~$3-5K. Assuming a healthy 50 year lifespan for those who are saved, this works out to $60-$100/DALY. GiveWell cautions not to take these cost-estimates literally, due to significant uncertainty around them.
A possible EA critique to kidney donation is to say that, "If donating a kidney will save ~14 QALYs and a developing world DALY only costs ~$100 to save, then donating a kidney is only worth $1,400 in donations." We think there is something true and important in this argument, and we wouldn't encourage someone to donate a kidney if it meant they would forego significant donations to GiveWell's top charities. But we don’t see why that should be the case, since giving a kidney is a complement to and not a replacement for monetary donation.
Put somewhat differently, "donating a kidney is only worth $1,400 in donations" takes our failure to donate the maximum we could and uses it to avoid orthogonal actions that would be extremely net beneficial but don't necessarily quite reach the benefit-cost threshold we've set for our financial donations. This would also imply an incorrect fungibility between moral choices and is a principle that could lead to worse behavior in the long-run (i.e. constantly forgoing positive-utility actions whose gains are not the same as charitable donations with the best expected value).
Donation may also have benefits personal to the donor, indirect to other transplant recipients, and supportive to the EA cause that are not accounted for in its direct QALY figure, and giving through non-financial donation may also be a useful diversification strategy for one’s altruism.
Now vs. Later: If one intends to donate a kidney, donating now is significantly more effective than donating much later. Currently, the transplant shortage is serious, but future policy or technological change could greatly decrease the marginal benefit of kidney donation. That same change could cause a donation today to be a bridge for the recipient into a period with significantly decreased mortality for kidney failure that they would not otherwise have survived to. Opportunity to influence future donors and the political system are likely also higher now than later.
From a risk perspective, donating when young is inferior to donating when old, because younger donors do not have as much information on future health conditions as older donors (i.e. a 50 year-old without chronic kidney disease will almost certainly never develop it, but a 25 year-old still could). For either group, the absolute risk of kidney failure is still small and less than that of the general population. However, we note that donors who are currently young are more likely to benefit from future advances in technology (such as artificial organs) that reduce the expected costs of donation.
EA Identity and Movement Building: One concern that has been expressed is that, since kidney donation is an unusual and serious choice, donation can marginally pull the EA movement away from the mainstream. This concern comes from a model of successful social movements as incremental and moderate.
A different model finds movements do better when they are dramatic, salient, and compelling. Kidney donation is a very admired choice with clear and identifiable beneficiaries. It is a visible choice and attractive narrative that demonstrates significantly stronger commitment than a non-binding pledge to donate money in the future. It also may open opportunities for a broader discussion of EA principles. The more EAs save the lives of identifiable strangers, the more notice the movement gets and the better it looks.
Additionally, much of the EA movement to date— GiveWell, Giving What We Can, and the earning to give discussion at 80,000 Hours—has focused on money as the core resource we have to allocate. We don't disagree that decisions about money are important, but we think that the EA movement will be stronger and more compelling to newcomers if it can offer a broader variety of ways to make the world a better place. Kidney donation should very much be considered as one of those activities.
Kidney donation confers direct benefits to recipients that clearly exceed the cost to the donor by a large factor (more than a hundred, according to our estimate). It is a tangible choice to significantly and publicly affirm one’s commitment as an effective altruist. For many effective altruists, kidney donation is a good choice, and as someone who identifies as an Effective Altruist, I’m glad I donated.
 Thanks to Josh Morrison for helping write this and to Ryan Carey, Howie Lempel, and Alexander Berger for reviewing it. All mistakes are Josh and my own.
 2012 OPTN Annual Report, figure 6.7. See 2012 Kidney Data for the exact figure. If you remove from consideration patients who died for reasons unrelated to graft failure, the half-life of a living donor graft that survives past the first year post-surgery is 26.6 years. OPTN Annual Report 2010
 2013 USRDS, p. 266. For comparison, the difference in life expectancy between 45-49 year olds on dialysis (8.3 years) and receiving a transplant (22.8 years) is 14.5 years. Id. These figures include living and deceased transplants and are thus likely to undercount the impact of a living donor.
 Dialysis involves several, uncomfortable treatments several times per week and leaves 80% of patients unable to work, whereas transplant patients can live a fairly normal life. It also has a very high rate of complication compared to transplant. “Adjusted rates of all-cause mortality are 6.5–7.9 times greater for dialysis patients than for individuals in the general population. For renal transplant patients, rates approach those of the general population, yet remain 1.0–1.5 times higher.” Id. To use a specific example, “female dialysis patients in their 30s, 40s, and 50s … are expected to live just one-fourth as long as their counterparts without ESRD. Transplant patients fare better, with expected remaining lifetimes 75–80 percent as long as those in the general population.” Id.
 Survival rates are projected forward from mortality rates in early years.
 OPTN Annual Report 2012
 2012 OPTN Annual Data Report
 Id. See also Ibrahim, NEJM and generally Muzaale, JAMA (2014). Note that a 2014 study indicated more serious health risks (Mjoen, 2013). But this piece had serious methodological problems-- for example, the mean age of the control group was 37.6 compared to 46.0 for the donors, and the study did not have BMI data for much of the control group. See Boudville, Kidney International and Kaplan, AJT for some of the limitations with the Mjoen piece.
 Muzaale, JAMA (2014). See Gill, JAMA (2014), Kasiske, AJKD, and Allegretti, AJT for critiques and commentary of the Muzaale piece. Kidney donation also increases the risk of preeclampsia (a disorder of pregnancy characterized by high blood pressure and large amounts of protein in the urine.) by 5-6%. Ibrahim, AJT, 2009. See also Potential Risks for Kidney Donation for a general overview on risk of donation.
 Muzaale, JAMA (2014)
 The 2014 Muzaale study found the risk of related donors developing ESRD within 15 years was 34.1/10,000 as compared to 15.1/10,000 in unrelated donors, but this result was not statistically significant.
 Discussed in Lavine, Science, Fountain, NY Times, and Krassenstein, 3D Printing News. UCSF is also trying to develop an implantable miniaturized dialysis machine. UCSF, 2013 and Kleffman, San Jose Mercury News.
Comments sorted by top scores.