I’m at a fascinating luncheon talk by Al Roth at CRCS with the irresistable title, “Kidney Exchange.” This can’t help but call to mind “Anonymous Philanthopist Donates 200 Kidneys“, in The Onion. which I hope Al has in this talk or puts in his next one.
So I’m taking notes here. Lots of good fodder for the Markets chapter of the book I’m wrirting.,,
Market Design is Al’s summary and collection of works on that subject, including Kidney Exchange. See his “Efficient Kidney Exchange” paper In the American Economic Review. Market design is around patient-donor pairs (usually relatives), even though one half the pair will not be donating a kidney to the other half of the pair. The two surgeries: transplants and nephrectomies. The latter is the removal of a kidney.
NEAD: Non-simultaneous, Extended Altruistic-Donor Chain. These can add up. Al shows how ten transplants came out of one donor starting the chain. In cost/benefit terms, this is also good. This is now cited in both the New England Journal of Medicine and People magazine.
There are currently ~86,000 people on the waiting list now. There is a question on the floor about calculations based on 100% of Americans on the donors list. Nor feasible, but one wonders if a much larger percentage could be recruited. Al calls these “non-directed donors.”
Wondering what kind of matches we might classify with personal RFPs. (Al is talking here about pairwise ones, because that’s how kidney donorship works.) Other considerations: “Markets must be thick, uncongested and safe.” Need to look at the long and short side of a market. Also matching efficiencies…
Interesting point: relationships between hospitals and surgeons are a context. Also that we are starting to see pairs withheld, meaning that a hospital or two may not inform other hospitals about donor candidates. Arcanum: strategyproof industry rational (IR) mechanism.
(I’m surprised that all this stuff, most of which is new to me, does not hurt my brain.)
When creating the large exchange, respect the small exchanges that operated independenly before. “Mechanisms that give priority to internally matachable pairs have good incentive and efficiency properties in large markets. Theorem for k=2 (pairwise exchange), full participation is an E (the Greek letter epsilon) equilibrium for E(n) = 0(1/n).
“It could be that some regulation in order.” e.g., “If you participate, you must show us all our patients.”
Why do we have laws against simply buying and selling kidneys? The same reason we have the Ontario Dwarf Tossing Prevention Ban Act of 2003. The answer is, because it’s repugnant.(At least to lawmakers.) X may not be repugnant, but X+$ is repugnant — n some conditions, anyway, such when selling kidneys. Or kids. (Hmm… kidney sounds like a diminutive term for a kid. Or a kid part.)
Three repugnancies: Objectification, Coercion and Slippery Slope. (Hmm, don’t we have that in lots of business settings already? Such as excessive gathering user data online?)
Leave a Reply