Kidney Donations — Optimizing Them
Did you know that there is a country in which kidneys can be legally bought and sold? That country is Iran, and there are actually two models for payment. In one part of the country, donors receive a flat fee, and in the other part of the country prices are negotiated.
This point was actually an aside in John Dickerson’s talk, “Swapping Kidneys: Better Matching Market Design via Optimization” on algorithms for kidney swapping at Data Science DC Dec. 13, 2016, in Washington DC. Dickerson, a member of the University of Maryland’s Computer Science Dept., focused on non-monetary mechanisms for kidney exchange.
In most countries, a person with renal failure can get a kidney in one of two ways:
- By making it through the qualification and waitlist process to obtain a kidney from the “deceased donors list.”
- By obtaining a kidney from a friend, relative, or other personally-arranged contact.
Unfortunately, demand far outstrips supply for option 1. And in option 2, the probability is low that a willing donor’s kidney will be compatible for the recipient. Let’s say that patients A and B have identified willing donors 1 and 2 and neither of the matches is compatible, BUT A is compatible with 2 and B with 1. A simultaneous donation could then be arranged 1 > B and 2 > A.
Another approach starts with an altruistic donation — one not tied in advance to a particular recipient. The recipient of that donation is then selected from among donors who have arranged their own willing donor. That willing donor then donates to a different waiting recipient, also selected from among those who have identified a willing donor, and so on.
Both these approaches can be depicted in a network graph, and that graph can then be used in optimization algorithms to substantially increase the number of donations, compared to those that take place in an ad-hoc, more local environment.
One big obstacle remains. Due to the low probabiity of matches, for this system to work well a large, hence widespread, pool of willing donors is needed. However, hospitals and donation centers are reluctant to turn their donors over to a pool outside their control, preferring to keep matters “in house.”
Dickerson’s excellent talk is worth a look — get an overview of the algorithms that underlie this new approach that has the potential to save many lives.