New England’s kidney exchange: matching incompatible donor–recipient pairs

In 2003 roughly 60,000 Americans were waiting for a kidney; only 8,665 deceased-donor transplants happened that year, and 3,436 people died waiting. Live donation could fill some of the gap — a healthy person has two kidneys and can live with one — except that a willing donor is frequently the wrong blood type or tissue match for the specific patient they love, and money cannot legally fix that: buying or selling organs is illegal everywhere transplant medicine is regulated.

A handful of hospitals had begun swapping kidneys between two incompatible pairs — my incompatible donor gives to your patient, yours gives to mine — and Johns Hopkins had even managed a few three-way versions. But as of December 2004 only five such exchanges had happened across all fourteen transplant centers in New England, because no one had built a database of incompatible pairs to match against; unmatched donors were simply sent home.

Treated each incompatible pair as a node in a matching market, then searched for cycles and chains where mismatches cancel, turning scarcity into an allocation problem.

Each incompatible pair was a node, not a failure, so mismatches could cancel out.

A registry let the algorithm find cycles and chains across all pairs at once.

Non-simultaneous chains, started by altruistic donors, removed the need for simultaneous surgery.

Transplants grew with pool size, not new kidney supply, so the shortage became an allocation problem.

When direct exchange is blocked, build a clearinghouse that matches complementary mismatches across many parties.

Thousands of transplants that could not otherwise happen; the 2012 economics Nobel cited the work.