Medical Residency Matching Process
In class we have extensively talked about matching markets, and how the use of certain valuations can help to create a match that is socially optimal. There are many examples of matching markets in the real world, including Uber and Lyft for matching travelers to drivers and Airbnb for matching homeowners to renters and vacationers. A particularly interesting matching market is the medical residency market. Medical school graduates are matched to residency programs based on the preferences on both sides. The residency system was not always a matching market. Between the 1900s and the 1940s, hospitals would compete by signing contracts with medical students before other hospitals to claim the best talent. This style of hiring resulted in a process that got more and more expedited, until the process was changed to a matching market. At the worst point, medical students were signing residency offers 2 years in advance of graduating from medical school.
Today, the process works using a computer algorithm. Students rank their top choices for residency programs, and residency programs rank their top students for their open positions. The algorithm goes down each student’s preference list and tentatively matches them to a position. If there is another student who is ranked higher by the residency program, they will replace the tentatively matched student and the program will continue going down their ranking list. Once the program finishes iterating through all of the preference lists, the tentative matchings become finalized. The result is a stable matching that prioritizes the preferences of students. No student would want to be switched into a worse program further down their list, and no residency program would want to hire someone lower on their list. The dominant strategy for this process is also to rank truthfully, otherwise the chance of getting an unwanted match increases.
http://stanford.edu/~alroth/jama.html