Wednesday, November 18, 2009

Gaming the Liver Transplant Market (by Jason Snyder)

When my colleagues and I began talking to transplant surgeons about the design of kidney exchanges, it was initially sometimes hard to convince them that incentives played a big role in organ allocation. (I sometimes heard a variation of "Professor, incentives may be important in economics, but not in medicine; no one chooses to become sick.") But explanations were made easier by a 2003 legal settlement in which some hospitals paid fines for pretending their patients were sicker than they were, to give them increased priority on the waiting list for deceased donor liver transplants: Illinois: Prosecutor's Diagnosis Is Fraud.

By the time of the settlement, the rules for determining priority on the waiting list for livers had already been changed to depend on more objectively verifiable criteria, to reduce the ability of hospitals to game the system on behalf of their patients. A recent paper by Jason Snyder of the UCLA Anderson School of Management looks at the effect of this change:

"Gaming the Liver Transplant Market" Forthcoming at The Journal of Law, Economics, & Organization
"Approximately 6,000 transplants are performed annually and, on average, 2,500 people die while waiting for a liver. There is substantial variation in the number of transplant centers across markets; some markets have only one firm while other markets have multiple participants. Prior to March 1, 2002, a major determinant of whether a patient would obtain a liver was whether they were in the intensive care unit (ICU). Patients in the ICU jumped to the top of the priority list regardless of how sick they actually were. There is considerable anecdotal evidence suggesting that in order to obtain livers for their patients the transplant centers created faux-ICUs where relatively healthy people were put in the ICU to strategically advance their positions on the waiting list. After March 1, 2002, the allocation of livers changed to a system where livers were allocated solely on clinical indicators of sickness. ICU status was no longer a factor in determining whether a patient obtained a liver or not. This policy resulted in, if anything, an increase in the sickness of the average patient at transplant and a dramatic discontinuous decrease in the number of patients who were in the ICU at the time of their transplant. This seemingly contradictory behavior is consistent with centers strategically misrepresenting the health of their patients prior to the policy reforms.

"Using the policy change to examine changes in ICU admission behavior, I find that after the policy changed the use of the ICU decreased more in markets with more firms. I also find that after the policy changed the percentage of relatively healthy people in the ICU decreased more in markets with more firms. Finally I show that these results are non-linear in the number of firms in the market. Moving from one firm to two firms in the marketplace is associated with dramatic changes gaming behavior, but there is little difference between two firms and three or more firms."

No comments: