The Economic Analysis of Kidney-Exchange Networks

This dissertation investigates the effects of the largest national kidney-exchange network, National Kidney Registry (NKR), on the speed of finding a transplant and the quality of donors in participant hospitals. It also examines the effect of surgeons in the network-adoption decision of hospitals. The prohibition of monetary transactions for human organs under U.S. law generates a shortage of kidneys available for transplant. In 2017, about 100,000 patients were waiting on the long wait-lists to receive a kidney for transplant. Creation of distinct kidney-exchange networks that find compatible matches between patients who each have willing but incompatible living kidney donors reduced this shortage. This dissertation uses the data from the scientific registry of transplant recipients and the National Kidney Registry data on the list of participant hospitals by year. The first chapter estimates the change in the probability of receiving a transplant conditional on wait-time for patients after a hospital adopts the NKR network. Using survival analysis accounting for the competing risks, I find that the probability of finding a transplant from a living donor increases by 0.25 percentage points in hospitals participating in the network. This positive effect is mainly driven from the additional indirect-living transplants that these networks can accommodate through exchange transplants. The second chapter investigates how the quality of living donors changes as the use of the NKR network expands. I use the variation in the period before and after the adoption of NKR by hospitals and run a difference-in-differences method. Further, I use a Coarsend Exact Matching to correct for the imbalance between treatment and control groups. My finding suggests that the quality of living donors as measured by age, body mass index, and blood type decreases in participant hospitals. Specifically, I show that living donors in NKR affiliated hospitals are on average 8 months older, have 0.19 points higher body mass indexes, and are 3.8 percent less likely to have an O blood-type. The final chapter analyzes the fragmentation in the participation of hospitals in these networks. We (with Bobby W. Chung) investigate the influence of surgeons in expanding the use of the NKR by hospitals. We find that hospitals that are connected through mutual surgeons are more likely to adopt the NKR network. Specifically, we find that one more adoption by connected hospitals increases the probability of the focal hospital to adopt by about 4 percentage points. This trend shows a diminishing magnitude as the number of connected hospitals increases. This effect is stronger for surgeons that have performed a larger number of transplant surgeries and for hospitals that have more than one mutual surgeon