Liv-er Die: An Exploration of Bias in Liver Allocation Simulation

Abstract

This research presents an analysis of bias within allocation of liver transplants overseen by the United Network of Organ Sharing (UNOS) policies. In the past ten years, improving transplantation techniques, immunosuppression therapy, and other medical advances, in tandem with developments in data manipulation and prediction techniques have opened up new possibilities for modeling organ matching and accelerating the speed at which donor organs can be delivered to patients. With this potential, there is increasing pressure for policies to be in place to guide organ allocation decisions to save as many lives as possible while ensuring equity across all waiting list candidates. This project uses simulation techniques to model liver transplant allocation using data from the Organ Procurement and Transplantation Network (OPTN). It seeks to understand the impacts of implementing policies directly motivated to reduce bias between key waiting list candidate characteristics (age, gender, and BMI) on overall waiting list mortality rates. The results of the simulation reveal a relationship between the elimination of bias and a decrease in overall death rates on the wait list. These results raise the critical ethical questions surrounding the balance between equity and utility in organ allocation. While the results found here are not an answer to these questions by any means, they probe at the possibility of finding a balance between justice, utility, and respect for every human life.