Deceased donor organs in the United States are absolutely scarce- demand for transplantation greatly exceeds the supply of donors, and thousands of patients die on the waiting list each year. Deceased donor organ allocation policies take on the terrible task of distributing the limited supply of organs to the large pool of waiting candidates. Our Healthcare Ethics and Allocation Lab (HEAL) applies advanced empirical methods to evaluate and design organ allocation systems according to the underlying ethical principles.

Our Healthcare Ethics and Allocation Lab (HEAL) has multiple ongoing projects across all major organs. Below are the specific aims of a few examples:

1. Heart: Design a novel heart allocation score using deep learning. The current heart allocation system relies mainly on subjective treatment choices to rank-order candidates on the waitlist. Our lab is working to develop an objective score that is resistant to manipulation.

2. Liver: Determine the impact of the new "acuity circles" geographic sharing on the lives saved with liver transplantation. In response to legal actions by patients, geographic sharing of livers recently shifted from gerrymandered local regions (called "donor services areas") to 250 nautical mile rings around the donor hospital. This project aims to quantify the impact of this policy shift on the survival benefit of liver transplantation.

3. Lung: Design an alternative lung allocation score that is based on survival benefit. In contrast to the other major organs, lung transplantation does not improve survival in all patients with end-stage lung failure. Some patients, specifically those with Chronic Obstructive Pulmonary Disease, have worse survival with a transplant than without a transplant. However, the current lung allocation score overweights urgency (risk of death on the waitlist), leading to a system that does not maximize lives saved. This project aims to construct a new lung allocation score with different ethical objectives.

4. Kidney: Determine the accuracy of kidney allocation prediction models in a modern cohort. Kidney allocation relies on two prediction models- the Kidney Donor Profile Index (KDPI) and the Estimated Post-Transplant Survival (EPTS). Despite both models' critical role in allocating life-saving kidneys, the Organ Procurement and Transplant Network (OPTN) has not updated either in over a decade. This project aims to quantify the decaying accuracy of these models over time and demonstrate how updating the derivation cohorts can substantially improve allocation efficiency.

All projects will use the Scientific Registry of Transplant Recipients, a complete national registry of all US candidates, donors, and recipients. There are three principal empirical methodologies the lab employs.

1. Policy evaluation with advanced causal inference techniques from health services research

2. Machine learning to design novel allocation scores

3. Simulation modeling to evaluate the potential impact of novel allocation policies.

In parallel to these empirical techniques, the lab relies on a deep philosophical understanding of the ethical principles that guide the allocation of scarce healthcare resources.

Most projects use the open-source version of R with the RStudio IDE. Machine learning projects are typically in python. We also occasionally use Stata to estimate specific models. All software is either free or provided by the mentor.

International Society for Heart and Lung Transplantation

American Society for Transplantation