ESS - Collaborative Research: Modelling the Patient's Perspective on Organ Acceptance

本研究项目的资金由服务业工程探索性研究(ESS)项目公告提供。器官共享联合网络(UNOS)管理国家器官捐赠和分配系统。当UNOS将器官提供给移植患者时，患者决定要么接受它，要么继续等待。因此，手头的研究问题是：如果提供器官，患者应该在什么时候决定接受它？这个问题具有很大的现实意义，因为这是每年成千上万人面临的生死攸关的决定，而且还在不断增加。这项研究将只关注肝脏，但所使用的方法也适用于其他器官。对这个决策过程进行建模的关键问题是对患者可获得的信息做出的假设。在所有情况下，在提出报价时，肝脏的相关特征是已知的，患者的健康也是如此。不确定的是患者相对于等待名单上的其他患者的排名，以及未来提供服务的时间。这个问题最简单的版本是活着的捐赠者存在的情况。然而，通常情况下，捐赠者是身体。在后一种情况下，肝脏供应是随机发生的，患者在不知道自己在等待名单上的相对位置的情况下相互“竞争”肝脏。这项研究项目将为活体捐赠者以及身体捐赠者的情况制定数学模型，其中的目标是最大化，例如，“质量调整后的预期寿命”或一些其他目标。将为等待名单的构成是公开信息的假设情况开发一个模型。具体地说，REACH团队将把这些情况描述为马尔可夫决策过程，并对产生的最优策略进行数值和分析研究。数值研究将利用大量的临床数据，确定将指导分析研究的模式，并允许研究人员探索公布等待名单的潜在价值。分析研究将确定最优政策具有吸引人的、一致的结构的条件。有趣的是，现有的大多数文献都集中在从UNOS的角度对器官移植问题进行建模，而不是从患者的角度进行建模。然而，请注意，这项研究可能会对这一政策视角的研究产生影响，因为研究的目标之一是研究涉及公布等待名单的政策变化。

Funding for this research project is provided under the Exploratory Research on Engineering the Service Sector (ESS) program announcement. The United Network for Organ Sharing (UNOS) manages the national organ donation and allocation system. When UNOS offers an organ to a transplant patient, the patient decides to either accept it or to continue waiting. The research question at hand is therefore: if offered an organ, when should a patient decide to accept it? This question is of great practical importance, as it is a life-or-death decision faced by thousands of people per year, and growing. This research will focus only on livers, however the methodology used is applicable to other organs. The key issue in modeling this decision process is the assumption made regarding the information available to the patient. In all cases, the pertinent characteristics of the liver are known at the time an offer is made, as is the patient's health. What is uncertain is the ranking of the patient relative to the other patients on the waiting list, and the timing of future offers. The simplest version of this problem is the case in which a living donor exists. Typically, however, donors are cadaveric. In the latter situation, liver offerings occur randomly and patients "compete" with each other for livers without knowing their relative position on the waiting list. This research project will formulate mathematical models for the living-donor as well as the cadaveric-donor situation, in which the objective is to maximize, for instance, "quality-adjusted life expectancy" or some other objective. A model will be developed for the hypothetical situation in which the composition of the waiting list is public information. Specifically, the reach team will formulate these situations as Markov decision processes and engage in numerical and analytical study of the resulting optimal policies. Numerical study will make use of extensive clinical data, identify patterns that will guide the analytical study, and allow the researchers to explore the potential value in publishing the waiting list. Analytical study will determine conditions under which the optimal policies have appealing, consistent structure. It is interesting to note that most of the existing literature focuses on modeling the organ transplant problem from UNOS's perspective, rather than from the patient's perspective as considered here. Note, however, that the research may have an impact on this policy-perspective line of research, since one of the objectives is to study the policy change that involves the publication of the waiting list.