Santos Arteaga Francisco Javier, Di Caprio Debora, Cucchiari David, Campistol Josep M, Oppenheimer Federico, Diekmann Fritz, Revuelta Ignacio
Faculty of Economics and Management, Free University of Bolzano, Piazza Università, 1, 39100, Bolzano, Italy.
Department of Mathematics and Statistics, York University, Toronto, Canada.
Health Care Manag Sci. 2021 Mar;24(1):55-71. doi: 10.1007/s10729-020-09516-2. Epub 2020 Sep 18.
The main applications of Data Envelopment Analysis (DEA) to medicine focus on evaluating the efficiency of different health structures, hospitals and departments within them. The evolution of patients after undergoing a medical procedure or their response to a given treatment are not generally studied through this programming technique. In addition to the difficulty inherent to the collection of this type of data, the use of a technique that is mainly applied to evaluate the efficiency of decision making units representing industrial and production structures to analyze the evolution of human patients may seem inappropriate. In the current paper, we illustrate how this is not actually the case and implement a decision engineering approach to model kidney transplantation patients as decision making units. As such, patients undergo three different phases, each composed by specific as well as interrelated variables, determining the potential success of the transplantation process. DEA is applied to a set of 12 input and 6 output variables - retrieved over a 10-year period - describing the evolution of 485 patients undergoing kidney transplantation from living donors. The resulting analysis allows us to classify the set of patients in terms of the efficiency of the transplantation process and identify the specific characteristics across which potential improvements could be defined on a per patient basis.
数据包络分析（DEA）在医学领域的主要应用集中于评估不同卫生机构、医院及其内部科室的效率。一般而言，不会通过这种编程技术来研究患者在接受医疗程序后的病情演变或他们对特定治疗的反应。除了收集这类数据本身存在困难外，使用一种主要用于评估代表工业和生产结构的决策单元效率的技术来分析人类患者的病情演变似乎并不合适。在本文中，我们将说明实际情况并非如此，并实施一种决策工程方法，将肾移植患者建模为决策单元。这样一来，患者会经历三个不同阶段，每个阶段都由特定且相互关联的变量组成，这些变量决定了移植过程的潜在成功率。DEA应用于一组在10年期间获取的12个输入变量和6个输出变量，这些变量描述了485例接受活体供肾移植患者的病情演变。由此产生的分析使我们能够根据移植过程的效率对患者群体进行分类，并确定可以针对每位患者定义潜在改进措施的具体特征。
