Commentary: evaluating the efficiency of organ procurement - Public Policy Impact - the use of data envelopment analysis on Organ Procurement Organizations

Health Services Research, Oct, 1999 by Elizabeth Wrone

Meeting the increased demand for solid organs has emerged as a new challenge for transplantation. Advances in surgical technique, organ preservation, and anti-rejection therapy have made transplantation a life-saving option for thousands of patients. The care of patients with advanced chronic illnesses has also improved such that the number of potential recipients of solid organs has far outgrown the supply. For example, the number of transplantable kidneys has increased at a rate of 1.6 percent per year whereas the number of patients on the cadaveric renal transplant list increased 2.5-fold over the past decade (United States Renal Data System 1998 Annual Report). Examination of the current organ procurement process and its efficiency is the first step toward maximizing organ supply.

The organ procurement process is coordinated by Organ Procurement Organizations (OPOs), each with a designated service region. OPO involvement starts with a referral from a healthcare provider for evaluation of a potential donor. The procurement team assesses the eligibility of the potential donor, approaches the family for consent, and then organizes the retrieval and mobilization of organs. Community education and the fostering of relationships with practitioners in the hospital who initiate referrals, are two vital activities of OPOs that can affect the number of organs recovered.

Traditionally, the Health Care Financing Administration has assessed the performance of OPOs by measuring donor rates per million. However, the Association of Organ Procurement Organizations (AOPO) considers this performance evaluation to be a rough and potentially misleading representation of an OPO's effectiveness. Geographical variations in the demographics of donors and recipients, donor criteria, prevalence of chronic viral infections, cultural and racial variation in attitudes toward donation, use of advanced directives, and effects of population density contribute to difficulties in interpreting simple performance comparisons. The OPO must select from among a cohort of potential donors in order to yield suitable organs for transplantation. Three components of performance present challenges for an OPO that require unique resources and skills at each step: (1) identifying possible donors, (2) achieving family consent for the subgroup of those deemed suitable, and (3) surgically retrieving organs for transplantation. Although donation rates represent the overall effectiveness of a procurement system, they do not capture the effectiveness of the OPO at each crucial step in the process.

Expanding performance analysis to include utilization of resources improves on donation rate comparisons. A variety of statistical and economic techniques are available for more complex assessments. Standard analytic techniques, such as statistical regressions or least squares estimates, show how the distribution of one variable is affected by another and are useful for predicting the output of a system with a given set of inputs. Regression analyses, which rely on the mean and standard error of each input and draw an imaginary line through the middle of multidimensional clusters of points, have not been used extensively to evaluate transplantation programs. A relatively new analytic tool for economic efficiency modeling is data envelopment analysis (DEA). Developed in 1979 by Charnes, Cooper, and Rhodes, DEA has found applications in evaluating the relative efficiency of not-for-profit organizations, governmental units, and health services (Charnes, Cooper, and Rhodes 1979). DEA relies on the most efficient units in a data set by "enveloping" them, and it draws an imaginary line around the optimal points called the "efficiency frontier." Each service unit is then compared with the benchmark units on the frontier and given an efficiency score. Adjustments for factors, such as size and volume of activity, can be built into the model as well. The advantage of DEA over regression techniques is that it generates information for each inefficient unit that may suggest which inputs can be reduced or outputs increased to improve overall efficiency.

The study by Ozcan, Begun, and McKinney in this issue evaluates the relative efficiency of OPOs in the United States using DEA. Based on a data set built from a nationwide survey of executive directors of OPOs and secondary data from the AOPO and the United Network for Organ Sharing, the model identifies an efficiency frontier from a set of OPOs with optimal performance. Production of the OPOs was represented by two outputs: kidneys recovered and extrarenal organs recovered. Available resources of the OPOs were represented by four inputs concerning hospital development activities. The first input is an index of hospital development that incorporates a hospital development director, a transplantation department, and written standards for effectiveness. The other inputs include fulltime equivalent hospital development staff, other labor, and non-hospital development operating expenses. For this model, the number of referrals was considered to be independent of an OPO's activity, yet important for the efficient procurement of organs; the authors included it as a nondiscretionary variable.