PBDS: A Decision Support System for Regional Blood Management

Human blood is a perishable product: in the U.S. it has a legal lifetime of 21 days from collection, during which it can be used for transfusion to a patient of the same type, and after which it has to be discarded. It is collected in units of one pint from volunteer donors at various collection sites such as a Regional Blood Center, and after a series of typing and screening tests it is shipped to Hospital Blood Banks in the region. Once at the Hospital Blood Bank, a unit is stored and is available to satisfy the random daily demand for transfusions to patients. Since not all units demanded for a patient are generally used, a unit can be issued several times during its lifetime until transfused or outdated. Efficient management of the regional blood resources is a difficult task. In addition to the complexity of the problem, introduced by the factors outlined above, the Regional Blood Center is faced with some typical characteristics of a Health Care management problem: (i) the performance of a regional blood management system can be evaluated in terms of more than one criteria, some of which are conflicting (e.g., shortages vs. outdates), and (ii) quantitative cost measurement of the system's performance, and, therefore, comparison of alternative policies using simple cost criteria, is very difficult since the estimation of many costs involved (e.g., unavailability of blood) is purely subjective. The two most common performance measures of a blood region are the shortage rate (i.e., the percentage of days when "supplementary" deliveries have to be made to satisfy a hospital's demand), and the outdate rate (i.e., the percentage of a hospital's supply that becomes outdated) for the hospitals in the region. The system described in this paper (PBDS) has been designed so as to provide a decision support mechanism to the Regional Blood Center to address the following fundamental inventory management questions related to these measures: (i) what are the minimum achievable outdate and shortage targets that can be set for the region, (ii) what is the distribution policy to achieve those targets, and (iii) what should the level of regional supply be in order to achieve alternative targets. The system is based on a mathematical programming model whose primary objective is to optimize the allocation of the regional blood resources while observing policy constraints. It is characterized by (i) a centralized management of blood, rather man management by individual hospitals, (ii) prescheduled deliveries, and (iii) a distribution system according to which blood is "rotated" among the hospitals. PBDS has been implemented in the 38 hospital region of Long Island, New York, where it has been operational for 3 years. It has established a routine management, and has drastically reduced the outdating and shortage incidents in the region. By now, it is completely computerized, and plans are being made for its introduction to other regions in the U. S. and abroad.