Optimization of blood supply chain with shortened shelf lives and ABO compatibility

In clinical practice, red blood cells (RBCs) have a maximal shelf life (MSL) of 42 days. Recent studies suggest managing RBC inventory with a more restrictive MSL. To this end, a new simulation optimization (SO) framework is proposed for blood supply chain inventory management with ABO blood group compatibility. The inventory objective is to minimize the expected system outdate rate under a predetermined maximally allowable shortage level. The proposed SO framework is incorporated with a new metaheuristic optimization algorithm, TA–TS, to identify near-optimal inventory policies in reasonably acceptable computational time. The new SO framework is shown to offer a better tradeoff between solution accuracy and computational expense. The efficiency of the proposed framework is evaluated in detail for a single-hospital single-blood center supply chain system, in which the MSL of RBC units are shortened to 7, 14 and 21 days. A recently-developed replenishment policy based on old inventory ratio (the OIR policy developed by Duan and Liao (2013b)) is used to control the freshness of the entire inventory. Three different scenarios are investigated: (1) no ABO compatible substitution; (2) ABO compatible substitution at hospital only; (3) ABO compatible substitution at both hospital and blood center. Using the proposed SO framework, we are able to identify a near-optimal solution for each of these scenarios and quantify the potential savings offered by ABO compatible substitution. In the ABO compatible substitution scenarios, there is a clear trend of the increased use of group O blood. As the MSL decreases, group O blood are needed more frequently to substitute for other ABO/Rh(D) compatible blood types. Allowing ABO/Rh(D)-compatible blood substitution helps reduce the system-wide outdate at least by 16% even under the most restrictive MSL. For more relax MSL of 14 days and 21 days, the proposed framework is capable of keeping the highest system-wide outdate rate at as little as 2%.