A generic modelling framework to evaluate network blood management policies: The Canadian Blood Services experience

Abstract

Abstract Canadian Blood Services produces and distributes approximately 850,000 units of red cells annually. These units are distributed through ten production and/or distribution sites. Each distribution site acts as a regional hub serving between 20 and 110 hospital customers. Distribution sites hold a target inventory that is based on an integer number of median days demand on hand. In this paper, we report on the development and use of a simulation based methodology to evaluate network inventory policies for regional blood distribution sites in Canada. A generic framework was developed to represent each of the ten different regional networks. The modelling approach was validated by comparing model results against data from two networks. Once validated, ten instances were developed. For each model instance, a set of experiments was conducted, from which response surfaces were created. Non-linear optimization methods were applied to identify optimal supplier/consumer inventory policies using the response surfaces. We conclude that a generic modelling framework can be useful for regional blood supply chains, but suggest that at least four instances are necessary to recoup the efforts of building a reusable model.