This paper examines the use of simulations to develop decision procedures for the management of individual hospital blood banks and central blood bank systems. Due to the perishability of blood products and the stochastic nature of both the demand and supply processes, such multi-product, multiechelon systems can only be analyzed effectively by simulation. The hospital and central bank model structures are reviewed, as are the input, output, and validation analysis procedures. A number of issues relevant to the application of simulation to health care systems analysis are raised by the development and application of these models. Of particular interest is the integration of the simulations into a recursive statistics-optimization-simulation approach to systems analysis.