BACKGROUND: Preoperative anemia is associated with worse clinical postoperative outcomes and a higher risk of receiving red blood cell (RBC) transfusions. It is challenging to disentangle the effect of preoperative anemia from the effect of receiving RBC transfusions on postoperative clinical outcomes such as length of hospital stay (LOS). When analyzing the association of preoperative anemia on LOS, it is important to be able to analyze RBC transfusions as a mediator in this relationship. In this paper, the background and application of mediation analysis is outlined as a statistical methodology in transfusion medicine research. STUDY DESIGN AND METHODS: To explain the methodology of mediation analysis, a database from a previously reported clinical study was used (So-Osman C. et al. 2014) with anemia as the exposure variable and LOS as the primary outcome. Both the product-of-coefficients method and the change-in-coefficients method are used for mediation analysis, and linear regression models were used. RESULTS: In the example of a simplified analysis, two-thirds of the effect could be attributed to mediation. This result was obtained by both the product-of-coefficients method and the change-in-coefficients method. DISCUSSION: Mediation is assessed in a similar way as confounding, but the interpretation of the results is totally different. It is, therefore, of critical importance to distinguish between potential mediators and potential confounders in transfusion research. Since the calculation reported in the results is merely used as an example to show the methodology, e.g. ignoring confounding, the result should not be interpreted as scientific research data.