Construction of a decision model for donor testing in cases of suspected antibody-mediated transfusion-related-acute-lung-injury.

Abstract

Objective: To propose a rational basis for donor testing in cases of suspected antibody-mediated transfusion-related lung injury (AMT).

Background: Anti-leukocyte antibodies in donated blood are established causes of transfusion-related lung injury (TRALI). However, the question of whether to test donors for antibodies is not identical to whether the case meets definition criteria for TRALI. There is a balance between the potential benefits of testing and the costs of donor deferral and investigation. We propose that a decision-making process based on optimising the balance between risk and benefit requires a subjective choice of the relative value of different outcomes of testing.

Methods: We have developed a formal decision model to illustrate how these choices affect testing decisions.

Results: Using a Bayesian probability model, we show that the diagnostic benefit and TRALI prevention benefit of testing donors have a complex interrelationship with the number of implicated donors and clinical suspicion of antibody-mediated TRALI (AMT) and that rational testing choices vary according to value assigned to outcomes.

Conclusions: The challenges to the use of a formal decision model for clinical testing are discussed and conclude that a formal model is a useful consensus-building tool for improving consistency and openness in decision making.