A probabilistic method for the estimation of residual risk in donated blood

作者：Bish Ebru K^*; Ragavan Prasanna K; Bish Douglas R; Slonim Anthony D; Stramer Susan L

来源：Biostatistics, 2014, 15(4): 620-635.

DOI：10.1093/biostatistics/kxu017

• Blood donation
• Incidence/window period model
• Optimization
• Risk estimation

The residual risk (RR) of transfusion-transmitted infections, including the human immunodeficiency virus and hepatitis B and C viruses, is typically estimated by the incidence/window period model, which relies on the following restrictive assumptions: Each screening test, with probability 1, (1) detects an infected unit outside of the test%26apos;s window period; (2) fails to detect an infected unit within the window period; and (3) correctly identifies an infection-free unit. These assumptions need not hold in practice due to random or systemic errors and individual variations in the window period. We develop a probability model that accurately estimates the RR by relaxing these assumptions, and quantify their impact using a published cost-effectiveness study and also within an optimization model. These assumptions lead to inaccurate estimates in cost-effectiveness studies and to sub-optimal solutions in the optimization model. The testing solution generated by the optimization model translates into fewer expected infections without an increase in the testing cost.

• 出版日期2014-10
• 单位美国弗吉尼亚理工大学(Virginia Tech)