Abstract:
|
Group testing, through the use of pooling, has been widely implemented as a more efficient means to screen individuals for infectious diseases. Various testing strategies, such as hierarchical and square array-based testing algorithms, have been proposed. In this talk, I will present the comparison of the operating characteristics, including testing efficiency and classification accuracy, of these algorithms for the purpose of case identification, and also provide a general regression estimation framework for modeling all type group testing data. The differences between our approach and the previous ones are the assumptions regarding testing error rates. We relax previously made assumptions by acknowledging the mechanistic structure of the diagnostic assays. By doing this, we are able to account for the dilution effect; i.e., truly positive specimens could be diluted when they are pooled together with many truly negative ones, and thus cannot be detected. This methodology is illustrated by comparing different testing algorithms via the HIV, HBV and HCV data collected form a study involving Irish prisoners.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.