EPI/BIO Seminar Series: Ying Zhang, PhD

"Classification with Multiple Tests without a Gold Standard"

This work investigates a general problem of combining multiple continuous tests without any gold standard or a reference test motivated by determination of GBV-C E2 antibody in HIV patients for a HIV/AIDS study conducted in the University of Iowa. The existing methods for combining
multiple tests are thoroughly studied and the optimal composite classification rules corresponding to the methods are developed to incorporate the situation that no gold standard for E2 antibody exists. The proposed methods are theoretically justified and numerically compared via extensive simulation studies. Finally, the methods are applied to the motivating example to determine the HIV patients who had
GBV-C E2 antibody that is hypothesized to prolong the survival from HIV infection.

Click for event flyer