CMStatistics 2018: Start Registration
View Submission - CMStatistics
Title: Classification of competing risk outcomes using transition biomarkers Authors:  Feng-Chang Lin - University of North Carolina at Chapel Hill (United States) [presenting]
Quefeng Li - University of North Carolina - Chapel Hill (United States)
Jessica Lin - University of North Carolina at Chapel Hill (United States)
Abstract: In the Plasmodium vivax malaria infection, relapse from previous infections and reinfection from a new mosquito bite can be considered as competing risks. Classification of the recurrent infection to either relapse or new infection becomes critical when the researcher tries to detect genetic signatures of relapse that is key to evaluating anti-vivax interventions. While one can use baseline information to build up a nave classifier for the recurrent infection, little has been suggested to use transition biomarkers that appear in the recurrent infection for classification. We will introduce a newly developed classifier that uses the transition biomarker information to enhance the accuracy of classification. The approach was examined in extensive simulation experiments when the underlying outcome is known, with superior sensitivity and specificity. A real data from 78 Cambodian Plasmodium vivax malaria patients was analyzed to demonstrate the practical use of the proposed method.