All PhD Students are encouraged to attend!
Wednesday, October 5 2016
Building 2, Room 426 – Biostats Conference Room
12:30-1:30 PM
Sy Han (Steven) Chiou
Research
Department of Biostatistics
Harvard T.H. Chan School of Public Health
Permutation tests and survival estimation under general dependent truncation
Truncation is a mechanism that permits observation of selected subjects from a source population; subjects are included only if their event times are contained within subject-specific intervals. Standard survival analysis methods for estimation of the distribution of the event time require quasi-independence of failure and truncation. A series of nonparametric tests are developed for testing quasi-independence that are powerful against general, nonmonotone dependencies. Once a dependence is detected, a computationally efficient transformation model is proposed for survival estimation. These methods are generalized to handle independent right censoring. A real data from the National Alzheimer Coordinating Centers autopsy cohort study illustrate the performance of the proposed method.
