American Statistical Association New York City Metropolitan Area Chapter Memorial Sloan-Kettering Cancer Center Biostatistics Seminar

In this talk, I discuss some statistical challenges in analyzing neurodegenerative disease research data and the solutions. Developing biomarkers for early detection of neurodegenerative diseases (e.g. Alzheimer’s disease) has drawn great attentions recently. For example, cerebrospinal fluid (CSF) tau is a useful biomarker in detecting Alzheimer's disease (AD). However, the measurement error of these biomarkers cannot be ignored sometimes in the analysis. When an internal reliability sample is available for CSF tau, we have shown that averaging replicates and ignoring measurement error can sometimes lead to serious biases of sensitivity and specificity estimates as well as estimates of hazard ratios. We present bias-correction approaches to remove the biases of accuracy and hazard ratio estimates introduced by measurement error based on an internal reliability sample. Asymptotic distributions were obtained for the proposed estimators. Extensive simulations were conducted to evaluate the proposed approaches. All methods are illustrated using the University of Pennsylvania AD biomarker study or the Framingham study.