Evaluation of Cerebrospinal Fluid Assay Variability
 
   Lead Investigator:    Matthew White
   Institution      :    Boston Children's Hospital
   E-Mail           :    matthew.white@childrens.harvard.edu
   Proposal ID      :    564


   Publications:


1  White MT, Shaw LM, Xie SX
   Evaluation of Cerebrospinal Fluid Assay Variability in Alzheimer's Disease
   Journal Article  J Alzheimers Dis,51:463-470,2016,PMC4930357

   PUBLISHED





   Proposal Description:
   Studies of cerebrospinal fluid (CSF) biomarkers in Alzheimer's disease (AD)       have indicated that much of the variability observed in the biomarkers may be     due to measurement error. Biomarkers are often obtained with measurement error,   which may make the diagnostic biomarker appear less effective than it truly is.                                                                                                                                                                    We will utilize the longitudinal measurements of CSF biomarkers in the NACC       data to explicitly estimate the measurement error variability of the              biomarkers, which we define as the variability in the biomarker due to temporal                                                                                  variability within subjects.  We focus on the area under the receiver operating                                                                                  characteristic curve (AUC) as the measure of diagnostic effectiveness for         differentiating AD from normal cognition using CSF biomarkers.  We plan to        compare the AUC estimates obtained by a more standard, naive method (which uses                                                                                  a single observation per subject and ignores measurement error) and a             maximum-likelihood (ML) based method (which uses all replicates per subject and                                                                                adjusts for measurement error). If the methods produce nearly identical           estimates of the AUC, then we would conclude that using the naive method is       adequate for analysis of CSF biomarkers for this study.  If the methods do not    yield similar estimates, however, the ML-based estimation method is               recommended.