Correlation Coefficient Inference for Left-Censored Biomarker Data with Known Detection Limits

Hdl Handle:
http://hdl.handle.net/10675.2/317606
Title:
Correlation Coefficient Inference for Left-Censored Biomarker Data with Known Detection Limits
Authors:
McCracken, Courtney Elizabeth
Abstract:
Researchers are often interested in the relationship between biological concentrations obtained using two different assays, both of which may be biomarkers. Despite the continuing advances in biotechnology, the value of a particular biomarker may fall below some known limit of detection (LOD). Data values such as these are referred to as non-detects (NDs) and can be treated as left-censored observations. When attempting to measure the association between two concentrations, both of which are subject to NDs, serious complications can arise in the data analysis. Simple substitution, random imputation, and maximum likelihood estimation methods are just a few of the methods that have been proposed for handling NDs when estimating the correlation between two variables, both of which are subject to left-censoring. Unfortunately, many of the popular methods require that the data follow a bivariate normal distribution or that only a small percentage of the data for each variable are below the LOD. These assumptions are often violated with biomarker data. In this paper, we evaluate the performance of several methods, including Spearman’s rho, when the data do not follow a bivariate normal distribution and when there are moderate to large censoring proportions in one or both of the variables. We evaluate the performance of seven methods for estimating the correlation, ρ, between two left-censored variables using bias, median absolute deviation, 95% confidence interval width, and coverage probability under assumptions of various sample sizes, correlations, and censoring proportions. We show that using substitution and imputation methods yields biased estimates of ρ and less than nominal coverage probability under most of the simulation parameters we examined. We recommend the maximum likelihood method for general use even when the data significantly depart from bivariate normality.
Affiliation:
Department of Biostatistics and Epidemiology
Issue Date:
May-2013
URI:
http://hdl.handle.net/10675.2/317606
Additional Links:
http://ezproxy.gru.edu/login?url=http://search.proquest.com/docview/1372291839?accountid=12365
Type:
Dissertation
Language:
en_US
Appears in Collections:
Theses and Dissertations; Department of Biostatistics and Epidemiology Theses and Dissertations

Full metadata record

DC FieldValue Language
dc.contributor.authorMcCracken, Courtney Elizabethen
dc.date.accessioned2014-05-28T17:53:43Z-
dc.date.available2014-05-28T17:53:43Z-
dc.date.issued2013-05-
dc.identifier.urihttp://hdl.handle.net/10675.2/317606-
dc.description.abstractResearchers are often interested in the relationship between biological concentrations obtained using two different assays, both of which may be biomarkers. Despite the continuing advances in biotechnology, the value of a particular biomarker may fall below some known limit of detection (LOD). Data values such as these are referred to as non-detects (NDs) and can be treated as left-censored observations. When attempting to measure the association between two concentrations, both of which are subject to NDs, serious complications can arise in the data analysis. Simple substitution, random imputation, and maximum likelihood estimation methods are just a few of the methods that have been proposed for handling NDs when estimating the correlation between two variables, both of which are subject to left-censoring. Unfortunately, many of the popular methods require that the data follow a bivariate normal distribution or that only a small percentage of the data for each variable are below the LOD. These assumptions are often violated with biomarker data. In this paper, we evaluate the performance of several methods, including Spearman’s rho, when the data do not follow a bivariate normal distribution and when there are moderate to large censoring proportions in one or both of the variables. We evaluate the performance of seven methods for estimating the correlation, ρ, between two left-censored variables using bias, median absolute deviation, 95% confidence interval width, and coverage probability under assumptions of various sample sizes, correlations, and censoring proportions. We show that using substitution and imputation methods yields biased estimates of ρ and less than nominal coverage probability under most of the simulation parameters we examined. We recommend the maximum likelihood method for general use even when the data significantly depart from bivariate normality.en
dc.language.isoen_USen
dc.relation.urlhttp://ezproxy.gru.edu/login?url=http://search.proquest.com/docview/1372291839?accountid=12365en
dc.rightsCopyright protected. Unauthorized reproduction or use beyond the exceptions granted by the Fair Use clause of U.S. Copyright law may violate federal law.-
dc.subjectCorrelationen
dc.subjectLeft-Censoreden
dc.subjectBiomarkeren
dc.subjectNon-Bivariate Normal Dataen
dc.titleCorrelation Coefficient Inference for Left-Censored Biomarker Data with Known Detection Limitsen
dc.typeDissertationen
dc.contributor.departmentDepartment of Biostatistics and Epidemiologyen
dc.description.advisorLooney, Stephen W.-
dc.description.committeeWaller, Jennifer; Lan, Ling; Lyles, Robert; Hess, David.-
dc.description.degreeDoctor of Philosophy (Ph.D.)-
All Items in Scholarly Commons are protected by copyright, with all rights reserved, unless otherwise indicated.