Multiple imputation for left-censored biomarker data based on Gibbs sampling method

Minjae Lee, Lan Kong, Lisa Weissfeld

Research output: Contribution to journalArticle

10 Citations (Scopus)

Abstract

Biomarkers, increasingly used in biomedical studies for the diagnosis and prognosis of acute and chronic diseases, provide insight into the effectiveness of treatments and potential pathways that can be used to guide future treatment targets. The measurement of these markers is often limited by the sensitivity of the given assay, resulting in data that are censored either at the lower or at the upper limit of detection. For the Genetic and Inflammatory Markers of Sepsis (GenIMS) study, many different biomarkers were measured to examine the effect of different pathways on the development of sepsis. In this study, the left-censoring of several important inflammatory markers has led to the need for statistical methods that can incorporate this censoring into any analysis of the biomarker data. This paper focuses on the development of multiple imputation methods for the inclusion of multiple left-censored biomarkers in a logistic regression analysis. We assume a multivariate normal distribution to account for the correlations between biomarkers and use the Gibbs sampler for the estimation of the distributional parameters and the imputation of the censored markers. We evaluate and compare the proposed methods with some simple imputation methods through simulation. We use a data set of inflammatory and coagulation markers from the GenIMS study for illustration.

Original languageEnglish (US)
Pages (from-to)1838-1848
Number of pages11
JournalStatistics in Medicine
Volume31
Issue number17
DOIs
StatePublished - Jul 30 2012

Fingerprint

Multiple Imputation
Gibbs Sampling
Biomarkers
Sampling Methods
Sepsis
Imputation
Genetic Markers
Pathway
Left Censoring
Chronic Disease
Multivariate Normal Distribution
Gibbs Sampler
Coagulation
Normal Distribution
Prognosis
Acute Disease
Censoring
Logistic Regression
Regression Analysis
Acute

All Science Journal Classification (ASJC) codes

  • Epidemiology
  • Statistics and Probability

Cite this

Lee, Minjae ; Kong, Lan ; Weissfeld, Lisa. / Multiple imputation for left-censored biomarker data based on Gibbs sampling method. In: Statistics in Medicine. 2012 ; Vol. 31, No. 17. pp. 1838-1848.
@article{76128149e67e4a838497c5354f4e28bc,
title = "Multiple imputation for left-censored biomarker data based on Gibbs sampling method",
abstract = "Biomarkers, increasingly used in biomedical studies for the diagnosis and prognosis of acute and chronic diseases, provide insight into the effectiveness of treatments and potential pathways that can be used to guide future treatment targets. The measurement of these markers is often limited by the sensitivity of the given assay, resulting in data that are censored either at the lower or at the upper limit of detection. For the Genetic and Inflammatory Markers of Sepsis (GenIMS) study, many different biomarkers were measured to examine the effect of different pathways on the development of sepsis. In this study, the left-censoring of several important inflammatory markers has led to the need for statistical methods that can incorporate this censoring into any analysis of the biomarker data. This paper focuses on the development of multiple imputation methods for the inclusion of multiple left-censored biomarkers in a logistic regression analysis. We assume a multivariate normal distribution to account for the correlations between biomarkers and use the Gibbs sampler for the estimation of the distributional parameters and the imputation of the censored markers. We evaluate and compare the proposed methods with some simple imputation methods through simulation. We use a data set of inflammatory and coagulation markers from the GenIMS study for illustration.",
author = "Minjae Lee and Lan Kong and Lisa Weissfeld",
year = "2012",
month = "7",
day = "30",
doi = "10.1002/sim.4503",
language = "English (US)",
volume = "31",
pages = "1838--1848",
journal = "Statistics in Medicine",
issn = "0277-6715",
publisher = "John Wiley and Sons Ltd",
number = "17",

}

Multiple imputation for left-censored biomarker data based on Gibbs sampling method. / Lee, Minjae; Kong, Lan; Weissfeld, Lisa.

In: Statistics in Medicine, Vol. 31, No. 17, 30.07.2012, p. 1838-1848.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Multiple imputation for left-censored biomarker data based on Gibbs sampling method

AU - Lee, Minjae

AU - Kong, Lan

AU - Weissfeld, Lisa

PY - 2012/7/30

Y1 - 2012/7/30

N2 - Biomarkers, increasingly used in biomedical studies for the diagnosis and prognosis of acute and chronic diseases, provide insight into the effectiveness of treatments and potential pathways that can be used to guide future treatment targets. The measurement of these markers is often limited by the sensitivity of the given assay, resulting in data that are censored either at the lower or at the upper limit of detection. For the Genetic and Inflammatory Markers of Sepsis (GenIMS) study, many different biomarkers were measured to examine the effect of different pathways on the development of sepsis. In this study, the left-censoring of several important inflammatory markers has led to the need for statistical methods that can incorporate this censoring into any analysis of the biomarker data. This paper focuses on the development of multiple imputation methods for the inclusion of multiple left-censored biomarkers in a logistic regression analysis. We assume a multivariate normal distribution to account for the correlations between biomarkers and use the Gibbs sampler for the estimation of the distributional parameters and the imputation of the censored markers. We evaluate and compare the proposed methods with some simple imputation methods through simulation. We use a data set of inflammatory and coagulation markers from the GenIMS study for illustration.

AB - Biomarkers, increasingly used in biomedical studies for the diagnosis and prognosis of acute and chronic diseases, provide insight into the effectiveness of treatments and potential pathways that can be used to guide future treatment targets. The measurement of these markers is often limited by the sensitivity of the given assay, resulting in data that are censored either at the lower or at the upper limit of detection. For the Genetic and Inflammatory Markers of Sepsis (GenIMS) study, many different biomarkers were measured to examine the effect of different pathways on the development of sepsis. In this study, the left-censoring of several important inflammatory markers has led to the need for statistical methods that can incorporate this censoring into any analysis of the biomarker data. This paper focuses on the development of multiple imputation methods for the inclusion of multiple left-censored biomarkers in a logistic regression analysis. We assume a multivariate normal distribution to account for the correlations between biomarkers and use the Gibbs sampler for the estimation of the distributional parameters and the imputation of the censored markers. We evaluate and compare the proposed methods with some simple imputation methods through simulation. We use a data set of inflammatory and coagulation markers from the GenIMS study for illustration.

UR - http://www.scopus.com/inward/record.url?scp=84864000255&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84864000255&partnerID=8YFLogxK

U2 - 10.1002/sim.4503

DO - 10.1002/sim.4503

M3 - Article

C2 - 22359320

AN - SCOPUS:84864000255

VL - 31

SP - 1838

EP - 1848

JO - Statistics in Medicine

JF - Statistics in Medicine

SN - 0277-6715

IS - 17

ER -