Correcting for multivariate measurement error by regression calibration in meta-analyses of epidemiological studies

Amanda Lee, Fibrinogen Studies Collaboration

Research output: Contribution to journalArticle

44 Citations (Scopus)

Abstract

Within-person variability in measured values of multiple risk factors call bias their associations with disease. The multivariate regression calibration (RC) approach call correct for such measurement error and has been applied to Studies in which true Values or independent repeat measurements of the risk factors are observed oil a subsample. We extend the multivariate RC techniques to a meta-analysis framework where Multiple studies provide independent repeat measurements and information oil disease outcome. We consider the cases where sonic or all studies have repeat measurements, and compare study-specific, averaged and empirical Bayes estimates of RC parameters. Additionally, we allow for binary covariates (e.g. smoking status) and for uncertainty and time trends in the measurement error corrections. Our methods are illustrated using a subset of individual participant data from prospective long-term studies ill the Fibrinogen Studies Collaboration to assess file relationship between usual levels of plasma fibrinogen and the risk of coronary heart disease, allowing for measurement error in plasma fibrinogen and several confounders. Copyright (C) 2009 John Wiley & Soils, Ltd.
Original languageEnglish
Pages (from-to)1067-1092
Number of pages26
JournalStatistics in Medicine
Volume28
Issue number7
Early online date16 Feb 2009
DOIs
Publication statusPublished - 30 Mar 2009

Fingerprint

Regression Calibration
Measurement Error
Fibrinogen
Calibration
Meta-Analysis
Multivariate Calibration
Epidemiologic Studies
Multivariate Regression
Risk Factors
Oils
Plasma
Coronary Heart Disease
Bayes Estimate
Empirical Bayes
Smoking
Error Correction
Uncertainty
Coronary Disease
Soil
Covariates

Keywords

  • measurement error
  • within-person variation
  • meta-analysis
  • individual participant data
  • regression calibration

Cite this

Correcting for multivariate measurement error by regression calibration in meta-analyses of epidemiological studies. / Lee, Amanda; Fibrinogen Studies Collaboration.

In: Statistics in Medicine, Vol. 28, No. 7, 30.03.2009, p. 1067-1092.

Research output: Contribution to journalArticle

@article{43844359b31f4020b4ba9137a8f6c2b2,
title = "Correcting for multivariate measurement error by regression calibration in meta-analyses of epidemiological studies",
abstract = "Within-person variability in measured values of multiple risk factors call bias their associations with disease. The multivariate regression calibration (RC) approach call correct for such measurement error and has been applied to Studies in which true Values or independent repeat measurements of the risk factors are observed oil a subsample. We extend the multivariate RC techniques to a meta-analysis framework where Multiple studies provide independent repeat measurements and information oil disease outcome. We consider the cases where sonic or all studies have repeat measurements, and compare study-specific, averaged and empirical Bayes estimates of RC parameters. Additionally, we allow for binary covariates (e.g. smoking status) and for uncertainty and time trends in the measurement error corrections. Our methods are illustrated using a subset of individual participant data from prospective long-term studies ill the Fibrinogen Studies Collaboration to assess file relationship between usual levels of plasma fibrinogen and the risk of coronary heart disease, allowing for measurement error in plasma fibrinogen and several confounders. Copyright (C) 2009 John Wiley & Soils, Ltd.",
keywords = "measurement error, within-person variation, meta-analysis, individual participant data, regression calibration",
author = "Amanda Lee and {Fibrinogen Studies Collaboration}",
year = "2009",
month = "3",
day = "30",
doi = "10.1002/sim.3530",
language = "English",
volume = "28",
pages = "1067--1092",
journal = "Statistics in Medicine",
issn = "0277-6715",
publisher = "John Wiley and Sons Ltd",
number = "7",

}

TY - JOUR

T1 - Correcting for multivariate measurement error by regression calibration in meta-analyses of epidemiological studies

AU - Lee, Amanda

AU - Fibrinogen Studies Collaboration

PY - 2009/3/30

Y1 - 2009/3/30

N2 - Within-person variability in measured values of multiple risk factors call bias their associations with disease. The multivariate regression calibration (RC) approach call correct for such measurement error and has been applied to Studies in which true Values or independent repeat measurements of the risk factors are observed oil a subsample. We extend the multivariate RC techniques to a meta-analysis framework where Multiple studies provide independent repeat measurements and information oil disease outcome. We consider the cases where sonic or all studies have repeat measurements, and compare study-specific, averaged and empirical Bayes estimates of RC parameters. Additionally, we allow for binary covariates (e.g. smoking status) and for uncertainty and time trends in the measurement error corrections. Our methods are illustrated using a subset of individual participant data from prospective long-term studies ill the Fibrinogen Studies Collaboration to assess file relationship between usual levels of plasma fibrinogen and the risk of coronary heart disease, allowing for measurement error in plasma fibrinogen and several confounders. Copyright (C) 2009 John Wiley & Soils, Ltd.

AB - Within-person variability in measured values of multiple risk factors call bias their associations with disease. The multivariate regression calibration (RC) approach call correct for such measurement error and has been applied to Studies in which true Values or independent repeat measurements of the risk factors are observed oil a subsample. We extend the multivariate RC techniques to a meta-analysis framework where Multiple studies provide independent repeat measurements and information oil disease outcome. We consider the cases where sonic or all studies have repeat measurements, and compare study-specific, averaged and empirical Bayes estimates of RC parameters. Additionally, we allow for binary covariates (e.g. smoking status) and for uncertainty and time trends in the measurement error corrections. Our methods are illustrated using a subset of individual participant data from prospective long-term studies ill the Fibrinogen Studies Collaboration to assess file relationship between usual levels of plasma fibrinogen and the risk of coronary heart disease, allowing for measurement error in plasma fibrinogen and several confounders. Copyright (C) 2009 John Wiley & Soils, Ltd.

KW - measurement error

KW - within-person variation

KW - meta-analysis

KW - individual participant data

KW - regression calibration

U2 - 10.1002/sim.3530

DO - 10.1002/sim.3530

M3 - Article

VL - 28

SP - 1067

EP - 1092

JO - Statistics in Medicine

JF - Statistics in Medicine

SN - 0277-6715

IS - 7

ER -