Correlated biomarker measurement error: An important threat to inference in environmental epidemiology

A. Z. Pollack, N. J. Perkins, S. L. Mumford, A. Ye, E. F. Schisterman

Research output: Contribution to journalArticle

Abstract

Utilizing multiple biomarkers is increasingly common in epidemiology. However, the combined impact of correlated exposure measurement error, unmeasured confounding, interaction, and limits of detection (LODs) on inference for multiple biomarkers is unknown. We conducted data-driven simulations evaluating bias from correlated measurement error with varying reliability coefficients (R), odds ratios (ORs), levels of correlation between exposures and error, LODs, and interactions. Blood cadmium and lead levels in relation to anovulation served as the motivating example, based on findings from the BioCycle Study (2005-2007). For most scenarios, main-effect estimates for cadmium and lead with increasing levels of positively correlated measurement error created increasing downward or upward bias for OR > 1.00 and OR <1.00, respectively, that was also a function of effect size. Some scenarios showed bias for cadmium away from the null. Results subject to LODs were similar. Bias for main and interaction effects ranged from -130% to 36% and from -144% to 84%, respectively. A closed-form continuous outcome case solution provides a useful tool for estimating the bias in logistic regression. Investigators should consider how measurement error and LODs may bias findings when examining biomarkers measured in the same medium, prepared with the same process, or analyzed using the same method.

Original languageEnglish (US)
Pages (from-to)84-92
Number of pages9
JournalAmerican Journal of Epidemiology
Volume177
Issue number1
DOIs
StatePublished - Jan 2013
Externally publishedYes

Fingerprint

Limit of Detection
Epidemiology
Biomarkers
Cadmium
Odds Ratio
Anovulation
Logistic Models
Research Personnel

Keywords

  • biomarkers
  • cadmium
  • environmental epidemiology
  • lead
  • measurement error; reliability

ASJC Scopus subject areas

  • Epidemiology

Cite this

Correlated biomarker measurement error : An important threat to inference in environmental epidemiology. / Pollack, A. Z.; Perkins, N. J.; Mumford, S. L.; Ye, A.; Schisterman, E. F.

In: American Journal of Epidemiology, Vol. 177, No. 1, 01.2013, p. 84-92.

Research output: Contribution to journalArticle

Pollack, A. Z. ; Perkins, N. J. ; Mumford, S. L. ; Ye, A. ; Schisterman, E. F. / Correlated biomarker measurement error : An important threat to inference in environmental epidemiology. In: American Journal of Epidemiology. 2013 ; Vol. 177, No. 1. pp. 84-92.
@article{e959b7804df949c5a790d125cc8e9cbb,
title = "Correlated biomarker measurement error: An important threat to inference in environmental epidemiology",
abstract = "Utilizing multiple biomarkers is increasingly common in epidemiology. However, the combined impact of correlated exposure measurement error, unmeasured confounding, interaction, and limits of detection (LODs) on inference for multiple biomarkers is unknown. We conducted data-driven simulations evaluating bias from correlated measurement error with varying reliability coefficients (R), odds ratios (ORs), levels of correlation between exposures and error, LODs, and interactions. Blood cadmium and lead levels in relation to anovulation served as the motivating example, based on findings from the BioCycle Study (2005-2007). For most scenarios, main-effect estimates for cadmium and lead with increasing levels of positively correlated measurement error created increasing downward or upward bias for OR > 1.00 and OR <1.00, respectively, that was also a function of effect size. Some scenarios showed bias for cadmium away from the null. Results subject to LODs were similar. Bias for main and interaction effects ranged from -130{\%} to 36{\%} and from -144{\%} to 84{\%}, respectively. A closed-form continuous outcome case solution provides a useful tool for estimating the bias in logistic regression. Investigators should consider how measurement error and LODs may bias findings when examining biomarkers measured in the same medium, prepared with the same process, or analyzed using the same method.",
keywords = "biomarkers, cadmium, environmental epidemiology, lead, measurement error; reliability",
author = "Pollack, {A. Z.} and Perkins, {N. J.} and Mumford, {S. L.} and A. Ye and Schisterman, {E. F.}",
year = "2013",
month = "1",
doi = "10.1093/aje/kws209",
language = "English (US)",
volume = "177",
pages = "84--92",
journal = "American Journal of Epidemiology",
issn = "0002-9262",
publisher = "Oxford University Press",
number = "1",

}

TY - JOUR

T1 - Correlated biomarker measurement error

T2 - An important threat to inference in environmental epidemiology

AU - Pollack, A. Z.

AU - Perkins, N. J.

AU - Mumford, S. L.

AU - Ye, A.

AU - Schisterman, E. F.

PY - 2013/1

Y1 - 2013/1

N2 - Utilizing multiple biomarkers is increasingly common in epidemiology. However, the combined impact of correlated exposure measurement error, unmeasured confounding, interaction, and limits of detection (LODs) on inference for multiple biomarkers is unknown. We conducted data-driven simulations evaluating bias from correlated measurement error with varying reliability coefficients (R), odds ratios (ORs), levels of correlation between exposures and error, LODs, and interactions. Blood cadmium and lead levels in relation to anovulation served as the motivating example, based on findings from the BioCycle Study (2005-2007). For most scenarios, main-effect estimates for cadmium and lead with increasing levels of positively correlated measurement error created increasing downward or upward bias for OR > 1.00 and OR <1.00, respectively, that was also a function of effect size. Some scenarios showed bias for cadmium away from the null. Results subject to LODs were similar. Bias for main and interaction effects ranged from -130% to 36% and from -144% to 84%, respectively. A closed-form continuous outcome case solution provides a useful tool for estimating the bias in logistic regression. Investigators should consider how measurement error and LODs may bias findings when examining biomarkers measured in the same medium, prepared with the same process, or analyzed using the same method.

AB - Utilizing multiple biomarkers is increasingly common in epidemiology. However, the combined impact of correlated exposure measurement error, unmeasured confounding, interaction, and limits of detection (LODs) on inference for multiple biomarkers is unknown. We conducted data-driven simulations evaluating bias from correlated measurement error with varying reliability coefficients (R), odds ratios (ORs), levels of correlation between exposures and error, LODs, and interactions. Blood cadmium and lead levels in relation to anovulation served as the motivating example, based on findings from the BioCycle Study (2005-2007). For most scenarios, main-effect estimates for cadmium and lead with increasing levels of positively correlated measurement error created increasing downward or upward bias for OR > 1.00 and OR <1.00, respectively, that was also a function of effect size. Some scenarios showed bias for cadmium away from the null. Results subject to LODs were similar. Bias for main and interaction effects ranged from -130% to 36% and from -144% to 84%, respectively. A closed-form continuous outcome case solution provides a useful tool for estimating the bias in logistic regression. Investigators should consider how measurement error and LODs may bias findings when examining biomarkers measured in the same medium, prepared with the same process, or analyzed using the same method.

KW - biomarkers

KW - cadmium

KW - environmental epidemiology

KW - lead

KW - measurement error; reliability

UR - http://www.scopus.com/inward/record.url?scp=84871811890&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84871811890&partnerID=8YFLogxK

U2 - 10.1093/aje/kws209

DO - 10.1093/aje/kws209

M3 - Article

C2 - 23221725

AN - SCOPUS:84871811890

VL - 177

SP - 84

EP - 92

JO - American Journal of Epidemiology

JF - American Journal of Epidemiology

SN - 0002-9262

IS - 1

ER -