On the nondifferential misclassification of a binary confounder

Elizabeth Leigh Ogburn, Tyler J. VanderWeele

Research output: Contribution to journalArticle

Abstract

Consider a study with binary exposure, outcome, and confounder, where the confounder is nondifferentially misclassified. Epidemiologists have long accepted the unproven but oft-cited result that, if the confounder is binary, then odds ratios, risk ratios, and risk differences that control for the mismeasured confounder will lie between the crude and the true measures. In this paper, we provide an analytic proof of the result in the absence of a qualitative interaction between treatment and confounder, and we demonstrate via counterexample that the result need not hold when there is such a qualitative interaction. We also present an analytic proof of the result for the effect of treatment among the treated and describe extensions to measures conditional on or standardized over other covariates.

Original languageEnglish (US)
Pages (from-to)433-439
Number of pages7
JournalEpidemiology
Volume23
Issue number3
DOIs
StatePublished - May 2012
Externally publishedYes

Fingerprint

Odds Ratio
Therapeutics
Epidemiologists

ASJC Scopus subject areas

  • Epidemiology

Cite this

On the nondifferential misclassification of a binary confounder. / Ogburn, Elizabeth Leigh; VanderWeele, Tyler J.

In: Epidemiology, Vol. 23, No. 3, 05.2012, p. 433-439.

Research output: Contribution to journalArticle

Ogburn, Elizabeth Leigh ; VanderWeele, Tyler J. / On the nondifferential misclassification of a binary confounder. In: Epidemiology. 2012 ; Vol. 23, No. 3. pp. 433-439.
@article{a20d67609bdc43f892d6c526be4a0880,
title = "On the nondifferential misclassification of a binary confounder",
abstract = "Consider a study with binary exposure, outcome, and confounder, where the confounder is nondifferentially misclassified. Epidemiologists have long accepted the unproven but oft-cited result that, if the confounder is binary, then odds ratios, risk ratios, and risk differences that control for the mismeasured confounder will lie between the crude and the true measures. In this paper, we provide an analytic proof of the result in the absence of a qualitative interaction between treatment and confounder, and we demonstrate via counterexample that the result need not hold when there is such a qualitative interaction. We also present an analytic proof of the result for the effect of treatment among the treated and describe extensions to measures conditional on or standardized over other covariates.",
author = "Ogburn, {Elizabeth Leigh} and VanderWeele, {Tyler J.}",
year = "2012",
month = "5",
doi = "10.1097/EDE.0b013e31824d1f63",
language = "English (US)",
volume = "23",
pages = "433--439",
journal = "Epidemiology",
issn = "1044-3983",
publisher = "Lippincott Williams and Wilkins",
number = "3",

}

TY - JOUR

T1 - On the nondifferential misclassification of a binary confounder

AU - Ogburn, Elizabeth Leigh

AU - VanderWeele, Tyler J.

PY - 2012/5

Y1 - 2012/5

N2 - Consider a study with binary exposure, outcome, and confounder, where the confounder is nondifferentially misclassified. Epidemiologists have long accepted the unproven but oft-cited result that, if the confounder is binary, then odds ratios, risk ratios, and risk differences that control for the mismeasured confounder will lie between the crude and the true measures. In this paper, we provide an analytic proof of the result in the absence of a qualitative interaction between treatment and confounder, and we demonstrate via counterexample that the result need not hold when there is such a qualitative interaction. We also present an analytic proof of the result for the effect of treatment among the treated and describe extensions to measures conditional on or standardized over other covariates.

AB - Consider a study with binary exposure, outcome, and confounder, where the confounder is nondifferentially misclassified. Epidemiologists have long accepted the unproven but oft-cited result that, if the confounder is binary, then odds ratios, risk ratios, and risk differences that control for the mismeasured confounder will lie between the crude and the true measures. In this paper, we provide an analytic proof of the result in the absence of a qualitative interaction between treatment and confounder, and we demonstrate via counterexample that the result need not hold when there is such a qualitative interaction. We also present an analytic proof of the result for the effect of treatment among the treated and describe extensions to measures conditional on or standardized over other covariates.

UR - http://www.scopus.com/inward/record.url?scp=82055193052&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=82055193052&partnerID=8YFLogxK

U2 - 10.1097/EDE.0b013e31824d1f63

DO - 10.1097/EDE.0b013e31824d1f63

M3 - Article

C2 - 22450692

AN - SCOPUS:82055193052

VL - 23

SP - 433

EP - 439

JO - Epidemiology

JF - Epidemiology

SN - 1044-3983

IS - 3

ER -