Using Sensitivity Analyses for Unobserved Confounding to Address Covariate Measurement Error in Propensity Score Methods

Kara Rudolph, Elizabeth A. Stuart

Research output: Contribution to journalArticle

3 Citations (Scopus)

Abstract

Propensity score methods are a popular tool with which to control for confounding in observational data, but their bias-reduction properties - as well as internal validity, generally - are threatened by covariate measurement error. There are few easy-to-implement methods of correcting for such bias. In this paper, we describe and demonstrate how existing sensitivity analyses for unobserved confounding - propensity score calibration, VanderWeele and Arah's bias formulas, and Rosenbaum's sensitivity analysis - can be adapted to address this problem. In a simulation study, we examine the extent to which these sensitivity analyses can correct for several measurement error structures: classical, systematic differential, and heteroscedastic covariate measurement error. We then apply these approaches to address covariate measurement error in estimating the association between depression and weight gain in a cohort of adults in Baltimore, Maryland. We recommend the use of VanderWeele and Arah's bias formulas and propensity score calibration (assuming it is adapted appropriately for the measurement error structure), as both approaches perform well for a variety of propensity score estimators and measurement error structures.

Original languageEnglish (US)
Pages (from-to)604-613
Number of pages10
JournalAmerican Journal of Epidemiology
Volume187
Issue number3
DOIs
StatePublished - Mar 1 2018
Externally publishedYes

Fingerprint

Propensity Score
Calibration
Baltimore
Weight Gain

Keywords

  • confounding factors (epidemiology)
  • measurement error
  • propensity score
  • unobserved confounding

ASJC Scopus subject areas

  • Epidemiology

Cite this

Using Sensitivity Analyses for Unobserved Confounding to Address Covariate Measurement Error in Propensity Score Methods. / Rudolph, Kara; Stuart, Elizabeth A.

In: American Journal of Epidemiology, Vol. 187, No. 3, 01.03.2018, p. 604-613.

Research output: Contribution to journalArticle

@article{7d2e2f4b4b104469b2ad5ae7af375298,
title = "Using Sensitivity Analyses for Unobserved Confounding to Address Covariate Measurement Error in Propensity Score Methods",
abstract = "Propensity score methods are a popular tool with which to control for confounding in observational data, but their bias-reduction properties - as well as internal validity, generally - are threatened by covariate measurement error. There are few easy-to-implement methods of correcting for such bias. In this paper, we describe and demonstrate how existing sensitivity analyses for unobserved confounding - propensity score calibration, VanderWeele and Arah's bias formulas, and Rosenbaum's sensitivity analysis - can be adapted to address this problem. In a simulation study, we examine the extent to which these sensitivity analyses can correct for several measurement error structures: classical, systematic differential, and heteroscedastic covariate measurement error. We then apply these approaches to address covariate measurement error in estimating the association between depression and weight gain in a cohort of adults in Baltimore, Maryland. We recommend the use of VanderWeele and Arah's bias formulas and propensity score calibration (assuming it is adapted appropriately for the measurement error structure), as both approaches perform well for a variety of propensity score estimators and measurement error structures.",
keywords = "confounding factors (epidemiology), measurement error, propensity score, unobserved confounding",
author = "Kara Rudolph and Stuart, {Elizabeth A.}",
year = "2018",
month = "3",
day = "1",
doi = "10.1093/aje/kwx248",
language = "English (US)",
volume = "187",
pages = "604--613",
journal = "American Journal of Epidemiology",
issn = "0002-9262",
publisher = "Oxford University Press",
number = "3",

}

TY - JOUR

T1 - Using Sensitivity Analyses for Unobserved Confounding to Address Covariate Measurement Error in Propensity Score Methods

AU - Rudolph, Kara

AU - Stuart, Elizabeth A.

PY - 2018/3/1

Y1 - 2018/3/1

N2 - Propensity score methods are a popular tool with which to control for confounding in observational data, but their bias-reduction properties - as well as internal validity, generally - are threatened by covariate measurement error. There are few easy-to-implement methods of correcting for such bias. In this paper, we describe and demonstrate how existing sensitivity analyses for unobserved confounding - propensity score calibration, VanderWeele and Arah's bias formulas, and Rosenbaum's sensitivity analysis - can be adapted to address this problem. In a simulation study, we examine the extent to which these sensitivity analyses can correct for several measurement error structures: classical, systematic differential, and heteroscedastic covariate measurement error. We then apply these approaches to address covariate measurement error in estimating the association between depression and weight gain in a cohort of adults in Baltimore, Maryland. We recommend the use of VanderWeele and Arah's bias formulas and propensity score calibration (assuming it is adapted appropriately for the measurement error structure), as both approaches perform well for a variety of propensity score estimators and measurement error structures.

AB - Propensity score methods are a popular tool with which to control for confounding in observational data, but their bias-reduction properties - as well as internal validity, generally - are threatened by covariate measurement error. There are few easy-to-implement methods of correcting for such bias. In this paper, we describe and demonstrate how existing sensitivity analyses for unobserved confounding - propensity score calibration, VanderWeele and Arah's bias formulas, and Rosenbaum's sensitivity analysis - can be adapted to address this problem. In a simulation study, we examine the extent to which these sensitivity analyses can correct for several measurement error structures: classical, systematic differential, and heteroscedastic covariate measurement error. We then apply these approaches to address covariate measurement error in estimating the association between depression and weight gain in a cohort of adults in Baltimore, Maryland. We recommend the use of VanderWeele and Arah's bias formulas and propensity score calibration (assuming it is adapted appropriately for the measurement error structure), as both approaches perform well for a variety of propensity score estimators and measurement error structures.

KW - confounding factors (epidemiology)

KW - measurement error

KW - propensity score

KW - unobserved confounding

UR - http://www.scopus.com/inward/record.url?scp=85042916433&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85042916433&partnerID=8YFLogxK

U2 - 10.1093/aje/kwx248

DO - 10.1093/aje/kwx248

M3 - Article

C2 - 28992211

AN - SCOPUS:85042916433

VL - 187

SP - 604

EP - 613

JO - American Journal of Epidemiology

JF - American Journal of Epidemiology

SN - 0002-9262

IS - 3

ER -