Doubly robust estimation and causal inference in longitudinal studies with dropout and truncation by death

Michelle Shardell, Gregory E. Hicks, Luigi Ferrucci

Research output: Contribution to journalArticle

Abstract

Motivated by aging research, we propose an estimator of the effect of a time-varying exposure on an outcome in longitudinal studies with dropout and truncation by death. We use an inverse-probability weighted (IPW) estimator to derive a doubly robust augmented inverse-probability weighted (AIPW) estimator. IPW estimation involves weights for the exposure mechanism, dropout, and mortality; AIPW estimation additionally involves estimating data-generating models via regression. We demonstrate that the estimators identify a causal contrast that is a function of principal strata effects under a set of assumptions. Simulations show that AIPW estimation is unbiased when weights or outcome regressions are correct, and that AIPW estimation is more efficient than IPW estimation when all models are correct. We apply the method to a study of vitamin D and gait speed among older adults.

Original languageEnglish (US)
Pages (from-to)155-168
Number of pages14
JournalBiostatistics
Volume16
Issue number1
DOIs
StatePublished - Jan 1 2015
Externally publishedYes

Fingerprint

Causal Inference
Drop out
Longitudinal Study
Robust Estimation
Truncation
Longitudinal Studies
Estimator
Weights and Measures
Gait
Causal inference
Longitudinal study
Robust estimation
Vitamin D
Mortality
Regression Model
Time-varying
Regression
Research
Demonstrate

Keywords

  • Causal inference
  • Longitudinal data analysis
  • Missing data
  • Observational studies

ASJC Scopus subject areas

  • Medicine(all)
  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this

Doubly robust estimation and causal inference in longitudinal studies with dropout and truncation by death. / Shardell, Michelle; Hicks, Gregory E.; Ferrucci, Luigi.

In: Biostatistics, Vol. 16, No. 1, 01.01.2015, p. 155-168.

Research output: Contribution to journalArticle

Shardell, Michelle ; Hicks, Gregory E. ; Ferrucci, Luigi. / Doubly robust estimation and causal inference in longitudinal studies with dropout and truncation by death. In: Biostatistics. 2015 ; Vol. 16, No. 1. pp. 155-168.
@article{01245b3fafba4fc8892542e036633feb,
title = "Doubly robust estimation and causal inference in longitudinal studies with dropout and truncation by death",
abstract = "Motivated by aging research, we propose an estimator of the effect of a time-varying exposure on an outcome in longitudinal studies with dropout and truncation by death. We use an inverse-probability weighted (IPW) estimator to derive a doubly robust augmented inverse-probability weighted (AIPW) estimator. IPW estimation involves weights for the exposure mechanism, dropout, and mortality; AIPW estimation additionally involves estimating data-generating models via regression. We demonstrate that the estimators identify a causal contrast that is a function of principal strata effects under a set of assumptions. Simulations show that AIPW estimation is unbiased when weights or outcome regressions are correct, and that AIPW estimation is more efficient than IPW estimation when all models are correct. We apply the method to a study of vitamin D and gait speed among older adults.",
keywords = "Causal inference, Longitudinal data analysis, Missing data, Observational studies",
author = "Michelle Shardell and Hicks, {Gregory E.} and Luigi Ferrucci",
year = "2015",
month = "1",
day = "1",
doi = "10.1093/biostatistics/kxu032",
language = "English (US)",
volume = "16",
pages = "155--168",
journal = "Biostatistics",
issn = "1465-4644",
publisher = "Oxford University Press",
number = "1",

}

TY - JOUR

T1 - Doubly robust estimation and causal inference in longitudinal studies with dropout and truncation by death

AU - Shardell, Michelle

AU - Hicks, Gregory E.

AU - Ferrucci, Luigi

PY - 2015/1/1

Y1 - 2015/1/1

N2 - Motivated by aging research, we propose an estimator of the effect of a time-varying exposure on an outcome in longitudinal studies with dropout and truncation by death. We use an inverse-probability weighted (IPW) estimator to derive a doubly robust augmented inverse-probability weighted (AIPW) estimator. IPW estimation involves weights for the exposure mechanism, dropout, and mortality; AIPW estimation additionally involves estimating data-generating models via regression. We demonstrate that the estimators identify a causal contrast that is a function of principal strata effects under a set of assumptions. Simulations show that AIPW estimation is unbiased when weights or outcome regressions are correct, and that AIPW estimation is more efficient than IPW estimation when all models are correct. We apply the method to a study of vitamin D and gait speed among older adults.

AB - Motivated by aging research, we propose an estimator of the effect of a time-varying exposure on an outcome in longitudinal studies with dropout and truncation by death. We use an inverse-probability weighted (IPW) estimator to derive a doubly robust augmented inverse-probability weighted (AIPW) estimator. IPW estimation involves weights for the exposure mechanism, dropout, and mortality; AIPW estimation additionally involves estimating data-generating models via regression. We demonstrate that the estimators identify a causal contrast that is a function of principal strata effects under a set of assumptions. Simulations show that AIPW estimation is unbiased when weights or outcome regressions are correct, and that AIPW estimation is more efficient than IPW estimation when all models are correct. We apply the method to a study of vitamin D and gait speed among older adults.

KW - Causal inference

KW - Longitudinal data analysis

KW - Missing data

KW - Observational studies

UR - http://www.scopus.com/inward/record.url?scp=84922540923&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84922540923&partnerID=8YFLogxK

U2 - 10.1093/biostatistics/kxu032

DO - 10.1093/biostatistics/kxu032

M3 - Article

C2 - 24997309

AN - SCOPUS:84922540923

VL - 16

SP - 155

EP - 168

JO - Biostatistics

JF - Biostatistics

SN - 1465-4644

IS - 1

ER -