Estimating population treatment effects from a survey subsample

Kara Rudolph, Iván Díaz, Michael Rosenblum, Elizabeth A. Stuart

Research output: Contribution to journalArticle

9 Citations (Scopus)

Abstract

We considered the problem of estimating an average treatment effect for a target population using a survey subsample. Our motivation was to generalize a treatment effect that was estimated in a subsample of the National Comorbidity Survey Replication Adolescent Supplement (2001-2004) to the population of US adolescents. To address this problem, we evaluated easy-to-implement methods that account for both nonrandom treatment assignment and a nonrandom 2-stage selection mechanism. We compared the performance of a Horvitz-Thompson estimator using inverse probability weighting and 2 doubly robust estimators in a variety of scenarios. We demonstrated that the 2 doubly robust estimators generally outperformed inverse probability weighting in terms of mean-squared error even under misspecification of one of the treatment, selection, or outcome models. Moreover, the doubly robust estimators are easy to implement and provide an attractive alternative to inverse probability weighting for applied epidemiologic researchers. We demonstrated how to apply these estimators to our motivating example.

Original languageEnglish (US)
Pages (from-to)737-748
Number of pages12
JournalAmerican Journal of Epidemiology
Volume180
Issue number7
DOIs
StatePublished - Oct 1 2014
Externally publishedYes

Fingerprint

Population
Health Services Needs and Demand
Therapeutics
Comorbidity
Research Personnel
Surveys and Questionnaires

Keywords

  • causal inference
  • inverse probability weighting
  • survey
  • targeted maximum likelihood estimation

ASJC Scopus subject areas

  • Epidemiology

Cite this

Estimating population treatment effects from a survey subsample. / Rudolph, Kara; Díaz, Iván; Rosenblum, Michael; Stuart, Elizabeth A.

In: American Journal of Epidemiology, Vol. 180, No. 7, 01.10.2014, p. 737-748.

Research output: Contribution to journalArticle

Rudolph, Kara ; Díaz, Iván ; Rosenblum, Michael ; Stuart, Elizabeth A. / Estimating population treatment effects from a survey subsample. In: American Journal of Epidemiology. 2014 ; Vol. 180, No. 7. pp. 737-748.
@article{97a0bc447f444d09b08c021acf4c5133,
title = "Estimating population treatment effects from a survey subsample",
abstract = "We considered the problem of estimating an average treatment effect for a target population using a survey subsample. Our motivation was to generalize a treatment effect that was estimated in a subsample of the National Comorbidity Survey Replication Adolescent Supplement (2001-2004) to the population of US adolescents. To address this problem, we evaluated easy-to-implement methods that account for both nonrandom treatment assignment and a nonrandom 2-stage selection mechanism. We compared the performance of a Horvitz-Thompson estimator using inverse probability weighting and 2 doubly robust estimators in a variety of scenarios. We demonstrated that the 2 doubly robust estimators generally outperformed inverse probability weighting in terms of mean-squared error even under misspecification of one of the treatment, selection, or outcome models. Moreover, the doubly robust estimators are easy to implement and provide an attractive alternative to inverse probability weighting for applied epidemiologic researchers. We demonstrated how to apply these estimators to our motivating example.",
keywords = "causal inference, inverse probability weighting, survey, targeted maximum likelihood estimation",
author = "Kara Rudolph and Iv{\'a}n D{\'i}az and Michael Rosenblum and Stuart, {Elizabeth A.}",
year = "2014",
month = "10",
day = "1",
doi = "10.1093/aje/kwu197",
language = "English (US)",
volume = "180",
pages = "737--748",
journal = "American Journal of Epidemiology",
issn = "0002-9262",
publisher = "Oxford University Press",
number = "7",

}

TY - JOUR

T1 - Estimating population treatment effects from a survey subsample

AU - Rudolph, Kara

AU - Díaz, Iván

AU - Rosenblum, Michael

AU - Stuart, Elizabeth A.

PY - 2014/10/1

Y1 - 2014/10/1

N2 - We considered the problem of estimating an average treatment effect for a target population using a survey subsample. Our motivation was to generalize a treatment effect that was estimated in a subsample of the National Comorbidity Survey Replication Adolescent Supplement (2001-2004) to the population of US adolescents. To address this problem, we evaluated easy-to-implement methods that account for both nonrandom treatment assignment and a nonrandom 2-stage selection mechanism. We compared the performance of a Horvitz-Thompson estimator using inverse probability weighting and 2 doubly robust estimators in a variety of scenarios. We demonstrated that the 2 doubly robust estimators generally outperformed inverse probability weighting in terms of mean-squared error even under misspecification of one of the treatment, selection, or outcome models. Moreover, the doubly robust estimators are easy to implement and provide an attractive alternative to inverse probability weighting for applied epidemiologic researchers. We demonstrated how to apply these estimators to our motivating example.

AB - We considered the problem of estimating an average treatment effect for a target population using a survey subsample. Our motivation was to generalize a treatment effect that was estimated in a subsample of the National Comorbidity Survey Replication Adolescent Supplement (2001-2004) to the population of US adolescents. To address this problem, we evaluated easy-to-implement methods that account for both nonrandom treatment assignment and a nonrandom 2-stage selection mechanism. We compared the performance of a Horvitz-Thompson estimator using inverse probability weighting and 2 doubly robust estimators in a variety of scenarios. We demonstrated that the 2 doubly robust estimators generally outperformed inverse probability weighting in terms of mean-squared error even under misspecification of one of the treatment, selection, or outcome models. Moreover, the doubly robust estimators are easy to implement and provide an attractive alternative to inverse probability weighting for applied epidemiologic researchers. We demonstrated how to apply these estimators to our motivating example.

KW - causal inference

KW - inverse probability weighting

KW - survey

KW - targeted maximum likelihood estimation

UR - http://www.scopus.com/inward/record.url?scp=84930387469&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84930387469&partnerID=8YFLogxK

U2 - 10.1093/aje/kwu197

DO - 10.1093/aje/kwu197

M3 - Article

VL - 180

SP - 737

EP - 748

JO - American Journal of Epidemiology

JF - American Journal of Epidemiology

SN - 0002-9262

IS - 7

ER -