Assessing the Treatment Effect Heterogeneity with a Latent Variable

Yunjian Yin, Lan Liu, Zhi Geng

Research output: Contribution to journalArticle

Abstract

The average treatment effect (ATE) is commonly used to assess the effect of treatment. However, the ATE implicitly assumes a homogenous treatment effect even amongst individuals with different characteristics. In order to describe the magnitude of heterogeneity, we define the treatment benefit rate (TBR) as the proportion of individuals in different subgroups who benefit from the treatment and define the treatment harm rate (THR) as the proportion harmed. These rates involve the joint distribution of the potential outcomes and cannot be identified without further assumptions, even in randomized clinical trials. Under the assumption that the potential outcomes are independent conditional on the observed covariates and an unmeasured latent variable, we show the identification of the TBR and THR in non-separable (generalized) linear mixed models for both continuous and binary outcomes. We then propose estimators and derive their asymptotic distributions. The proposed methods are implemented in an extensive simulation study and two randomized controlled trials.

Original languageEnglish (US)
Pages (from-to)115-135
Number of pages21
JournalStatistica Sinica
Volume28
Issue number1
DOIs
StatePublished - 2016

Fingerprint

Latent Variables
Treatment Effects
Average Treatment Effect
Potential Outcomes
Proportion
Randomized Controlled Trial
Randomized Clinical Trial
Binary Outcomes
Generalized Linear Mixed Model
Nonseparable
Joint Distribution
Asymptotic distribution
Treatment effects
Latent variables
Covariates
Simulation Study
Subgroup
Estimator

Cite this

Assessing the Treatment Effect Heterogeneity with a Latent Variable. / Yin, Yunjian; Liu, Lan; Geng, Zhi.

In: Statistica Sinica, Vol. 28, No. 1, 2016, p. 115-135.

Research output: Contribution to journalArticle

Yin, Yunjian ; Liu, Lan ; Geng, Zhi. / Assessing the Treatment Effect Heterogeneity with a Latent Variable. In: Statistica Sinica. 2016 ; Vol. 28, No. 1. pp. 115-135.
@article{ef1a0828949042eca45133940b5c2193,
title = "Assessing the Treatment Effect Heterogeneity with a Latent Variable",
abstract = "The average treatment effect (ATE) is commonly used to assess the effect of treatment. However, the ATE implicitly assumes a homogenous treatment effect even amongst individuals with different characteristics. In order to describe the magnitude of heterogeneity, we define the treatment benefit rate (TBR) as the proportion of individuals in different subgroups who benefit from the treatment and define the treatment harm rate (THR) as the proportion harmed. These rates involve the joint distribution of the potential outcomes and cannot be identified without further assumptions, even in randomized clinical trials. Under the assumption that the potential outcomes are independent conditional on the observed covariates and an unmeasured latent variable, we show the identification of the TBR and THR in non-separable (generalized) linear mixed models for both continuous and binary outcomes. We then propose estimators and derive their asymptotic distributions. The proposed methods are implemented in an extensive simulation study and two randomized controlled trials.",
author = "Yunjian Yin and Lan Liu and Zhi Geng",
year = "2016",
doi = "10.5705/ss.202016.0150",
language = "English (US)",
volume = "28",
pages = "115--135",
journal = "Statistica Sinica",
issn = "1017-0405",
publisher = "Institute of Statistical Science",
number = "1",

}

TY - JOUR

T1 - Assessing the Treatment Effect Heterogeneity with a Latent Variable

AU - Yin, Yunjian

AU - Liu, Lan

AU - Geng, Zhi

PY - 2016

Y1 - 2016

N2 - The average treatment effect (ATE) is commonly used to assess the effect of treatment. However, the ATE implicitly assumes a homogenous treatment effect even amongst individuals with different characteristics. In order to describe the magnitude of heterogeneity, we define the treatment benefit rate (TBR) as the proportion of individuals in different subgroups who benefit from the treatment and define the treatment harm rate (THR) as the proportion harmed. These rates involve the joint distribution of the potential outcomes and cannot be identified without further assumptions, even in randomized clinical trials. Under the assumption that the potential outcomes are independent conditional on the observed covariates and an unmeasured latent variable, we show the identification of the TBR and THR in non-separable (generalized) linear mixed models for both continuous and binary outcomes. We then propose estimators and derive their asymptotic distributions. The proposed methods are implemented in an extensive simulation study and two randomized controlled trials.

AB - The average treatment effect (ATE) is commonly used to assess the effect of treatment. However, the ATE implicitly assumes a homogenous treatment effect even amongst individuals with different characteristics. In order to describe the magnitude of heterogeneity, we define the treatment benefit rate (TBR) as the proportion of individuals in different subgroups who benefit from the treatment and define the treatment harm rate (THR) as the proportion harmed. These rates involve the joint distribution of the potential outcomes and cannot be identified without further assumptions, even in randomized clinical trials. Under the assumption that the potential outcomes are independent conditional on the observed covariates and an unmeasured latent variable, we show the identification of the TBR and THR in non-separable (generalized) linear mixed models for both continuous and binary outcomes. We then propose estimators and derive their asymptotic distributions. The proposed methods are implemented in an extensive simulation study and two randomized controlled trials.

UR - http://www.scopus.com/inward/record.url?scp=85040095902&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85040095902&partnerID=8YFLogxK

U2 - 10.5705/ss.202016.0150

DO - 10.5705/ss.202016.0150

M3 - Article

AN - SCOPUS:85040095902

VL - 28

SP - 115

EP - 135

JO - Statistica Sinica

JF - Statistica Sinica

SN - 1017-0405

IS - 1

ER -