Estimating the complier average causal effect via a latent class approach using gsem

Patricio Troncoso; Ana Morales-Gomez

doi:10.1177/1536867X221106416

Estimating the complier average causal effect via a latent class approach using gsem

Patricio Troncoso^*, Ana Morales-Gomez

^*Corresponding author for this work

Research output: Contribution to journal › Article › peer-review

Abstract / Description of output

In randomized control trials (RCT), intention-to-treat (ITT) analysis is customarily used to estimate the effect of the trial; however, in the presence of noncompliance, this can often lead to biased estimates because ITT completely ignores varying levels of actual treatment received. This is a known issue that can be overcome by adopting the complier average causal effect (CACE) approach, which estimates the effect the trial had on the individuals who complied with the protocol. When compliance is unobserved in the control group, the CACE estimate can be obtained via a latent class specification using the gsem command.

Original language	English
Pages (from-to)	404-415
Number of pages	12
Journal	The Stata Journal
Volume	22
Issue number	2
Early online date	30 Jun 2022
DOIs	https://doi.org/10.1177/1536867X221106416
Publication status	Published - Jun 2022

Keywords / Materials (for Non-textual outputs)

randomized control trial
complier average causal effect
mixture modeling
latent class modeling
compliance
adherence
gsem
st0677

Access to Document

10.1177/1536867X221106416Licence: Creative Commons: Attribution (CC-BY)

TroncosoPMorales-GomezA2022TSJEstimatingTheComplierAverageFinal published version, 234 KBLicence: Creative Commons: Attribution (CC-BY)

Understanding Children's Lives and Outcomes
Dibben, C. & Williamson, L.
ESRC
1/10/20 → 30/09/22
Project: Research

Cite this

@article{89af137b0074469ca356b8f186caa410,

title = "Estimating the complier average causal effect via a latent class approach using gsem",

abstract = "In randomized control trials (RCT), intention-to-treat (ITT) analysis is customarily used to estimate the effect of the trial; however, in the presence of noncompliance, this can often lead to biased estimates because ITT completely ignores varying levels of actual treatment received. This is a known issue that can be overcome by adopting the complier average causal effect (CACE) approach, which estimates the effect the trial had on the individuals who complied with the protocol. When compliance is unobserved in the control group, the CACE estimate can be obtained via a latent class specification using the gsem command.",

keywords = "randomized control trial, complier average causal effect, mixture modeling, latent class modeling, compliance, adherence, gsem, st0677",

author = "Patricio Troncoso and Ana Morales-Gomez",

note = "Funding Information: This work was supported by the Economic and Social Research Council through the following grants: 1) Understanding Children{\textquoteright}s Lives and Outcomes and 2) Scottish Centre for Administrative Data Research (grant numbers ES/V011243/1 and ES/S007407/1, respectively). This work was also supported in its early stages by the National Institute for Health Research (grant number 14/52/38). The authors are affiliated with the Scottish Centre for Administrative Data Research ( scadr ), which is part of the Administrative Data Research U.K. partnership, funded by the Economic and Social Research Council. Publisher Copyright: {\textcopyright} StataCorp LLC 2022.",

year = "2022",

month = jun,

doi = "10.1177/1536867X221106416",

language = "English",

volume = "22",

pages = "404--415",

journal = "The Stata Journal",

issn = "1536-867X",

publisher = "SAGE Publications",

number = "2",

}

TY - JOUR

T1 - Estimating the complier average causal effect via a latent class approach using gsem

AU - Troncoso, Patricio

AU - Morales-Gomez, Ana

N1 - Funding Information: This work was supported by the Economic and Social Research Council through the following grants: 1) Understanding Children’s Lives and Outcomes and 2) Scottish Centre for Administrative Data Research (grant numbers ES/V011243/1 and ES/S007407/1, respectively). This work was also supported in its early stages by the National Institute for Health Research (grant number 14/52/38). The authors are affiliated with the Scottish Centre for Administrative Data Research ( scadr ), which is part of the Administrative Data Research U.K. partnership, funded by the Economic and Social Research Council. Publisher Copyright: © StataCorp LLC 2022.

PY - 2022/6

Y1 - 2022/6

N2 - In randomized control trials (RCT), intention-to-treat (ITT) analysis is customarily used to estimate the effect of the trial; however, in the presence of noncompliance, this can often lead to biased estimates because ITT completely ignores varying levels of actual treatment received. This is a known issue that can be overcome by adopting the complier average causal effect (CACE) approach, which estimates the effect the trial had on the individuals who complied with the protocol. When compliance is unobserved in the control group, the CACE estimate can be obtained via a latent class specification using the gsem command.

AB - In randomized control trials (RCT), intention-to-treat (ITT) analysis is customarily used to estimate the effect of the trial; however, in the presence of noncompliance, this can often lead to biased estimates because ITT completely ignores varying levels of actual treatment received. This is a known issue that can be overcome by adopting the complier average causal effect (CACE) approach, which estimates the effect the trial had on the individuals who complied with the protocol. When compliance is unobserved in the control group, the CACE estimate can be obtained via a latent class specification using the gsem command.

KW - randomized control trial

KW - complier average causal effect

KW - mixture modeling

KW - latent class modeling

KW - compliance

KW - adherence

KW - gsem

KW - st0677

U2 - 10.1177/1536867X221106416

DO - 10.1177/1536867X221106416

M3 - Article

SN - 1536-867X

VL - 22

SP - 404

EP - 415

JO - The Stata Journal

JF - The Stata Journal

IS - 2

ER -

Estimating the complier average causal effect via a latent class approach using gsem

Abstract / Description of output

Keywords / Materials (for Non-textual outputs)

Access to Document

Fingerprint

Projects

Understanding Children's Lives and Outcomes

Cite this