Targeted Maximum Likelihood Estimation of Natural Direct Effect
In many causal inference problems, one is interested in the direct causal effect of an exposure on an outcome of interest that is not mediated by certain intermediate variables. Robins and Greenland (1992) and Pearl (2000) formalized the definition of two types of direct effects (natural and control...
Main Authors: | , |
---|---|
Format: | Text |
Language: | unknown |
Published: |
Collection of Biostatistics Research Archive
2011
|
Subjects: | |
Online Access: | https://biostats.bepress.com/ucbbiostat/paper288 https://biostats.bepress.com/cgi/viewcontent.cgi?article=1291&context=ucbbiostat |
id |
ftcobra:oai:biostats.bepress.com:ucbbiostat-1291 |
---|---|
record_format |
openpolar |
spelling |
ftcobra:oai:biostats.bepress.com:ucbbiostat-1291 2023-05-15T16:29:41+02:00 Targeted Maximum Likelihood Estimation of Natural Direct Effect Zheng, Wenjing van der Laan, Mark J. 2011-07-19T07:00:00Z application/pdf https://biostats.bepress.com/ucbbiostat/paper288 https://biostats.bepress.com/cgi/viewcontent.cgi?article=1291&context=ucbbiostat unknown Collection of Biostatistics Research Archive https://biostats.bepress.com/ucbbiostat/paper288 https://biostats.bepress.com/cgi/viewcontent.cgi?article=1291&context=ucbbiostat U.C. Berkeley Division of Biostatistics Working Paper Series Natural direct effects mediation analysis mediation formula mediator direct effects asymptotic efficiency Biostatistics text 2011 ftcobra 2022-02-01T14:06:10Z In many causal inference problems, one is interested in the direct causal effect of an exposure on an outcome of interest that is not mediated by certain intermediate variables. Robins and Greenland (1992) and Pearl (2000) formalized the definition of two types of direct effects (natural and controlled) under the counterfactual framework. Since then, identifiability conditions for these effects have been studied extensively. By contrast, considerably fewer efforts have been invested in the estimation problem of the natural direct effect. In this article, we propose a semiparametric efficient, multiply robust estimator for the natural direct effect of a binary treatment using the targeted maximum likelihood framework of van der Laan and Rubin (2006) and van der Laan and Rose (2011). The proposed estimator is asymptotically unbiased if either one of the following holds: i) the conditional outcome expectation given exposure, mediator, and confounders, and the mediated mean outcome difference are consistently estimated; (ii) the exposure mechanism given confounders, and the conditional outcome expectation are consistently estimated; or (iii) the exposure mechanism given confounders, and a ratio of conditional mediator densities are consistently estimated. Moreover, case (iii) implies in particular that estimation of the conditional mediator density may be replaced by consistent estimation of the exposure mechanism and the conditional distribution of exposure given confounders and mediator. If all three conditions hold, then the effect estimate is asymptotically efficient. Text Greenland Collection of Biostatistics Research Archive (COBRA) Greenland Rubin ENVELOPE(65.493,65.493,-73.438,-73.438) |
institution |
Open Polar |
collection |
Collection of Biostatistics Research Archive (COBRA) |
op_collection_id |
ftcobra |
language |
unknown |
topic |
Natural direct effects mediation analysis mediation formula mediator direct effects asymptotic efficiency Biostatistics |
spellingShingle |
Natural direct effects mediation analysis mediation formula mediator direct effects asymptotic efficiency Biostatistics Zheng, Wenjing van der Laan, Mark J. Targeted Maximum Likelihood Estimation of Natural Direct Effect |
topic_facet |
Natural direct effects mediation analysis mediation formula mediator direct effects asymptotic efficiency Biostatistics |
description |
In many causal inference problems, one is interested in the direct causal effect of an exposure on an outcome of interest that is not mediated by certain intermediate variables. Robins and Greenland (1992) and Pearl (2000) formalized the definition of two types of direct effects (natural and controlled) under the counterfactual framework. Since then, identifiability conditions for these effects have been studied extensively. By contrast, considerably fewer efforts have been invested in the estimation problem of the natural direct effect. In this article, we propose a semiparametric efficient, multiply robust estimator for the natural direct effect of a binary treatment using the targeted maximum likelihood framework of van der Laan and Rubin (2006) and van der Laan and Rose (2011). The proposed estimator is asymptotically unbiased if either one of the following holds: i) the conditional outcome expectation given exposure, mediator, and confounders, and the mediated mean outcome difference are consistently estimated; (ii) the exposure mechanism given confounders, and the conditional outcome expectation are consistently estimated; or (iii) the exposure mechanism given confounders, and a ratio of conditional mediator densities are consistently estimated. Moreover, case (iii) implies in particular that estimation of the conditional mediator density may be replaced by consistent estimation of the exposure mechanism and the conditional distribution of exposure given confounders and mediator. If all three conditions hold, then the effect estimate is asymptotically efficient. |
format |
Text |
author |
Zheng, Wenjing van der Laan, Mark J. |
author_facet |
Zheng, Wenjing van der Laan, Mark J. |
author_sort |
Zheng, Wenjing |
title |
Targeted Maximum Likelihood Estimation of Natural Direct Effect |
title_short |
Targeted Maximum Likelihood Estimation of Natural Direct Effect |
title_full |
Targeted Maximum Likelihood Estimation of Natural Direct Effect |
title_fullStr |
Targeted Maximum Likelihood Estimation of Natural Direct Effect |
title_full_unstemmed |
Targeted Maximum Likelihood Estimation of Natural Direct Effect |
title_sort |
targeted maximum likelihood estimation of natural direct effect |
publisher |
Collection of Biostatistics Research Archive |
publishDate |
2011 |
url |
https://biostats.bepress.com/ucbbiostat/paper288 https://biostats.bepress.com/cgi/viewcontent.cgi?article=1291&context=ucbbiostat |
long_lat |
ENVELOPE(65.493,65.493,-73.438,-73.438) |
geographic |
Greenland Rubin |
geographic_facet |
Greenland Rubin |
genre |
Greenland |
genre_facet |
Greenland |
op_source |
U.C. Berkeley Division of Biostatistics Working Paper Series |
op_relation |
https://biostats.bepress.com/ucbbiostat/paper288 https://biostats.bepress.com/cgi/viewcontent.cgi?article=1291&context=ucbbiostat |
_version_ |
1766019397802000384 |