Conditioning on intermediates in perinatal epidemiology

Tyler J. Vanderweele, Sunni L. Mumford, Enrique F. Schisterman

Research output: Contribution to journalArticle

Abstract

It is common practice in perinatal epidemiology to calculate gestational-age-specific or birth-weight-specific associations between an exposure and a perinatal outcome. Gestational age or birth weight, for example, might lie on a pathway from the exposure to the outcome. This practice of conditioning on a potential intermediate has come under critique for various reasons. First, if one is interested in assessing the overall effect of an exposure on an outcome, it is not necessary to stratify, and indeed, it is important not to stratify, on an intermediate. Second, if one does condition on an intermediate, to try to obtain what might conceived of as a "direct effect" of the exposure on the outcome, then various biases and paradoxical results can arise. It is now well documented theoretically and empirically that, when there is an unmeasured common cause of the intermediate and the outcome, associations adjusted for the intermediate are subject to bias. In this paper, we propose 3 approaches to facilitate valid inference when effects conditional on an intermediate are in view. These 3 approaches correspond to (i) conditioning on the predicted risk of the intermediate, (ii) conditioning on the intermediate itself in conjunction with sensitivity analysis, and (iii) conditioning on the subgroup of individuals for whom the intermediate would occur irrespective of the exposure received. The second and third approaches both require sensitivity analysis, and they result in a range of estimates. Each of the 3 approaches can be used to resolve the "birth-weight paradox" that exposures such as maternal smoking seem to have a protective effect among low-birth-weight infants. The various methodologic approaches described in this paper are applicable to a number of similar settings in perinatal epidemiology.

Original languageEnglish (US)
Pages (from-to)1-9
Number of pages9
JournalEpidemiology
Volume23
Issue number1
DOIs
StatePublished - Jan 2012
Externally publishedYes

Fingerprint

Epidemiology
Birth Weight
Gestational Age
Low Birth Weight Infant
Smoking
Mothers
Conditioning (Psychology)

ASJC Scopus subject areas

  • Epidemiology

Cite this

Vanderweele, T. J., Mumford, S. L., & Schisterman, E. F. (2012). Conditioning on intermediates in perinatal epidemiology. Epidemiology, 23(1), 1-9. https://doi.org/10.1097/EDE.0b013e31823aca5d

Conditioning on intermediates in perinatal epidemiology. / Vanderweele, Tyler J.; Mumford, Sunni L.; Schisterman, Enrique F.

In: Epidemiology, Vol. 23, No. 1, 01.2012, p. 1-9.

Research output: Contribution to journalArticle

Vanderweele, TJ, Mumford, SL & Schisterman, EF 2012, 'Conditioning on intermediates in perinatal epidemiology', Epidemiology, vol. 23, no. 1, pp. 1-9. https://doi.org/10.1097/EDE.0b013e31823aca5d
Vanderweele, Tyler J. ; Mumford, Sunni L. ; Schisterman, Enrique F. / Conditioning on intermediates in perinatal epidemiology. In: Epidemiology. 2012 ; Vol. 23, No. 1. pp. 1-9.
@article{94390f3757024e37963eeb79629c55a3,
title = "Conditioning on intermediates in perinatal epidemiology",
abstract = "It is common practice in perinatal epidemiology to calculate gestational-age-specific or birth-weight-specific associations between an exposure and a perinatal outcome. Gestational age or birth weight, for example, might lie on a pathway from the exposure to the outcome. This practice of conditioning on a potential intermediate has come under critique for various reasons. First, if one is interested in assessing the overall effect of an exposure on an outcome, it is not necessary to stratify, and indeed, it is important not to stratify, on an intermediate. Second, if one does condition on an intermediate, to try to obtain what might conceived of as a {"}direct effect{"} of the exposure on the outcome, then various biases and paradoxical results can arise. It is now well documented theoretically and empirically that, when there is an unmeasured common cause of the intermediate and the outcome, associations adjusted for the intermediate are subject to bias. In this paper, we propose 3 approaches to facilitate valid inference when effects conditional on an intermediate are in view. These 3 approaches correspond to (i) conditioning on the predicted risk of the intermediate, (ii) conditioning on the intermediate itself in conjunction with sensitivity analysis, and (iii) conditioning on the subgroup of individuals for whom the intermediate would occur irrespective of the exposure received. The second and third approaches both require sensitivity analysis, and they result in a range of estimates. Each of the 3 approaches can be used to resolve the {"}birth-weight paradox{"} that exposures such as maternal smoking seem to have a protective effect among low-birth-weight infants. The various methodologic approaches described in this paper are applicable to a number of similar settings in perinatal epidemiology.",
author = "Vanderweele, {Tyler J.} and Mumford, {Sunni L.} and Schisterman, {Enrique F.}",
year = "2012",
month = "1",
doi = "10.1097/EDE.0b013e31823aca5d",
language = "English (US)",
volume = "23",
pages = "1--9",
journal = "Epidemiology",
issn = "1044-3983",
publisher = "Lippincott Williams and Wilkins",
number = "1",

}

TY - JOUR

T1 - Conditioning on intermediates in perinatal epidemiology

AU - Vanderweele, Tyler J.

AU - Mumford, Sunni L.

AU - Schisterman, Enrique F.

PY - 2012/1

Y1 - 2012/1

N2 - It is common practice in perinatal epidemiology to calculate gestational-age-specific or birth-weight-specific associations between an exposure and a perinatal outcome. Gestational age or birth weight, for example, might lie on a pathway from the exposure to the outcome. This practice of conditioning on a potential intermediate has come under critique for various reasons. First, if one is interested in assessing the overall effect of an exposure on an outcome, it is not necessary to stratify, and indeed, it is important not to stratify, on an intermediate. Second, if one does condition on an intermediate, to try to obtain what might conceived of as a "direct effect" of the exposure on the outcome, then various biases and paradoxical results can arise. It is now well documented theoretically and empirically that, when there is an unmeasured common cause of the intermediate and the outcome, associations adjusted for the intermediate are subject to bias. In this paper, we propose 3 approaches to facilitate valid inference when effects conditional on an intermediate are in view. These 3 approaches correspond to (i) conditioning on the predicted risk of the intermediate, (ii) conditioning on the intermediate itself in conjunction with sensitivity analysis, and (iii) conditioning on the subgroup of individuals for whom the intermediate would occur irrespective of the exposure received. The second and third approaches both require sensitivity analysis, and they result in a range of estimates. Each of the 3 approaches can be used to resolve the "birth-weight paradox" that exposures such as maternal smoking seem to have a protective effect among low-birth-weight infants. The various methodologic approaches described in this paper are applicable to a number of similar settings in perinatal epidemiology.

AB - It is common practice in perinatal epidemiology to calculate gestational-age-specific or birth-weight-specific associations between an exposure and a perinatal outcome. Gestational age or birth weight, for example, might lie on a pathway from the exposure to the outcome. This practice of conditioning on a potential intermediate has come under critique for various reasons. First, if one is interested in assessing the overall effect of an exposure on an outcome, it is not necessary to stratify, and indeed, it is important not to stratify, on an intermediate. Second, if one does condition on an intermediate, to try to obtain what might conceived of as a "direct effect" of the exposure on the outcome, then various biases and paradoxical results can arise. It is now well documented theoretically and empirically that, when there is an unmeasured common cause of the intermediate and the outcome, associations adjusted for the intermediate are subject to bias. In this paper, we propose 3 approaches to facilitate valid inference when effects conditional on an intermediate are in view. These 3 approaches correspond to (i) conditioning on the predicted risk of the intermediate, (ii) conditioning on the intermediate itself in conjunction with sensitivity analysis, and (iii) conditioning on the subgroup of individuals for whom the intermediate would occur irrespective of the exposure received. The second and third approaches both require sensitivity analysis, and they result in a range of estimates. Each of the 3 approaches can be used to resolve the "birth-weight paradox" that exposures such as maternal smoking seem to have a protective effect among low-birth-weight infants. The various methodologic approaches described in this paper are applicable to a number of similar settings in perinatal epidemiology.

UR - http://www.scopus.com/inward/record.url?scp=83655183679&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=83655183679&partnerID=8YFLogxK

U2 - 10.1097/EDE.0b013e31823aca5d

DO - 10.1097/EDE.0b013e31823aca5d

M3 - Article

C2 - 22157298

AN - SCOPUS:83655183679

VL - 23

SP - 1

EP - 9

JO - Epidemiology

JF - Epidemiology

SN - 1044-3983

IS - 1

ER -