Cargando…

Universal Difference-in-Differences for Causal Inference in Epidemiology

Difference-in-differences is undoubtedly one of the most widely used methods for evaluating the causal effect of an intervention in observational (i.e., nonrandomized) settings. The approach is typically used when pre- and postexposure outcome measurements are available, and one can reasonably assum...

Descripción completa

Detalles Bibliográficos
Autores principales: Tchetgen Tchetgen, Eric J., Park, Chan, Richardson, David B.
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Lippincott Williams & Wilkins 2023
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10683972/
https://www.ncbi.nlm.nih.gov/pubmed/38032801
http://dx.doi.org/10.1097/EDE.0000000000001676
_version_ 1785151299205988352
author Tchetgen Tchetgen, Eric J.
Park, Chan
Richardson, David B.
author_facet Tchetgen Tchetgen, Eric J.
Park, Chan
Richardson, David B.
author_sort Tchetgen Tchetgen, Eric J.
collection PubMed
description Difference-in-differences is undoubtedly one of the most widely used methods for evaluating the causal effect of an intervention in observational (i.e., nonrandomized) settings. The approach is typically used when pre- and postexposure outcome measurements are available, and one can reasonably assume that the association of the unobserved confounder with the outcome has the same absolute magnitude in the two exposure arms and is constant over time; a so-called parallel trends assumption. The parallel trends assumption may not be credible in many practical settings, for example, if the outcome is binary, a count, or polytomous, as well as when an uncontrolled confounder exhibits nonadditive effects on the distribution of the outcome, even if such effects are constant over time. We introduce an alternative approach that replaces the parallel trends assumption with an odds ratio equi-confounding assumption under which an association between treatment and the potential outcome under no treatment is identified with a well-specified generalized linear model relating the pre-exposure outcome and the exposure. Because the proposed method identifies any causal effect that is conceivably identified in the absence of confounding bias, including nonlinear effects such as quantile treatment effects, the approach is aptly called universal difference-in-differences. We describe and illustrate both fully parametric and more robust semiparametric universal difference-in-differences estimators in a real-world application concerning the causal effects of a Zika virus outbreak on birth rate in Brazil. A supplementary digital video is available at: http://links.lww.com/EDE/C90
format Online
Article
Text
id pubmed-10683972
institution National Center for Biotechnology Information
language English
publishDate 2023
publisher Lippincott Williams & Wilkins
record_format MEDLINE/PubMed
spelling pubmed-106839722023-11-30 Universal Difference-in-Differences for Causal Inference in Epidemiology Tchetgen Tchetgen, Eric J. Park, Chan Richardson, David B. Epidemiology Methods Difference-in-differences is undoubtedly one of the most widely used methods for evaluating the causal effect of an intervention in observational (i.e., nonrandomized) settings. The approach is typically used when pre- and postexposure outcome measurements are available, and one can reasonably assume that the association of the unobserved confounder with the outcome has the same absolute magnitude in the two exposure arms and is constant over time; a so-called parallel trends assumption. The parallel trends assumption may not be credible in many practical settings, for example, if the outcome is binary, a count, or polytomous, as well as when an uncontrolled confounder exhibits nonadditive effects on the distribution of the outcome, even if such effects are constant over time. We introduce an alternative approach that replaces the parallel trends assumption with an odds ratio equi-confounding assumption under which an association between treatment and the potential outcome under no treatment is identified with a well-specified generalized linear model relating the pre-exposure outcome and the exposure. Because the proposed method identifies any causal effect that is conceivably identified in the absence of confounding bias, including nonlinear effects such as quantile treatment effects, the approach is aptly called universal difference-in-differences. We describe and illustrate both fully parametric and more robust semiparametric universal difference-in-differences estimators in a real-world application concerning the causal effects of a Zika virus outbreak on birth rate in Brazil. A supplementary digital video is available at: http://links.lww.com/EDE/C90 Lippincott Williams & Wilkins 2023-11-27 2024-01 /pmc/articles/PMC10683972/ /pubmed/38032801 http://dx.doi.org/10.1097/EDE.0000000000001676 Text en Copyright © 2023 The Author(s). Published by Wolters Kluwer Health, Inc. https://creativecommons.org/licenses/by-nc-nd/4.0/This is an open-access article distributed under the terms of the Creative Commons Attribution-Non Commercial-No Derivatives License 4.0 (CCBY-NC-ND) (https://creativecommons.org/licenses/by-nc-nd/4.0/) , where it is permissible to download and share the work provided it is properly cited. The work cannot be changed in any way or used commercially without permission from the journal.
spellingShingle Methods
Tchetgen Tchetgen, Eric J.
Park, Chan
Richardson, David B.
Universal Difference-in-Differences for Causal Inference in Epidemiology
title Universal Difference-in-Differences for Causal Inference in Epidemiology
title_full Universal Difference-in-Differences for Causal Inference in Epidemiology
title_fullStr Universal Difference-in-Differences for Causal Inference in Epidemiology
title_full_unstemmed Universal Difference-in-Differences for Causal Inference in Epidemiology
title_short Universal Difference-in-Differences for Causal Inference in Epidemiology
title_sort universal difference-in-differences for causal inference in epidemiology
topic Methods
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10683972/
https://www.ncbi.nlm.nih.gov/pubmed/38032801
http://dx.doi.org/10.1097/EDE.0000000000001676
work_keys_str_mv AT tchetgentchetgenericj universaldifferenceindifferencesforcausalinferenceinepidemiology
AT parkchan universaldifferenceindifferencesforcausalinferenceinepidemiology
AT richardsondavidb universaldifferenceindifferencesforcausalinferenceinepidemiology