Handling missing data when estimating causal effects with targeted maximum likelihood estimation
American Journal of Epidemiology · 08.07.2024 · 193 (7) · 1019-1030
DOI: 10.1093/aje/kwae012 ↗Kurz gefasst
Fachartikel in der Zeitschrift American Journal of Epidemiology, online erschienen am 22.02.2024. Verfasst von S Ghazaleh Dashti und 5 weiteren Autor:innen. Hauptthemen laut MEDLINE: Causality. Die Arbeit wurde bisher 19-mal zitiert (Crossref) und nennt 40 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Operational Infrastructure Support Program, National Health and Medical Research Council, Medical Research Council.
Abstract (Originalsprache)
Targeted maximum likelihood estimation (TMLE) is increasingly used for doubly robust causal inference, but how missing data should be handled when using TMLE with data-adaptive approaches is unclear. Based on data (1992-1998) from the Victorian Adolescent Health Cohort Study, we conducted a simulation study to evaluate 8 missing-data methods in this context: complete-case analysis, extended TMLE incorporating an outcome-missingness model, the missing covariate missing indicator method, and 5 multiple imputation (MI) approaches using parametric or machine-learning models. We considered 6 scenarios that varied in terms of exposure/outcome generation models (presence of confounder-confounder interactions) and missingness mechanisms (whether outcome influenced missingness in other variables and presence of interaction/nonlinear terms in missingness models). Complete-case analysis and extended TMLE had small biases when outcome did not influence missingness in other variables. Parametric MI without interactions had large bias when exposure/outcome generation models included interactions. Parametric MI including interactions performed best in bias and variance reduction across all settings, except when missingness models included a nonlinear term. When choosing a method for handling missing data in the context of TMLE, researchers must consider the missingness mechanism and, for MI, compatibility with the analysis method. In many settings, a parametric MI approach that incorporates interactions and nonlinearities is expected to perform well.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
ORCID-Profile: S Ghazaleh Dashti, Katherine J Lee, Julie A Simpson, Ian R White, John B Carlin, Margarita Moreno-Betancur
Förderung
- Operational Infrastructure Support Program
- National Health and Medical Research Council
- Medical Research Council
Zitieren
APA
Dashti, S. G., Lee, K. J., Simpson, J. A., White, I. R., Carlin, J. B., & Moreno-Betancur, M. (2024). Handling missing data when estimating causal effects with targeted maximum likelihood estimation. American Journal of Epidemiology, 193(7), 1019-1030. https://doi.org/10.1093/aje/kwae012
Vancouver
Dashti SG, Lee KJ, Simpson JA, White IR, Carlin JB, Moreno-Betancur M. Handling missing data when estimating causal effects with targeted maximum likelihood estimation. American Journal of Epidemiology. 2024;193(7):1019-1030. doi:10.1093/aje/kwae012
BibTeX
@article{Dashti2024,
author = {Dashti, S Ghazaleh and Lee, Katherine J and Simpson, Julie A and White, Ian R and Carlin, John B and Moreno-Betancur, Margarita},
title = {Handling missing data when estimating causal effects with targeted maximum likelihood estimation},
journal = {American Journal of Epidemiology},
year = {2024},
volume = {193},
number = {7},
pages = {1019--1030},
doi = {10.1093/aje/kwae012}
}
EUVIMED