Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2012 Jun 15;31(13):1380-404.
doi: 10.1002/sim.4469. Epub 2012 Feb 23.

Causal inference in epidemiological studies with strong confounding

Kelly L Moore  1 , Romain Neugebauer, Mark J van der Laan, Ira B Tager
Affiliations

Causal inference in epidemiological studies with strong confounding

Kelly L Moore et al. Stat Med. .

Abstract

One of the identifiability assumptions of causal effects defined by marginal structural model (MSM) parameters is the experimental treatment assignment (ETA) assumption. Practical violations of this assumption frequently occur in data analysis when certain exposures are rarely observed within some strata of the population. The inverse probability of treatment weighted (IPTW) estimator is particularly sensitive to violations of this assumption; however, we demonstrate that this is a problem for all estimators of causal effects. This is due to the fact that the ETA assumption is about information (or lack thereof) in the data. A new class of causal models, causal models for realistic individualized exposure rules (CMRIER), is based on dynamic interventions. CMRIER generalize MSM, and their parameters remain fully identifiable from the observed data, even when the ETA assumption is violated, if the dynamic interventions are set to be realistic. Examples of such realistic interventions are provided. We argue that causal effects defined by CMRIER may be more appropriate in many situations, particularly those with policy considerations. Through simulation studies, we examine the performance of the IPTW estimator of the CMRIER parameters in contrast to that of the MSM parameters. We also apply the methodology to a real data analysis in air pollution epidemiology to illustrate the interpretation of the causal effects defined by CMRIER.

PubMed Disclaimer

Figures

Figure 1
Figure 1
Motivational simulation study: consistency and MSE of IPTW and G-computation estimates. ‘GCOMP CORRECT’ is the G-computation estimator and equivalently in this simulation study the traditional estimator when the model for E(Y | A, W) is correctly specified; ‘GCOMP STEPWISE’ is the G-computation/traditional estimator when A is not forced into the selected model for E(Y | A, W); and ‘IPTW’ is the IPTW estimator.
Figure 2
Figure 2
Consistency and MSE of the IPTW estimator with the correctly specified treatment mechanism (‘IPTW’) and the IPTW estimator where the weights based on the correctly specified treatment mechanism are truncated between 1.1 and 10.
Figure 3
Figure 3
Consistency results of IPTW estimator with correctly specified treatment mechanism for simulation study with discretized categorical exposure (A*) with three levels where E(Ya*): m(a* | β) = β0+β1I(a* = 1)+β2I(a* = 2).
Figure 4
Figure 4
Consistency results of IPTW estimator with correctly specified treatment mechanism for simulation study with discretized binary exposure (A*) where E(Ya*): m(a* | β) = β0 + β1a*.
Figure 5
Figure 5
Consistency results of the IPTW estimator of the MSM and CMRIER parameters for the simulation study with binary exposure.
Figure 6
Figure 6
MSE results of the IPTW estimator of the MSM and CMRIER parameters for the simulation study with binary exposure.
Figure 7
Figure 7
Consistency results of the IPTW estimator of the MSM and CMRIER parameters for the simulation study with categorical (three levels) exposure.
Figure 8
Figure 8
MSE results of the IPTW estimator of the MSM and CMRIER parameters for the simulation study with categorical (three levels) exposure.
See this image and copyright information in PMC

References

    1. Robins JM. Marginal structural models versus structural nested models as tools for causal inference. AAAI Technical Report; Symposium on prospects for a common sense theory of causation; Stanford, CA. 1998.
    1. Robins JM, Hernan MA, Brumback B. Marginal structural models and causal inference in epidemiology. Epidemiology. 2000;11(5):550–560. - PubMed
    1. Neugebauer R, van der Laan MJ. Why prefer double robust estimators in causal inference? Journal of Statistical Planning and Inference. 2005;129(1–2):405–426.
    1. van der Laan MJ, Robins J. Unified Methods for Censored Longitudinal Data and Causality. Springer; New York: 2003.
    1. Moore K, Neugebauer R, Lurmann F, Hall J, Brajer V, Alcorn S, Tager I. Ambient ozone concentrations cause increased hospitalizations for asthma in children: an 18-year study in Southern California. Environmental Health Perspectives. 2008;116(8):1063–1070. - PMC - PubMed

LinkOut - more resources