This article describes a new command, gformula, that is an
implementation of the g-computation procedure. It is used to estimate the causal
effect of time-varying exposures on an outcome in the presence of time-varying
confounders that are themselves also affected by the exposures. The procedure
also addresses the related problem of estimating direct and indirect effects
when the causal effect of the exposures on an outcome is mediated by
intermediate variables, and in particular when confounders of the
mediator–outcome relationships are themselves affected by the exposures. A brief
overview of the theory and a description of the command and its options are
given, and illustrations using two simulated examples are provided.
AbbringJ. H.2003. Dynamic econometric program evaluation.
Discussion Paper No. 804, Institute for the Study of Labor (IZA). http://ftp.iza.org/dp804.pdf.
2.
AngristJ. D., and
PischkeJ.-S.2009. Mostly Harmless Econometrics: An Empiricist's
Companion.Princeton, NJ: Princeton
University Press.
3.
CainL. E., RobinsJ. M., LanoyE., LoganR., CostagliolaD., and HernánM. A.2010. When to start treatment? A systematic approach
to the comparison of dynamic regimes using observational
data. International Journal of
Biostatistics6(2): Article 18.
4.
D'AgostinoR. B., LeeM.-L., BelangerA. J., CupplesL. A., AndersonK., and KannelW. B.1990. Relation of pooled logistic regression to time
dependent Cox regression analysis: The Framingham heart
study. Statistics in Medicine9: 1501–1515.
5.
DanielR. M., CousensS. N., De StavolaB. L., KenwardM. G., and SterneJ. A. C.2011. Methods for dealing with time-varying
confounding. Unpublished manuscript.
6.
DanielR. M., De StavolaB. L., and
CousensS. N. Forthcoming.
Time-varying confounding: Some practical considerations in a
likelihood framework. InCausality: Statistical Perspectives and Applications, ed.
BerzuiniC., DawidA. P., and
BernardinelliL.Wiley.
7.
DidelezV., DawidA. P., and
GenelettiS.2006. Direct and indirect effects of sequential
treatments. InProceedings of the Twenty-Second Conference on Uncertainty in Artificial
Intelligence, 138–146.
Arlington, VA: AUAI
Press.
8.
FewellZ., HernanM. A., WolfeF., TillingK., ChoiH., and SterneJ. A. C.2004. Controlling for time-dependent confounding
using marginal structural models. Stata
Journal4: 402–420.
9.
GillR. D., and RobinsJ. M.2001. Causal inference for complex longitudinal
data: The continuous case. Annals of
Statistics29: 1785–1811.
10.
GreenlandS., PearlJ., and RobinsJ. M.1999. Causal diagrams for epidemiological
research. Epidemiology10: 37–48.
11.
HafemanD. M.2009. “Proportion explained”: A causal
interpretation for standard measures of indirect effect?American Journal of Epidemiology170: 1443–1448.
12.
HafemanD. M., and
VanderWeeleT. J.2011. Alternative assumptions for the identification
of direct and indirect effects.
Epidemiology22: 753–764.
13.
HernánM. A.,
Hernández-DíazS., and RobinsJ. M.2004. A structural approach to selection
bias. Epidemiology15: 615–625.
14.
HubbardA. E., and van der
LaanM. J.2008. Population intervention models in causal
inference. Biometrika95: 35–47.
15.
LechnerM.2001. Identification and estimation of causal
effects of multiple treatments under the conditional independence
assumption. InEconometric Evaluation of Labour Market Policies, ed.
LechnerM., and PfeifferF., 43–58.
Heidelberg:
Physica-Verlag.
16.
LechnerM.2004. Sequential matching estimation of dynamic causal
models. Discussion Paper No. 1042, Institute for
the Study of Labor (IZA). http://ftp.iza.org/dp1042.pdf.
17.
LechnerM., and MiquelR.2001. A potential outcome approach to dynamic programme
evaluation: Nonparametric identification. Discussion Paper No.
2001–07, Department of Economics, University of St.
Gallen.http://www.alexandria.unisg.ch/export/DL/33987.pdf.
18.
LittleR. J. A., and
RubinD. B.2002. Statistical Analysis with Missing Data. 2nd
ed. Hoboken, NJ:
Wiley.
19.
MiquelR.2002. Identification of dynamic treatments effects by
instrumental variables. Discussion Paper No. 2002–11,
University of St. Gallen.
MorganS. L., and
WinshipC.2007. Counterfactuals and Causal Inference: Methods and
Principles for Social Research.New York: Cambridge University
Press.
22.
MurphyS. A.2003. Optimal dynamic treatment
regimes. Journal of the Royal Statistical Society,
Series B65: 331–366.
23.
PearlJ.2001. Direct and indirect effects.
InProceedings of the Seventeenth Conference in Uncertainty in Artificial
Intelligence, 411–420.
San Francisco: Morgan
Kaufmann.
24.
PearlJ.2009. Causality: Models, Reasoning, and Inference.
2nd ed. New York: Cambridge
University Press.
25.
PetersenM. L., SinisiS. E., and van der
LaanM. J.2006. Estimation of direct causal
effects. Epidemiology17: 276–284.
26.
RobinsJ. M.1986. A new approach to causal inference in
mortality studies with a sustained exposure period—application to control of
the healthy worker survivor effect. Mathematical
Modelling7: 1393–1512.
27.
RobinsJ. M.1987a. Addendum to “A new approach to causal
inference in mortality studies with a sustained exposure period—application
to control of the healthy worker survivor effect”.
Computers and Mathematics with Applications14: 923–945.
28.
RobinsJ. M.1987b. Errata to “A new approach to causal inference
in mortality studies with a sustained exposure period—application to control
of the healthy worker survivor effect”. Computers
and Mathematics with Applications14: 917–921.
29.
RobinsJ. M.1989a. The analysis of randomized and non-randomized
AIDS treatment trials using a new approach to causal inference in
longitudinal studies. In Health Services Research
Methodology: A Focus on AIDS, ed. SechrestL., FreemanH., and MulleyA.Washington, DC: U.S. Public
Health Service.
30.
RobinsJ. M.1989b. Errata to “Addendum to ‘A new approach to
causal inference in mortality studies with a sustained exposure
period—application to control of the healthy worker survivor
effect’”. Computers and Mathematics with
Applications18: 477.
31.
RobinsJ. M.1997. Causal inference from complex longitudinal
data. In Latent Variable Modeling and Applications
to Causality, ed. BerkaneM., 69–117.
New York:
Springer.
32.
RobinsJ. M.1999. Association, causation, and marginal
structural models. Synthese121: 151–179.
33.
RobinsJ. M.2003. Semantics of causal DAG models and the
identification of direct and indirect effects. In
Highly Structured Stochastic Systems, ed.
GreenP., HjortN. L., and
RichardsonS., 70–81.
New York: Oxford University
Press.
34.
RobinsJ. M., BlevinsD., RitterG., and WulfsohnM.1992. G-estimation of the effect of prophylaxis
therapy for Pneumocystis carinii pneumonia on the survival of AIDS
patients. Epidemiology3: 319–336.
35.
RobinsJ. M., and GillR. D.1997. Non-response models for the analysis of
nonmonotone ignorable missing data. Statistics in
Medicine16: 39–56.
36.
RobinsJ. M., and
GreenlandS.1992. Identifiability and exchangeability for direct
and indirect effects. Epidemiology3: 143–155.
37.
RobinsJ. M., GreenlandS., and HuF.-C.1999. Estimation of the causal effect of a
time-varying exposure on the marginal mean of a repeated binary
outcome. Journal of the American Statistical
Association94: 687–700.
38.
RobinsJ. M., and HernánM. A.2009. Estimation of the causal effects of
time-varying exposures. In Longitudinal Data
Analysis, ed. FitzmauriceG., DavidianM., VerbekeG., and
MolenberghsG., 553–599.
Boca Raton, FL: Chapman &
Hall/CRC.
39.
RobinsJ. M., HernánM. A., and
BrumbackB.2000. Marginal structural models and causal
inference in epidemiology. Epidemiology11: 550–560.
40.
RobinsJ. M., and
RichardsonT. S.2010. Alternative graphical causal models and the
identification of direct effects. InCausality and Psychopathology: Finding the Determinants of Disorders and
Their Cures, ed. ShroutP.Oxford: Oxford University
Press.
41.
RothmanK. J., GreenlandS., and LashT. L.2008. Modern Epidemiology. 3rd ed.
Philadelphia: Lippincott
Williams & Wilkins.
42.
SterneJ. A. C., and
TillingK.2002. G-estimation of causal effects, allowing for
time-varying confounding. Stata Journal2: 164–182.
43.
TaubmanS. L., RobinsJ. M., MittlemanM. A., and HernánM. A.2009. Intervening on risk factors for coronary heart
disease: An application of the parametric g-formula.
International Journal of Epidemiology38: 1599–1611.
44.
TsiatisA. A.2006. Semiparametric Theory and Missing Data.New York:
Springer.
45.
van BuurenS., BoshuizenH. C., and KnookD. L.1999. Multiple imputation of missing blood pressure
covariates in survival analysis. Statistics in
Medicine18: 681–694.
