Skip to main content

Causal mediation analyses for randomized trials

Health Services and Outcomes Research Methodology volume 8, pages 57–76 (2008)Cite this article

Abstract

In the context of randomized intervention trials, we describe causal methods for analyzing how post-randomization factors constitute the process through which randomized baseline interventions act on outcomes. Traditionally, such mediation analyses have been undertaken with great caution, because they assume that the mediating factor is also randomly assigned to individuals in addition to the randomized baseline intervention (i.e., sequential ignorability). Because the mediating factors are typically not randomized, such analyses are unprotected from unmeasured confounders that may lead to biased inference. We review several causal approaches that attempt to reduce such bias without assuming that the mediating factor is randomized. However, these causal approaches require certain interaction assumptions that may be assessed if there is enough treatment heterogeneity with respect to the mediator. We describe available estimation procedures in the context of several examples from the literature and provide resources for software code.

This is a preview of subscription content, access via your institution.

Access options

Buy single article

Instant access to the full article PDF.

43,34 €

Price includes VAT (Finland)
Tax calculation will be finalised during checkout.

Fig. 1
Fig. 2
Fig. 3

References

  • Angrist, J., Imbens, G., Rubin, D.: Identification of causal effects using instrumental variables. J. Am. Stat. Assoc. 91, 444–455 (1996)

    Article  Google Scholar 

  • Baranowski, T.: Theory as mediating variables: why aren’t community interventions working as desired? Ann. Epidemiol. S7, S89–S95 (1997)

    Article  Google Scholar 

  • Baron, R.M., Kenny, D.A.: The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations. J. Pers. Soc. Psychol. 51, 1173–1182 (1986)

    PubMed  Article  CAS  Google Scholar 

  • Brown, G., Ten Have, T., Henriques, G., Xie, S.X., Hollander, J.E., Beck, A.T.: Cognitive therapy for the prevention of suicide attempts: a randomized controlled trial. J. Am. Med. Assoc. 294, 2847–2848 (2005)

    Google Scholar 

  • Bruce, M., Ten Have, T., Reynolds, C., et al.: A randomized trial to reduce suicidal ideation and depressive symptoms in depressed older primary care patients: the PROSPECT study. J. Am. Med. Assoc. 291, 1081–1091 (2004)

    Article  CAS  Google Scholar 

  • Cole, D., Maxwell, S.: Testing mediational models with longitudinal data: questions and tips in the use of structural equation modeling. J. Abnorm. Psychol. 112, 558–577 (2003)

    PubMed  Article  Google Scholar 

  • Frangakis, C., Brookmeyer, R., Varadhan, R., Safaeian, M., Vlahov, D., Strathdee, S.: Methodology for evaluating a partially controlled longitudinal treatment using principal stratification, with application to a needle exchange program. J. Am. Stat. Assoc. 97, 284–292 (2004)

    Google Scholar 

  • Frangakis, C., Rubin, D.: Principal stratification in causal inference. Biometrics 58, 21–29 (2002)

    PubMed  Article  Google Scholar 

  • Frangakis, C., Rubin, D., Zhao, X.-H.: Clustered encouragement designs with individual noncompliance: Bayesian inference with randomization, and application to advance directive forms. Biostatistics 3, 147–164 (2002)

    PubMed  Article  Google Scholar 

  • Gallop, R., Small, D., Ten Have, T.: Mediation analyses with principal stratification models. Submitted (2007)

  • Gollob, H.F., Reichardt, C.S.: Taking account of time lags in causal models. Child Dev. 58, 80–92 (1987)

    PubMed  Article  CAS  Google Scholar 

  • Gollob, H.F., Reichardt, C.S.: Interpreting and estimating indirect effects assuming time lags really matter. In: Collins, L.M., Horn, J.L. (eds.) Best Methods for the Analysis of Change: Recent Advances, Unanswered Questions, Future Directions, pp 243–259. American Psychological Association (1991)

  • Hirano, K., Imbens, G., Rubin, D., Zhou, X.: Assessing the effect of an influenza vaccine in an encouragement design. Biostatistics 1, 69–88 (2000)

    PubMed  Article  Google Scholar 

  • Holland, P.: Statistics and causal inference. J. Am. Stat. Assoc. 81, 945–960 (1986)

    Article  Google Scholar 

  • Imbens, G., Rubin, D.: Bayesian inference for causal effects in randomized experiments with noncompliance. Ann. Stat. 25, 305–327 (1997)

    Article  Google Scholar 

  • Joffe, M., Hoover, D., Jacobson, L., Kingsley, L., Chmiel, J., Fischer, B., Robins, J.: Estimating the effect of Ziduvodine on Kaposi’s sarcoma from observational data using a rank preserving failure time model. Stat. Med. 17, 1073–1102 (1998)

    PubMed  Article  CAS  Google Scholar 

  • Joffe, M., Small, D., Hsu, C.: Defining and estimating intervention effects for groups who will develop an auxiliary outcome. Stat. Sci. 22, 74–97 (2007)

    Article  Google Scholar 

  • Judd, C.M., Kenny, D.A.: Process analysis: Estimating mediation in treatment evaluations. Eval. Rev. 5, 602–619 (1981)

    Article  Google Scholar 

  • Kazdin, A.: Mediators and mechanisms of change in psychotherapy research. Ann. Rev. Clin. Psychol. 3, 1–27 (2007)

    Article  Google Scholar 

  • Kenny, D., Korchmaros, J., Bolger, N.: Lower level mediation in multi-level models. Psychol. Methods 8, 115–128 (2003)

    PubMed  Article  Google Scholar 

  • Kraemer, H., Stice, E., Kazdin, A., Offord, D., Kupfer, D.: How do risk factors work together? Mediators, moderators, and independent, overlapping, and proxy risk factors. Am. J. Psychiat. 158, 848–856 (2001)

    PubMed  CAS  Google Scholar 

  • Kraemer, H., Wilson, G., Fairburn, C.: Mediators and moderators of treatment effects in randomized clinical trials. Arch. Gen. Psychiatry 59, 877–883 (2002)

    PubMed  Article  Google Scholar 

  • Krull, J.L., MacKinnon, D.P.: Multilevel mediation modeling in group-based intervention studies. Eval. Rev. 23, 418–444 (1999)

    PubMed  CAS  Google Scholar 

  • Krull, J.L., MacKinnon, D.P.: Multilevel modeling of individual and group level mediated effects. Multivariate Behav. Res. 36, 249–277 (2001)

    Article  Google Scholar 

  • MacKinnon, D.P., Dwyer, J.H.: Estimating mediated effects in prevention studies. Eval. Rev. 17, 144–158 (1993)

    Article  Google Scholar 

  • MacKinnon, D.P., Lockwood, C.M., Hoffman, J.M., West, S.G., Sheets, V.: A comparison of methods to test mediation and other intervening variable effects. Psychol. Methods 7, 83–104 (2002)

    PubMed  Article  Google Scholar 

  • Mavandadi, S., Ten Have, T., Katz, I., Nalla, U., Durai, B., Krahn, D., Llorente, M., Kirchner, J., Olsen, E., Van Stone, W., Cooley, S., Oslin, D.: The effect of depression treatment on depressive symptoms in older adulthood: the moderating role of pain. J. Am. Geriatr. Soc. 55, 202–211 (2007)

    PubMed  Article  Google Scholar 

  • Mealli, F., Imbens, G., Ferro, S., Biggeri, A.: Analyzing a randomized trial on breast self-examination with noncompliance and missing outcomes. Biostatistics 5, 207–222 (2004)

    PubMed  Article  Google Scholar 

  • Mealli, F., Rubin, D.: Commentary: assumptions allowing the estimation of direct causal effects. J. Econom. 112, 79–87 (2003)

    Article  Google Scholar 

  • Neyman, J.: On the application of probability theory to agricultural experiments. Essay on principles. Stat. Sci. 5, 465–472 (1923), Translated by D.M. Dabrowska and edited by T.P. Speed (1990)

  • Pearl, J.: Causality. Cambridge University Press, Cambridge UK (1999)

    Google Scholar 

  • Pearl, J.: Direct and indirect effects. In: Besnard, P., Hanks, S. (eds.) Proceedings of the Seventeenth Conference on Uncertainty in Artificial Intelligence, pp. 411–420. Morgan Kaufmann, San Francisco (2001)

  • Prentice, R., Langer, M., Stefanick, B., et al.: Combined postmenopausal hormone therapy and cardiovascular disease: toward resolving the discrepancy between observational studies and the women’s health initiative clinical trial. Am. J. Epidemiol. 162, 404–414 (2005)

    Google Scholar 

  • Robins, J.: Correcting for non-compliance in randomized trials using structural nested mean models. Commun. Stat. Theory Methods 23, 2379–2412 (1994)

    Article  Google Scholar 

  • Robins, J.: Testing and estimation of direct effects by reparameterizing directed acyclic graphs with structural nested models. In: Glymour, C., Cooper, G. (eds.) Computation, Causation, and Discovery, pp 349–405. AAAI Press/The MIT Press, Menlo Park CA/Cambridge MA (1999)

    Google Scholar 

  • Robins, J.: Semantics of causal DAG models and the identification of direct and indirect effects. In: Green, P., Hjort, N., Richardson, S. (eds.) In Highly Structured Stochastic Systems, pp 70–81. Oxford University Press, New York (2003)

    Google Scholar 

  • Robins, J., Blevins, D., Ritter, G., Wulfsohn, M.: G-estimation of the effect of prophylaxis therapy for pneumocystis carinii pneumonia on the survival of AIDS patients. Epidemiology 3, 319–336 (1992)

    PubMed  Article  CAS  Google Scholar 

  • Robins, J., Rotnitzky, A.: Estimation of treatment effects in randomised trials with non-compliance and a dichotomous outcome using structural mean models. Biometrika 91, 763–783 (2005)

    Article  Google Scholar 

  • Rubin, D.: Estimating causal effects of treatment in randomized and nonrandomized studies. J. Educ. Psychol. 66, 688–701 (1974)

    Article  Google Scholar 

  • Rubin, D.: Statistics rs. J. Am. Stat. Assoc. 81, 961–962 (1986)

    Article  Google Scholar 

  • Rubin, D.: Direct and indirect causal effects via potential outcomes. Scand. J. Stat. 31, 161–170 (2004)

    Article  Google Scholar 

  • Ten Have, T., Joffe, M., Cary, M.: Causal logistic models for non-compliance under randomized treatment with univariate binary response. Stat. Med. 22, 1255–1284 (2003)

    Google Scholar 

  • Ten Have, T., Elliott, M., Joffe, M., Zanutto, E., Datto, C.: Causal models for randomized physician encouragement trials in treating primary care depression. J. Am. Stat. Assoc. 99, 8–16 (2004)

    Google Scholar 

  • Ten Have, T., Joffe, M., Lynch, K., Maisto, S., Brown, G., Beck, A.: Causal mediation analyses with rank preserving models. Biometrics 63, 926–934 (2007)

    Article  Google Scholar 

  • Vansteelandt, S., Goetghebeur, E.: Causal inference with generalized structural mean models. J. Roy. Stat. Soc. B 65, 817–835 (2003)

    Article  Google Scholar 

  • Vansteelandt, S., Goetghebeur, E.: Using potential outcomes as predictors of treatment activity via strong structural mean models. Stat. Sinica 14, 907–925 (2004)

    Google Scholar 

Download references

Acknowledgments

The authors thank Marshall Joffe, Dylan Small, Michael Elliott, Knashawn Morales, Joseph Gallo, and two reviewers for very insightful comments that improved the paper tremendously. Funding was provided by NIMH grants: R01-MH61892, R01-MH59380, R01-CA095415, P30-MH066270, R01-MH60915, P20-MH71905, and R37-CCR316866.

Author information

Affiliations

  1. Department of Psychiatry, University of Pennsylvania School of Medicine, Philadelphia, PA, 19104, USA

    Kevin G. Lynch

  2. Department of Biostatistics and Epidemiology, University of Pennsylvania School of Medicine, Philadelphia, PA, 19104, USA

    Mark Cary, Robert Gallop & Thomas R. Ten Have

Authors
  1. Kevin G. Lynch
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Mark Cary
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Robert Gallop
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Thomas R. Ten Have
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Thomas R. Ten Have.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Lynch, K.G., Cary, M., Gallop, R. et al. Causal mediation analyses for randomized trials. Health Serv Outcomes Res Method 8, 57–76 (2008). https://doi.org/10.1007/s10742-008-0028-9

Download citation

  • Received:

  • Revised:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s10742-008-0028-9

Keywords

  • Structural mean models
  • Principal stratification
  • Direct effects
  • Unmeasured confounding
  • Baseline randomization
  • Sequential ignorability