Purpose We review and compare two causal modeling approaches that correspond to two major and distinct classes of inference – efficacy and interventionbased inference – in the context of randomized trials with subject noncompliance. Methods We review the definitions of efficacy and intervention-based effects in the clinical trials literature and relate these to two separate and distinct causal modeling approaches: the structural mean modeling (SMM) approach and the principal stratification, instrumental variable approach.
Results The SMM-based efficacy approach focuses on the effect of actually receiving treatment. In contrast, the principal stratification method addresses the effect of treatment assignment within partially unobserved latent subgroups defined by compliance behavior. While these approaches differ in terms of philosophy, model definitions, and estimation, they estimate the same causal effect under certain assumptions, but estimate very different causal effects when those assumptions are relaxed. We illustrate these results using a randomized psychiatry trial where the focus is physician compliance to the designated protocol and the other examines patient compliance to the designated protocol, both from the same trial.
Limitations The validity of the models under the instrumental variable, SMM and principal stratification approaches depends on modeling assumptions, some of which may not be verifiable from the observed data and potentially less realistic than the no-confounding assumption made by non-causal approaches.
Conclusions This comparison in terms of efficacy versus intervention-based effects in causal modeling parallels the explanatory versus pragmatic approaches in clinical trials research; therefore researchers should weigh carefully when choosing causal modeling methodology based on whether efficacy or intervention-based effects are of interest. Clinical Trials 2007; 4: 58–73; http://ctj.sagepub.com
Get full access to this article
View all access options for this article.
References
1.
Sommer A, Zeger S. On estimating efficacy from clinical trials. Statistics in Medicine1991; 10: 45–52.
2.
Sheiner LB, Rubin DB. Intentions-to-treat analysis and the goal of clinical trials. Clinical Pharmacology Therapy1995; 56: 6–10
3.
Robins JM.Correcting for non-compliance in randomized trials using structural nested mean models. Communications in Statistics, Theory and Methods1923; 23: 2379–2412.
4.
Fischer-Lapp K, Goetghebeur E. Practical properties of some structural mean analyses of the effect of compliance in randomized trial. Randomized Controlled Trial Controlled Clinical Trials1999; 20: 531–546.
5.
Angrist JD, Imbens GW, Rubin DB. Identification of causal effects using instrumental variables. Journal of the American Statistical Association1996; 91: 444–455.
6.
Hirano K, Imbens GW, Rubin DB, Zhou XH. Assessing the effect of an influenza vaccine in an encouragement design. Biostatistics2000; 1: 69–88.
7.
Frangakis CE, Rubin DB. Principal stratification in causal inference. Biometrics2002; 58: 21–29.
8.
Frangakis CE, Brookmeyer RS, Varadhan R, Safaeian M, Vlahov D, Strathdee SA. Methodology for evaluating a partially controlled longitudinal treatment using principal stratification, with application to a Needle Exchange Program. Journal of the American Statistical Association2004; 97: 284–292.
9.
Robins JM, Greenland S. Comment on “Identification of causal effects using instrumental variables”.Journal of the American Statistical Association1996; 91: 456–458.
10.
Frangakis CE, Rubin DB. Addressing complications of intent-to-treat analysis in the combined presence of allor-none treatment non-compliance and subsequent missing outcomes. Biometrika1999; 86: 365–379.
11.
Neyman J.On the application of probability theory to agricultural experiments: essay on principles, section 9, translated in Statistical Science1923; 5: 165–480, 1990.
12.
Mulsant BH, Ganbuli M. Epidemiology and diagnosis of depression late in life. Journal of Clinical Psychiatry1999; 60(Suppl 20): 9–15.
13.
Ten Have TR, Elliott MR, Joffe M, Zanutto E, Datto C. Causal models for randomized physician encouragement trials in treating in primary care depression. Journal of the American Statistical Association2004; 99: 16–25.
14.
Robins JM, Rotnitzky A. Estimation of treatment effects in randomised trials with non-compliance and a dichotomous outcome using structural mean models, Biometrika2004; 91: 763–783.
15.
Vansteelandt S, Goetghebeur E. Sense and sensitivity when correcting for observed exposures in randomized clinical trials. Statistics in Medicine2005; 24: 191–210.
16.
Ten Have TR, Joffe MM, Cary M. Causal logistic models for non-compliance under randomized treatment with univariate binary response, Statistics in Medicine2003; 22: 1255–1284.
17.
Rubin DB.Bayesian inference for causal effects: the role of randomization. The Annals of Statistics1978; 6: 34–58.
18.
Rubin DB.Statistics and causal inference. Comment: which ifs have causal answers. Journal of the American Statistical Association1986; 81: 961–962.
19.
Robins JM, Greenland S. Adjusting for differential rates of PCP prophylaxis in high-versus low-dose AZT treatment arms. Journal of the American Statistical Association1994; 89: 737–749.
20.
Imbens GW, Rubin DB. Bayesian inference for causal effects in randomized experiments with noncompliance. The Annals of Statistics1997; 25: 305–327.
21.
Imbens GW, Rubin DB. Estimating outcome distributions for compliers in instrumental variable models. The Review of Economic Studies1997; 64: 555–574.
22.
Jo B, Muthén B. Modeling of intervention effects with noncompliance: A latent variable modeling approach for randomized trials. In Marcoulides GA and Schumacker RE (eds). New developments and techniques in structural equation modeling. Lawrence Erlbaum Associates, 2001; 57–87.
23.
Nagelkerke N, Fidler V, Bernsen R, Borgdorff M. Estimating treatment effects in randomized clinical trials in the presence of non-compliance. Statistics in Medicine2000; 19: 1849–1864.
24.
Stock J.Instrumental variables in statistics and economics. In Baltes P and Smelser N (eds). Elsevier Science. International encyclopedia of the social & behavioral sciences.
25.
Last JM, Spasoff RA, Harris SS, Thuriaux MC (eds) A dictionary of epidemiology. 4th edition. Oxford University Press, 2000.
26.
Schwartz D, Pejovic M. The explanatory and pragmatic approaches in clinical trials. In Steichele C, Abshagen U and Koch-Wester J (eds). Drugs between research and regulations. Springer-Verglag, 1985; 113–120.
27.
Joffe M, Brensinger C. Weighting in instrumental variables and g-estimation. Statistics in Medicine2005; 22: 1285–1303.
28.
Small D, Ten Have TR, Cheng J. Random effects logistic models for analyzing efficacy of a longitudinal randomized treatment with non-adherence. Statistics in Medicine, 2006; 25: 1981–2007.
29.
Little RJ, Rubin DB. Causal effects in clinical and epidemiological studies via potential outcomes: concept and analytical approaches. Annual Reviews of Public Health2000; 21: 121–145.
30.
Angrist J, Imbens GW. Identification and two-stage least squares estimation of average causal effects in models with variable treatment intensityJournal of the American Statistical Association1995; 90: 431–442.
31.
Morales KH, Ten Have TR, Joffe M, Elliot M, Heo M, Bogner H, Mulsant B. Causal linear models for noncompliance under randomized encouragement with univariate continuous response. Unpublished manuscript, University of Pennsylvania, Department of Biostatistics and Epidemiology, 2004.
