Effects of multiple interventions

The purpose of this chapter is (i) to describe some currently available analytical methods for using individual level epidemiological data to estimate the impact of multiple risk factor interventions on health and (ii) to carefully review the conditions under which these methods deliver unbiased estimates of impact. The chapter is organized as follows. In sections 2 and 3, we discuss estimation of effects of short-term, timeindependent interventions. Specifically, we discuss estimating the effect of a single risk factor intervention on life expectancy or quality-adjusted life expectancy over a specified period of follow-up in a single population, when essentially ideal epidemiologic data are available. That is, we assume a random sample of the population is randomly assigned to different levels of the risk factor and followed prospectively for a fixed time period. Second, we consider the same study design, except now we are interested in the joint effect of an intervention on several risk factors. Third, we consider the problem of extrapolation of the results to longer periods of follow-up and to other populations for which no primary epidemiologic data are available. Sections 2 and 3 serve to indicate the possibilities and limitations of even ideal epidemiologic data for estimating the effects of multiple time-independent risk factor interventions. In sections 4 and 5 we turn to the main topic of this chapter: the estimation of the effect of multiple time-dependent interventions from observational data, possibly plagued by confounding, selection bias, measurement error, information bias and ill-defined interventions. In sections 6 to 8, we illustrate our methods by attempting to estimate the effects of various time-varying interventions on subjects entered in the Framingham Offspring cohort study. Finally, in section 9 we offer some conclusions. Chapter 28

[1]  D. Rubin.  [On the Application of Probability Theory to Agricultural Experiments. Essay on Principles. Section 9.] Comment: Neyman (1923) and Causal Inference in Experiments and Observational Studies , 1990 .

[2]  S. Zeger,et al.  Longitudinal data analysis using generalized linear models , 1986 .

[3]  J. Pearl,et al.  Causal diagrams for epidemiologic research. , 1999, Epidemiology.

[4]  J. Robins,et al.  Sensitivity Analysis for Selection bias and unmeasured Confounding in missing Data and Causal inference models , 2000 .

[5]  J. Robins.  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” , 1987 .

[6]  J. Robins,et al.  Marginal Structural Models to Estimate the Joint Causal Effect of Nonrandomized Treatments , 2001 .

[7]  J. Robins,et al.  G-Estimation of the Effect of Prophylaxis Therapy for Pneumocystis carinii Pneumonia on the Survival of AIDS Patients , 1992, Epidemiology.

[8]  Tom Burr,et al.  Causation, Prediction, and Search , 2003, Technometrics.

[9]  J. Robins.  A new approach to causal inference in mortality studies with a sustained exposure period—application to control of the healthy worker survivor effect , 1986 .

[10]  J. Robins,et al.  Estimation of the Causal Effect of a Time-Varying Exposure on the Marginal Mean of a Repeated Binary Outcome , 1999 .

[11]  J. Robins,et al.  Marginal structural models to estimate the causal effect of zidovudine on the survival of HIV-positive men. , 2000, Epidemiology.

[12]  K G Manton,et al.  Projecting the future size and health status of the U.S. elderly population. , 1992, International journal of forecasting.

[13]  T. Speed,et al.  On the Application of Probability Theory to Agricultural Experiments. Essay on Principles. Section 9 , 1990 .

[14]  James M. Robins,et al.  Marginal Structural Models versus Structural nested Models as Tools for Causal inference , 2000 .

[15]  Donald B. Rubin,et al.  Bayesian Inference for Causal Effects: The Role of Randomization , 1978 .