Sensitivity analysis for interactions under unmeasured confounding

We develop a sensitivity analysis technique to assess the sensitivity of interaction analyses to unmeasured confounding. We give bias formulas for sensitivity analysis for interaction under unmeasured confounding on both additive and multiplicative scales. We provide simplified formulas in the case in which either one of the two factors does not interact with the unmeasured confounder in its effects on the outcome. An interesting consequence of the results is that if the two exposures of interest are independent (e.g., gene-environment independence), even under unmeasured confounding, if the estimate of the interaction is nonzero, then either there is a true interaction between the two factors or there is an interaction between one of the factors and the unmeasured confounder; an interaction must be present in either scenario. We apply the results to two examples drawn from the literature.

[1]  Sander Greenland,et al.  Modern Epidemiology 3rd edition , 1986 .

[2]  Tyler J. VanderWeele,et al.  Empirical tests for compositional epistasis , 2010, Nature Reviews Genetics.

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

[4]  S Lemeshow,et al.  Confidence interval estimation of interaction. , 1992, Epidemiology.

[5]  Tyler J VanderWeele,et al.  Sufficient Cause Interactions and Statistical Interactions , 2009, Epidemiology.

[6]  R. Kronmal,et al.  Assessing the sensitivity of regression results to unmeasured confounders in observational studies. , 1998, Biometrics.

[7]  N. Day,et al.  Synergism and interaction: are they equivalent? , 1979, American journal of epidemiology.

[8]  D. Hosmer,et al.  Confidence Intervals for Measures of Interaction , 1996, Epidemiology.

[9]  E. C. Hammond,et al.  Smoking and lung cancer: recent evidence and a discussion of some questions. , 1959, Journal of the National Cancer Institute.

[10]  E. Bowman,et al.  Environmental tobacco smoke, genetic susceptibility, and risk of lung cancer in never-smoking women. , 1999, Journal of the National Cancer Institute.

[11]  James M Robins,et al.  The Identification of Synergism in the Sufficient-Component-Cause Framework , 2007, Epidemiology.

[12]  D. Rubin,et al.  Assessing Sensitivity to an Unobserved Binary Covariate in an Observational Study with Binary Outcome , 1983 .

[13]  S Greenland,et al.  Tests for interaction in epidemiologic studies: a review and a study of power. , 1983, Statistics in medicine.

[14]  W D Flanders,et al.  Nontraditional epidemiologic approaches in the analysis of gene-environment interaction: case-control studies with no controls! , 1996, American journal of epidemiology.

[15]  T. VanderWeele.  On the Distinction Between Interaction and Effect Modification , 2009, Epidemiology.

[16]  S Greenland,et al.  Concepts of interaction. , 1980, American journal of epidemiology.

[17]  Jack A. Taylor,et al.  Non-hierarchical logistic models and case-only designs for assessing susceptibility in population-based case-control studies. , 1994, Statistics in medicine.

[18]  Terence P. Speed,et al.  Introductory Remarks on Neyman (1923) , 1990 .

[19]  T. VanderWeele,et al.  Interpretation of Subgroup Analyses in Randomized Trials: Heterogeneity Versus Secondary Interventions , 2011, Annals of Internal Medicine.

[20]  Onyebuchi A Arah,et al.  Bias Formulas for Sensitivity Analysis of Unmeasured Confounding for General Outcomes, Treatments, and Confounders , 2011, Epidemiology.

[21]  R. Saracci.  Interaction and synergism. , 1980, American journal of epidemiology.

[22]  D. Rubin.  Estimating causal effects of treatments in randomized and nonrandomized studies. , 1974 .

[23]  W. Flanders,et al.  Indirect Assessment of Confounding: Graphic Description and Limits on Effect of Adjusting for Covariates , 1990, Epidemiology.

[24]  Zhongqi Cheng,et al.  Arsenic exposure from drinking water and risk of premalignant skin lesions in Bangladesh: baseline results from the Health Effects of Arsenic Longitudinal Study. , 2006, American journal of epidemiology.