Dataset Information

A Semiparametric Approach to Model-Based Sensitivity Analysis in Observational Studies.

ABSTRACT: When drawing causal inference from observational data, there is almost always concern about unmeasured confounding. One way to tackle this is to conduct a sensitivity analysis. One widely-used sensitivity analysis framework hypothesizes the existence of a scalar unmeasured confounder U and asks how the causal conclusion would change were U measured and included in the primary analysis. Work along this line often makes various parametric assumptions on U, for the sake of mathematical and computational convenience. In this article, we further this line of research by developing a valid sensitivity analysis that leaves the distribution of U unrestricted. Compared to many existing methods in the literature, our method allows for a larger and more flexible family of models, mitigates observable implications (Franks et al., 2019), and works seamlessly with any primary analysis that models the outcome regression parametrically. We construct both pointwise confidence intervals and confidence bands that are uniformly valid over a given sensitivity parameter space, thus formally accounting for unknown sensitivity parameters. We apply our proposed method on an influential yet controversial study of the causal relationship between war experiences and political activeness using observational data from Uganda.

SUBMITTER: Zhang B

PROVIDER: S-EPMC9916707 | biostudies-literature | 2022 Dec

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

A Semiparametric Approach to Model-Based Sensitivity Analysis in Observational Studies.

Zhang Bo B Tchetgen Tchetgen Eric J EJ

Journal of the Royal Statistical Society. Series A, (Statistics in Society) 20221124 Suppl 2

When drawing causal inference from observational data, there is almost always concern about unmeasured confounding. One way to tackle this is to conduct a sensitivity analysis. One widely-used sensitivity analysis framework hypothesizes the existence of a scalar unmeasured confounder U and asks how the causal conclusion would change were U measured and included in the primary analysis. Work along this line often makes various parametric assumptions on U, for the sake of mathematical and computat ...[more]

PMID: 36777968

Similar Datasets

Project description:The two-phase sampling design is a cost-efficient way of collecting expensive covariate information on a judiciously selected sub-sample. It is natural to apply such a strategy for collecting genetic data in a sub-sample enriched for exposure to environmental factors for gene-environment interaction (G × E) analysis. In this paper, we consider two-phase studies of G × E interaction where phase I data are available on exposure, covariates and disease status. Stratified sampling is done to prioritize individuals for genotyping at phase II conditional on disease and exposure. We consider a Bayesian analysis based on the joint retrospective likelihood of phase I and phase II data. We address several important statistical issues: (i) we consider a model with multiple genes, environmental factors and their pairwise interactions. We employ a Bayesian variable selection algorithm to reduce the dimensionality of this potentially high-dimensional model; (ii) we use the assumption of gene-gene and gene-environment independence to trade-off between bias and efficiency for estimating the interaction parameters through use of hierarchical priors reflecting this assumption; (iii) we posit a flexible model for the joint distribution of the phase I categorical variables using the non-parametric Bayes construction of Dunson and Xing (2009). We carry out a small-scale simulation study to compare the proposed Bayesian method with weighted likelihood and pseudo likelihood methods that are standard choices for analyzing two-phase data. The motivating example originates from an ongoing case-control study of colorectal cancer, where the goal is to explore the interaction between the use of statins (a drug used for lowering lipid levels) and 294 genetic markers in the lipid metabolism/cholesterol synthesis pathway. The sub-sample of cases and controls on which these genetic markers were measured is enriched in terms of statin users. The example and simulation results illustrate that the proposed Bayesian approach has a number of advantages for characterizing joint effects of genotype and exposure over existing alternatives and makes efficient use of all available data in both phases.

Project description:ObjectivesIn observational studies, epidemiologists often attempt to estimate the total effect of an exposure on an outcome of interest. However, when the underlying diagram is unknown and limited knowledge is available, dissecting bias performances is essential to estimating the total effect of an exposure on an outcome when mistakenly adjusting for mediators under logistic regression. Through simulation, we focused on six causal diagrams concerning different roles of mediators. Sensitivity analysis was conducted to assess the bias performances of varying across exposure-mediator effects and mediator-outcome effects when adjusting for the mediator.SettingBased on the causal relationships in the real world, we compared the biases of varying across the effects of exposure-mediator with those of varying across the effects of mediator-outcome when adjusting for the mediator. The magnitude of the bias was defined by the difference between the estimated effect (using logistic regression) and the total effect of the exposure on the outcome.ResultsIn four scenarios (a single mediator, two series mediators, two independent parallel mediators or two correlated parallel mediators), the biases of varying across the effects of exposure-mediator were greater than those of varying across the effects of mediator-outcome when adjusting for the mediator. In contrast, in two other scenarios (a single mediator or two independent parallel mediators in the presence of unobserved confounders), the biases of varying across the effects of exposure-mediator were less than those of varying across the effects of mediator-outcome when adjusting for the mediator.ConclusionsThe biases were more sensitive to the variation of effects of exposure-mediator than the effects of mediator-outcome when adjusting for the mediator in the absence of unobserved confounders, while the biases were more sensitive to the variation of effects of mediator-outcome than those of exposure-mediator in the presence of an unobserved confounder.

Dataset Information

A Semiparametric Approach to Model-Based Sensitivity Analysis in Observational Studies.

Publications

A Semiparametric Approach to Model-Based Sensitivity Analysis in Observational Studies.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets