Bayesian sensitivity analysis for unmeasured confounding in observational studies

2007 ◽  
Vol 26 (11) ◽  
pp. 2331-2347 ◽  
Author(s):  
Lawrence C. McCandless ◽  
Paul Gustafson ◽  
Adrian Levy
Keyword(s):  
Sensitivity Analysis ◽  
Observational Studies ◽  
Unmeasured Confounding

Sensitivity Analysis for Unmeasured Confounding: E-Values for Observational Studies

2017 ◽  
Vol 167 (4) ◽  
pp. 285 ◽  
Author(s):  
A. Russell Localio ◽  
Catherine B. Stack ◽  
Michael E. Griswold
Keyword(s):  
Sensitivity Analysis ◽  
Observational Studies ◽  
Unmeasured Confounding

scholarly journals A note on a sensitivity analysis for unmeasured confounding, and the related E-value

2020 ◽  
Vol 8 (1) ◽  
pp. 229-248
Author(s):  
Arvid Sjölander
Keyword(s):  
Sensitivity Analysis ◽  
Observational Studies ◽  
Real Data ◽  
Feasible Region ◽  
Maximal Strength ◽  
Unmeasured Confounding ◽  
Unmeasured Confounder ◽  
Two Parameters ◽  
Strength Of Association

Abstract Unmeasured confounding is one of the most important threats to the validity of observational studies. In this paper we scrutinize a recently proposed sensitivity analysis for unmeasured confounding. The analysis requires specification of two parameters, loosely defined as the maximal strength of association that an unmeasured confounder may have with the exposure and with the outcome, respectively. The E-value is defined as the strength of association that the confounder must have with the exposure and the outcome, to fully explain away an observed exposure-outcome association. We derive the feasible region of the sensitivity analysis parameters, and we show that the bounds produced by the sensitivity analysis are not always sharp. We finally establish a region in which the bounds are guaranteed to be sharp, and we discuss the implications of this sharp region for the interpretation of the E-value. We illustrate the theory with a real data example and a simulation.

Bayesian sensitivity analysis for unmeasured confounding in causal mediation analysis

2017 ◽  
Vol 28 (2) ◽  
pp. 515-531 ◽  
Author(s):  
Lawrence C McCandless ◽  
Julian M Somers
Keyword(s):  
Sensitivity Analysis ◽  
Mediation Analysis ◽  
Indirect Effects ◽  
Direct And Indirect Effects ◽  
Unmeasured Confounding ◽  
Intermediate Variable ◽  
Causal Mediation ◽  
Important Concern ◽  
Analysis Technique ◽  
Causal Mediation Analysis

Causal mediation analysis techniques enable investigators to examine whether the effect of the exposure on an outcome is mediated by some intermediate variable. Motivated by a data example from epidemiology, we consider estimation of natural direct and indirect effects on a survival outcome. An important concern is bias from confounders that may be unmeasured. Estimating natural direct and indirect effects requires an elaborate series of assumptions in order to identify the target quantities. The analyst must carefully measure and adjust for important predictors of the exposure, mediator and outcome. Omitting important confounders may bias the results in a way that is difficult to predict. In recent years, several methods have been proposed to explore sensitivity to unmeasured confounding in mediation analysis. However, many of these methods limit complexity by relying on a handful of sensitivity parameters that are difficult to interpret, or alternatively, by assuming that specific patterns of unmeasured confounding are absent. Instead, we propose a simple Bayesian sensitivity analysis technique that is indexed by four bias parameters. Our method has the unique advantage that it is able to simultaneously assess unmeasured confounding in the mediator–outcome, exposure–outcome and exposure–mediator relationships. It is a natural Bayesian extension of the sensitivity analysis methodologies of VanderWeele, which have been widely used in the epidemiology literature. We present simulation findings, and additionally, we illustrate the method in an epidemiological study of mortality rates in criminal offenders from British Columbia.

scholarly journals Novel bounds for causal effects based on sensitivity parameters on the risk difference scale

2021 ◽  
Vol 9 (1) ◽  
pp. 190-210
Author(s):  
Arvid Sjölander ◽  
Ola Hössjer
Keyword(s):  
Risk Ratio ◽  
Simulation Study ◽  
Observational Studies ◽  
Causal Effect ◽  
Real Data ◽  
Risk Difference ◽  
Causal Effects ◽  
Ratio Scale ◽  
Unmeasured Confounding ◽  
Range Of Values

Abstract Unmeasured confounding is an important threat to the validity of observational studies. A common way to deal with unmeasured confounding is to compute bounds for the causal effect of interest, that is, a range of values that is guaranteed to include the true effect, given the observed data. Recently, bounds have been proposed that are based on sensitivity parameters, which quantify the degree of unmeasured confounding on the risk ratio scale. These bounds can be used to compute an E-value, that is, the degree of confounding required to explain away an observed association, on the risk ratio scale. We complement and extend this previous work by deriving analogous bounds, based on sensitivity parameters on the risk difference scale. We show that our bounds can also be used to compute an E-value, on the risk difference scale. We compare our novel bounds with previous bounds through a real data example and a simulation study.

scholarly journals Causal Association Between Heart Failure and Alzheimer’s Disease: A Two-Sample Bidirectional Mendelian Randomization Study

2022 ◽  
Vol 12 ◽  
Author(s):  
Chenglin Duan ◽  
Jingjing Shi ◽  
Guozhen Yuan ◽  
Xintian Shou ◽  
Ting Chen ◽  
...  
Keyword(s):  
Heart Failure ◽  
Alzheimer’S Disease ◽  
Alzheimer's Disease ◽  
Sensitivity Analysis ◽  
Causal Relationship ◽  
Observational Studies ◽  
Mendelian Randomization ◽  
Causal Effect ◽  
Causal Relationships ◽  
Simple Mode

Background: Traditional observational studies have demonstrated an association between heart failure and Alzheimer’s disease. The strengths of observational studies lie in their speed of implementation, cost, and applicability to rare diseases. However, observational studies have several limitations, such as uncontrollable confounders. Therefore, we employed Mendelian randomization of genetic variants to evaluate the causal relationships existing between AD and HF, which can avoid these limitations.Materials and Methods: A two-sample bidirectional MR analysis was employed. All datasets were results from the UK’s Medical Research Council Integrative Epidemiology Unit genome-wide association study database, and we conducted a series of control steps to select the most suitable single-nucleotide polymorphisms for MR analysis, for which five primary methods are offered. We reversed the functions of exposure and outcomes to explore the causal direction of HF and AD. Sensitivity analysis was used to conduct several tests to avoid heterogeneity and pleiotropic bias in the MR results.Results: Our MR studies did not support a meaningful causal relationship between AD on HF (MR-Egger, p = 0.634 > 0.05; weighted median (WM), p = 0.337 > 0.05; inverse variance weighted (IVW), p = 0.471 > 0.05; simple mode, p = 0.454 > 0.05; weighted mode, p = 0.401 > 0.05). At the same time, we did not find a significant causal relationship between HF and AD with four of the methods (MR-Egger, p = 0.195 > 0.05; IVW, p = 0.0879 > 0.05; simple mode, p = 0.170 > 0.05; weighted mode, p = 0.110 > 0.05), but the WM method indicated a significant effect of HF on AD (p = 0.025 < 0.05). Because the statistical powers of IVW and MR-Egger are more than that of WM, we think that there is no causal effect of HF on AD. Sensitivity analysis and horizontal pleiotropy were not detected in the MR analysis.Conclusion: Our results did not provide significant evidence indicating any causal relationships between HF and AD in the European population. Therefore, more large-scale datasets or datasets related to similar factors are expected for further MR analysis.

Sign in / Sign up

Export Citation Format

Share Document