A Structural Approach to Selection Bias

@article{Hernn2004ASA,
  title={A Structural Approach to Selection Bias},
  author={Miguel A. Hern{\'a}n and Sonia Hern{\'a}ndez-D{\'i}az and James M. Robins},
  journal={Epidemiology},
  year={2004},
  volume={15},
  pages={615-625}
}
The term “selection bias” encompasses various biases in epidemiology. We describe examples of selection bias in case-control studies (eg, inappropriate selection of controls) and cohort studies (eg, informative censoring). We argue that the causal structure underlying the bias in each example is essentially the same: conditioning on a common effect of 2 variables, one of which is either exposure or a cause of exposure and the other is either the outcome or a cause of the outcome. This structure… Expand
Limits for the Magnitude of M-bias and Certain Other Types of Structural Selection Bias.
TLDR
A bound for (the maximum) M-bias is derived and examples illustrate how to apply it with other types of selection bias, and suggest that M- bias may often, but not uniformly, be small. Expand
Adjusting for selection bias in retrospective, case-control studies.
TLDR
The method can help to determine whether selection bias is present and thus confirm the validity of study conclusions when no evidence of selection bias can be found, and is demonstrated using simulations that the estimates of the odds ratios produced by the method are consistently closer to the true odds ratio than standard odds ratio estimates using logistic regression. Expand
Instrumental Variable Analyses and Selection Bias
TLDR
The presence of bias is demonstrated in IV analyses on a sample from a simulated dataset, where selection into the sample was a collider on a noncausal path from E to Y, by applying inverse probability of selection weights which can minimize bias. Expand
Selection Mechanisms and Their Consequences: Understanding and Addressing Selection Bias
TLDR
This work used causal directed acyclic graphs to demonstrate the types of bias that can result when selection into a sample is associated with the exposure or outcome of interest, or with both. Expand
Overadjustment Bias and Unnecessary Adjustment in Epidemiologic Studies
TLDR
This work uses causal diagrams and an empirical example (the effect of maternal smoking on neonatal mortality) to illustrate and clarify the definition of overadjustment bias, and to distinguish over adjustment bias from unnecessary adjustment. Expand
Reflection on modern methods: selection bias-a review of recent developments.
Selection bias remains a more difficult bias to understand than confounding or measurement error. Past definitions have not always been illuminating and a simple method (such as theExpand
Selection bias modeling using observed data augmented with imputed record-level probabilities.
TLDR
This internal adjustment technique using user-supplied bias parameters and inverse probability weighting for selection bias can be applied to any type of observational study. Expand
Bias due to differential participation in case-control studies and review of available approaches for adjustment
TLDR
This paper formally describes the problem and shows its potential extent, reviews existing approaches for bias adjustment applicable under certain conditions, compares and applies them, and demonstrates the approaches’ application via data from a case-control study on stroke. Expand
Implications of M bias in epidemiologic studies: a simulation study.
TLDR
In scenarios resembling those the authors examined, M bias had a small impact unless associations between the collider and unmeasured confounders were very large (relative risk > 8). Expand
Collider scope: when selection bias can substantially influence observed associations
TLDR
In simulations, it is shown that even modest influences on selection into, or attrition from, a study can generate biased and potentially misleading estimates of both phenotypic and genotypic associations. Expand
...
1
2
3
4
5
...

References

SHOWING 1-10 OF 40 REFERENCES
Quantifying Biases in Causal Models: Classical Confounding vs Collider-Stratification Bias
TLDR
The results suggest that bias from stratifying on variables affected by exposure and disease may often be comparable in size with bias from classical confounding, whereas other biases from collider stratification may tend to be much smaller. Expand
Causation of Bias: The Episcope
TLDR
For epidemiologists, this article serves as a review of ideas about confounding, information bias, and selection bias and underscores the need for routinely analyzing the sensitivity of study findings to multiple hypothesized biases. Expand
Fallibility in estimating direct effects.
We use causal graphs and a partly hypothetical example from the Physicians' Health Study to explain why a common standard method for quantifying direct effects (i.e. stratifying on the intermediateExpand
A new approach to causal inference in mortality studies with a sustained exposure period—application to control of the healthy worker survivor effect
TLDR
A graphical approach to the identification and computation of causal parameters in mortality studies with sustained exposure periods is offered and an adverse effect of arsenic exposure on all-cause and lung cancer mortality which standard methods fail to detect is found. Expand
Marginal Structural Models and Causal Inference in Epidemiology
In observational studies with exposures or treatments that vary over time, standard approaches for adjustment of confounding are biased when there exist time-dependent confounders that are alsoExpand
An analysis of detection bias and proposed corrections in the study of estrogens and endometrial cancer.
TLDR
It can be said that the problem of detection bias produces a selection bias in case-control studies and a misclassification bias in follow-up studies when there is empirical information on detection possibilities however the degree of bias present in conventional results can be estimated and corrected. Expand
Marginal structural models to estimate the causal effect of zidovudine on the survival of HIV-positive men.
TLDR
The marginal structural Cox proportional hazards model is described and used to estimate the causal effect of zidovudine on the survival of human immunodeficiency virus-positive men participating in the Multicenter AIDS Cohort Study. Expand
Identifiability and Exchangeability for Direct and Indirect Effects
TLDR
It is shown that adjustment for the intermediate variable, which is the most common method of estimating direct effects, can be biased, and that, even in a randomized crossover trial of exposure, direct and indirect effects cannot be separated without special assumptions. Expand
Causal knowledge as a prerequisite for confounding evaluation: an application to birth defects epidemiology.
TLDR
Findings are presented from the Slone Epidemiology Unit Birth Defects Study, 1992-1997, a case-control study on folic acid supplementation and risk of neural tube defects, which suggests that the crude odds ratio should be used because the adjusted odds ratio is invalid. Expand
Causality theory for policy uses of epidemiological measures
This paper provides an introduction to measures of causal effects and focuses on underlying conceptual models, definitions and drawbacks of special relevance to policy formulation based onExpand
...
1
2
3
4
...