Roadmap for systematic identification and analysis of multiple biases in causal inference
Abstract
Observational studies examining causal effects rely on unverifiable assumptions, the violation of which can induce multiple biases. Quantitative bias analysis (QBA) methods examine the sensitivity of findings to such violations, generally, by producing estimates under alternative assumptions, incorporating additional information. Although substantial guidance exists for implementing QBA, there is limited guidance on how to systematically determine the assumptions underlying a primary causal analysis and the potential violations that should guide bias analysis. Consequently, many assumptions remain implicit, leading to selective and therefore misleading QBA. To address this gap, we propose a roadmap for systematically identifying and analyzing multiple biases. Briefly, this consists of (1) articulating the assumptions underlying the primary analysis through specification of the “ideal trial” that defines the causal estimand and the assumptions required by its (possibly implicit) emulation, depicting these assumptions using a causal diagram; (2) extending the diagram to depict alternative assumptions, under which biases may arise; (3) obtaining a single estimate that simultaneously corrects for all potential biases. We illustrate the roadmap using an investigation of the effect of breastfeeding on the risk of childhood asthma and, through simulations, demonstrate the need for analyzing multiple biases jointly rather than one at a time.
Related
- Identification and Robust Inference for Multiple Treatments with Possibly Invalid Instruments
- Debiased Inference for AI-Generated Data without Gold-Standard Labels: Identification via Multiple Imperfect Measurements
- Deep Bayesian Filtering for Nonlinear Structural System Identification: A Tutorial on Structural and Physical Inductive Biases
- Efficient Bayesian inference for multiple network data
- Quantile-Only Regression: Inference and Partial Identification under Quantile Disclosure
BibTeX
@article{wijesuriya2026roadmap,
title = {{Roadmap for systematic identification and analysis of multiple biases in causal inference}},
author = {Rushani Wijesuriya and Rachael Hughes and John B Carlin and Rachel L Peters and Jennifer J Koplin and Margarita Moreno-Betancur},
journal = {American Journal of Epidemiology},
year = {2026},
month = jul,
doi = {10.1093/aje/kwag166},
url = {https://doi.org/10.1093/aje/kwag166},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Roadmap for systematic identification and analysis of multiple biases in causal inference
AU - Rushani Wijesuriya
AU - Rachael Hughes
AU - John B Carlin
AU - Rachel L Peters
AU - Jennifer J Koplin
AU - Margarita Moreno-Betancur
PY - 2026
DA - 2026/07/16
JO - American Journal of Epidemiology
DO - 10.1093/aje/kwag166
UR - https://doi.org/10.1093/aje/kwag166
AB - Observational studies examining causal effects rely on unverifiable assumptions, the violation of which can induce multiple biases. Quantitative bias analysis (QBA) methods examine the sensitivity of findings to such violations, generally, by producing estimates under alternative assumptions, incorporating additional information. Although substantial guidance exists for implementing QBA, there is limited guidance on how to systematically determine the assumptions underlying a primary causal analysis and the potential violations that should guide bias analysis. Consequently, many assumptions remain implicit, leading to selective and therefore misleading QBA. To address this gap, we propose a roadmap for systematically identifying and analyzing multiple biases. Briefly, this consists of (1) articulating the assumptions underlying the primary analysis through specification of the “ideal trial” that defines the causal estimand and the assumptions required by its (possibly implicit) emulation, depicting these assumptions using a causal diagram; (2) extending the diagram to depict alternative assumptions, under which biases may arise; (3) obtaining a single estimate that simultaneously corrects for all potential biases. We illustrate the roadmap using an investigation of the effect of breastfeeding on the risk of childhood asthma and, through simulations, demonstrate the need for analyzing multiple biases jointly rather than one at a time.
SN - 0002-9262
SN - 1476-6256
ER -