Bayesian data fusion: Probabilistic sensitivity analysis for unmeasured confounding using informative priors based on secondary data
- 16 February 2021
- journal article
- research article
- Published by Oxford University Press (OUP) in Biometrics
- Vol. 78 (2), 730-741
- https://doi.org/10.1111/biom.13436
Abstract
Bayesian causal inference offers a principled approach to policy evaluation of proposed interventions on mediators or time‐varying exposures. Building on the Bayesian g‐formula method introduced by Keil et al., we outline a general approach for the estimation of population‐level causal quantities involving dynamic and stochastic treatment regimes, including regimes related to mediation estimands such as natural direct and indirect effects. We further extend this approach to propose a Bayesian data fusion (BDF), an algorithm for performing probabilistic sensitivity analysis when a confounder unmeasured in a primary data set is available in an external data source. When the relevant relationships are causally transportable between the two source populations, BDF corrects confounding bias and supports causal inference and decision‐making within the main study population without sharing of the individual‐level external data set. We present results from a simulation study comparing BDF to two common frequentist correction methods for unmeasured mediator‐outcome confounding bias in the mediation setting. We use these methods to analyze data on the role of stage at cancer diagnosis in contributing to Black–White colorectal cancer survival disparities.Keywords
Funding Information
- National Institute of Environmental Health Sciences (P30ES000002, R01ES026217, T32ES007142)
- Environmental Protection Agency (RD‐835872)
- National Institute of Mental Health (K01MH118477)
- National Cancer Institute (P01CA134294, T32CA009337)
- National Institute of General Medical Sciences (R01GM111339)
This publication has 32 references indexed in Scilit:
- On model selection and model misspecification in causal inferenceStatistical Methods in Medical Research, 2010
- Bias Formulas for Sensitivity Analysis for Direct and Indirect EffectsEpidemiology, 2010
- Effects of Socioeconomic Status and Treatment Disparities in Colorectal Cancer SurvivalCancer Epidemiology, Biomarkers & Prevention, 2008
- The relationship between the power prior and hierarchical modelsBayesian Analysis, 2006
- Improving ecological inference using individual‐level dataStatistics in Medicine, 2005
- Understanding Cancer Treatment and Outcomes: The Cancer Care Outcomes Research and Surveillance ConsortiumJournal of Clinical Oncology, 2004
- Power prior distributions for regression modelsStatistical Science, 2000
- Adjusting for Nonignorable Drop-Out Using Semiparametric Nonresponse ModelsJournal of the American Statistical Association, 1999
- Inference from Iterative Simulation Using Multiple SequencesStatistical Science, 1992
- The Bayesian BootstrapThe Annals of Statistics, 1981