Skip to main content

# Causal Mediation Analysis

2 related articles

Subscribe to updates

arXiv Nov 10, 2026
Mediation analysis stops refitting per table: MediPFN learns interval error quantiles in advance and runs 59.1 times faster than bootstrap in paired timing
Causal mediation analysis splits a treatment effect into an indirect pathway through mediators and a direct pathway through other processes, and attaching a confidence interval to the indirect effect usually means refitting several nuisance functions on every table and resampling repeatedly. MediPFN adapts a pretrained tabular foundation model to predict those nuisance functions, then a separate head learns quantiles of the estimator's standardized error from simulated tables with known indirect effects, with symmetrization and independent calibration producing the interval. On synthetic, distribution-shift and real-covariate tables its coverage sits closer to the 95% nominal level than several comparators, and paired timing gives a median 59.1-fold speedup over bootstrap.
Causal mediation analysis splits a treatment effect into an indirect pathway through mediators and a direct pathway through other processes, and attaching a confidence interval to the indirect effect usually means refitting several nuisance functions on every table and resampling repeatedly. MediPFN adapts a pretrained tabular foundation model to predict those nuisance functions, then a separate head learns quantiles of the estimator's standardized error from simulated tables with known indirect effects, with symmetrization and independent calibration producing the interval. On synthetic, distribution-shift and real-covariate tables its coverage sits closer to the 95% nominal level than several comparators, and paired timing gives a median 59.1-fold speedup over bootstrap.
Bo Shi（First author）
MediPFN Causal Mediation Analysis Natural Indirect Effect Estimation Tabular Foundation Model Uncertainty Quantification
medRxiv Oct 23, 2026
Compressing health records for causal analysis can drop a rare code, cutting mediation interval coverage to 2%
Multimodal health records are often compressed into low-dimensional representations and used as adjustment sets, yet variational autoencoders are trained to reconstruct data rather than to preserve what causal identification needs. In simulations, a representation that discarded a rare code affecting both treatment and mediator gave 95% intervals for the interventional indirect effect with 2% coverage at n = 3,000 under moderate confounding, while reconstruction loss did not reveal the problem; a diagnostic comparing nuisance-function predictive loss flagged it in every replicate.
Multimodal health records are often compressed into low-dimensional representations and used as adjustment sets, yet variational autoencoders are trained to reconstruct data rather than to preserve what causal identification needs. In simulations, a representation that discarded a rare code affecting both treatment and mediator gave 95% intervals for the interventional indirect effect with 2% coverage at n = 3,000 under moderate confounding, while reconstruction loss did not reveal the problem; a diagnostic comparing nuisance-function predictive loss flagged it in every replicate.
Stanford University
Cong Cao（First author, Corresponding author）
Causal Mediation Analysis Interventional Direct and Indirect Effects Variational Autoencoder Doubly Robust Estimation Digital Twin

You’re all caught up