Proximal indirect comparison

2024-05-17

AI Summary

This paper addresses the challenge of estimating the average treatment effect (ATE) in a target randomized controlled trial (RCT) where the key intervention arm is missing, and only a source RCT includes that intervention—under the presence of unobserved effect modifiers. Departing from conventional conditional transportability assumptions that require full covariate observability, we propose the first proximal identification framework leveraging proxy variables, which accommodates unobserved confounding-induced distributional shift. Our method integrates bridge function modeling, inverse probability weighting, and doubly robust estimation to construct a semiparametric estimator that is consistent and asymptotically normal. Theoretical guarantees are rigorously established. Empirically, we apply our approach to two clinical trials on weight management, successfully enabling indirect comparative analysis even with missing outcomes in the target trial.

Technology Category

Reasoning under Uncertainty: Relational Probabilistic ModelsIntelligent Robots: State EstimationMachine Learning: Causal Learning

Application Category

User Modeling, Personalization and Recommendation: User modeling for targeted and personalized online advertisingGraph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphsSecurity and Privacy: Large-scale security measurements

Abstract

We consider the problem of indirect comparison, where a treatment arm of interest is absent by design in the target randomized control trial (RCT) but available in a source RCT. The identifiability of the target population average treatment effect often relies on conditional transportability assumptions. However, it is a common concern whether all relevant effect modifiers are measured and controlled for. We highlight a new proximal identification result in the presence of shifted, unobserved effect modifiers based on proxies: an adjustment proxy in both RCTs and an additional reweighting proxy in the source RCT. We propose an estimator which is doubly-robust against misspecifications of the so-called bridge functions and asymptotically normal under mild consistency of the nuisance models. An alternative estimator is presented to accommodate missing outcomes in the source RCT, which we then apply to conduct a proximal indirect comparison analysis using two weight management trials.

Problem

• Indirect treatment comparison with missing data
• Identifiability with unobserved effect modifiers
• Doubly-robust estimation using proxy variables

Innovation

• Proximal identification with shifted unobserved modifiers
• Doubly-robust estimator for bridge functions
• Proxy-based adjustment and reweighting technique