TALKs

Speaker: Sihyung Park, Department of Statistics, North Carolina State University

Title:  Estimating the long-term treatment effect under complex unmeasured confounding: a proximal data integration approach

Abstract: 

Estimating the long-term treatment effect is crucial for cost-effectiveness analysis and pricing negotiation. The limited follow-up duration of typical randomized controlled trials (RCTs), however, presents a significant challenge. A common strategy is to supplement RCT data with external real-world cohorts, but existing methods for this integration rely on strong, untestable assumptions. Specifically, they often assume transportability that implies unmeasured confounding effects are non-existent or constant over time, which limits their validity, particularly under common data-generating processes such as linear factor models.

To address these restrictions, this work develops a novel semiparametric data integration methodology inspired by proximal causal inference. Our framework successfully integrates heterogeneous data sources even in the presence of time-varying latent confounding. The result is a more robust and valid estimation of the long-term average treatment effect in settings where traditional approaches are compromised by their rigid assumptions.

Speaker: Siqi Cao, Department of Statistics, North Carolina State University

Title:  Enhanced Treatment Effect Evaluation Leveraging Trial and External Control with Missing Data (Summer Project at Merck)

Slides are available here.