Surrogate-powered Causal Inference with Censored Outcomes

2026-10-04

AI Summary

This study addresses information loss due to censoring in survival endpoints of clinical trials, which limits the precision of causal inference. We propose a novel framework that recovers censored information by leveraging post-treatment disease history. Specifically, we derive the influence function and efficiency gain identity for observed data to construct target parameter-preserving estimators. We demonstrate that efficiency improvements are jointly driven by censoring risk, state splitting, and residual survival separation, achieving root-N consistent convergence and robust inference. Empirical evaluations using the Aalen-Johansen estimator, Cox-Breslow models, and semi-synthetic breast cancer data confirm that our method substantially enhances estimation efficiency for marginal survival effects. This work provides an efficient and generalizable inference paradigm for censored survival analysis involving intermediate events.

Abstract

Clinical trials with survival endpoints lose information when participants are censored before death is observed. We develop target-preserving estimators that use posttreatment disease history, such as recurrence or progression, to recover information lost to censoring for marginal survival and restricted mean survival effects. The difficulty is that the intermediate event is downstream of treatment: naive adjustment can change the causal estimand, and the useful information enters only through the observed coarsening. We derive observed-data influence functions with and without recurrence history and obtain an exact gain identity. The identity shows that efficiency improvement is driven by the censoring hazard, the split of the alive risk set into recurrence states, and the residual-survival separation between those states. In the no-covariate illness-death model, the Aalen-Johansen estimator realizes the recurrence-augmented efficient score after standardization to the marginal target. With covariates, correctly specified Cox-Breslow transition hazards provide a root-N plug-in benchmark, while a hazard-induced one-step estimator gives rate robustness and, under primitive transition and censoring learner rates, canonical inference with a second-order product remainder. A semi-synthetic metastatic breast cancer study calibrated from digitized progression-free survival and overall survival curves illustrates the gain identity. The framework applies broadly to censored time-to-event studies with informative intermediate histories.

Problem

• surrogate outcomes
• censored data
• causal inference
• survival analysis
• intermediate events

Innovation

• surrogate-powered causal inference
• censored outcomes
• target-preserving estimators
• exact gain identity
• hazard-induced one-step estimator