Statistical methods for cost-effectiveness analysis of left-truncated censored survival data with treatment delays

2025-05-09

AI Summary

In public health, treatment initiation delay induces left-truncated survival data, complicating cost-effectiveness evaluation of interventions (e.g., antiretroviral therapy switching) due to biased effect estimation and inadequate handling of timing-related confounding. Method: We propose a unified statistical inference framework integrating a stratified semiparametric Cox model with incremental cost-effectiveness ratio (ICER) and incremental net benefit (INB) estimation, explicitly modeling treatment delay time to ensure causal interpretability and small-sample robustness. Contribution/Results: Simulation studies and analysis of a real-world Tanzanian cohort demonstrate superior finite-sample performance across diverse delay scenarios. The framework maintains valid type-I error control, achieves high statistical power, and yields reliable ICER and INB estimates even under substantial truncation. It has been successfully implemented in clinical decision support for HIV treatment optimization, enabling evidence-based policy translation while preserving methodological rigor.

Technology Category

• Reasoning under Uncertainty: Causality
• Search and Optimization: Sampling/Simulation-based Search
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for incomplete, noisy, or partially observed Web-related graphs

Abstract

The incremental cost-effectiveness ratio (ICER) and incremental net benefit (INB) are widely used for cost-effectiveness analysis. We develop methods for estimation and inference for the ICER and INB which use the semiparametric stratified Cox proportional hazard model, allowing for adjustment for risk factors. Since in public health settings, patients often begin treatment after they become eligible, we account for delay times in treatment initiation. Excellent finite sample properties of the proposed estimator are demonstrated in an extensive simulation study under different delay scenarios. We apply the proposed method to evaluate the cost-effectiveness of switching treatments among AIDS patients in Tanzania.

Problem

• Estimating cost-effectiveness ratios for survival data with delays
• Adjusting for treatment initiation delays in public health
• Evaluating AIDS treatment cost-effectiveness using new methods

Innovation

• Uses semiparametric stratified Cox model
• Adjusts for treatment delay times
• Estimates cost-effectiveness with risk factors