On Response-Adaptive Targeting Strategies for Multi-Treatment Experiments

2026-06-16

AI Summary

This study addresses the challenge in multi-arm clinical trials where response-adaptive randomization struggles to ensure continuous exploration across all treatment arms under sparse target allocations and lacks a unified theoretical framework. The authors propose the α-rebalanced target strategy (αRTS), which establishes, for the first time, a general framework for response-adaptive randomization applicable to any number of treatment arms \( K \geq 2 \). By incorporating a forced exploration mechanism, αRTS guarantees that every arm is sampled infinitely often. Theoretical analysis demonstrates that αRTS possesses strong consistency, asymptotic normality, and statistical efficiency. Simulation studies in a three-arm setting confirm its favorable finite-sample performance and highlight the critical role of forced exploration in maintaining robustness under sparse target allocations.

Abstract

Response-adaptive randomization (RAR) in clinical trials aims to improve ethical and statistical efficiency by dynamically allocating patients to treatments based on observed outcomes. While RAR based on a target optimal allocation have been extensively studied for two-arms settings, their extension to multi-treatment experiments ($K \geq 2$) remains theoretically fragmented, with most existing methods focusing on specific algorithms or restricted target allocations. In this paper, we introduce a unified framework for response-adaptive targeting, the $α$-Rebalancing Targeting Strategies ($α$RTS), which generalize the ERADE two-armed strategy of Hu et al. [2009]. We prove that all designs in this family share fundamental asymptotic properties: strong consistency, asymptotic normality of allocation proportions and treatment effect estimators, and asymptotic efficiency. To address sparse target regimes (where some treatments are asymptotically eliminated), we further propose $α$RTS with Forced Exploration, a variant that guarantees infinite sampling for all treatments while preserving the asymptotic guarantees. Extensive simulations illustrate the finite-sample behavior of $α$RTS variants in a 3-armed context, highlighting in particular the critical role of forced exploration in sparse settings.

Problem

• response-adaptive randomization
• multi-treatment experiments
• target allocation
• sparse regimes
• clinical trials

Innovation

• response-adaptive randomization
• multi-treatment experiments
• α-Rebalancing Targeting Strategies
• forced exploration
• asymptotic efficiency