More powerful logrank permutation tests for two-sample survival data

TL;DR

This paper introduces a flexible, combined weighted logrank test for two-sample survival data that offers broader power, better finite sample performance, and maintains theoretical properties like asymptotic exactness and consistency.

Contribution

It proposes a novel combination approach for weighted logrank tests, enhancing power and finite sample performance in survival analysis.

Findings

Permutation version is finitely exact under exchangeability.

The new test outperforms traditional methods in simulations.

The method is demonstrated on real data.

Abstract

Weighted logrank tests are a popular tool for analyzing right censored survival data from two independent samples. Each of these tests is optimal against a certain hazard alternative, for example the classical logrank test for proportional hazards. But which weight function should be used in practical applications? We address this question by a flexible combination idea leading to a testing procedure with broader power. Beside the test's asymptotic exactness and consistency its power behaviour under local alternatives is derived. All theoretical properties can be transferred to a permutation version of the test, which is even finitely exact under exchangeability and showed a better finite sample performance in our simulation study. The procedure is illustrated in a real data example.