arXiv subjects

Dominic Magirr

Publications and source records attributed to Dominic Magirr.

14 recordsLinked to original sources

arXiv2026

Finite-sample correction for the covariate-adjusted log-rank test

The covariate-adjusted log-rank test is a novel method for covariate adjustment in randomized trials with time-to-event endpoints, offering guaranteed efficiency gains compared to the standard log-rank test. However, it has been noted that, in small samples, this method may lead to type I error rate inflation. This issue is particularly pronounced in trials with imbalanced allocation and settings where the number of adjustment covariates is large relative to the sample size. We propose a finite-sample correction for the denominator of the covariate-adjusted log-rank test statistic that accounts for the loss of the residual degrees of freedom as well as the uncertainty in the unknown regression coefficients. In simulations, we show that applying this correction leads to a substantial reduction in the type I error rate inflation across multiple scenarios.

stat.ME↗

arXiv2026

Confidence intervals for two-stage adaptive designs with subpopulation selection

We consider clinical trials in which an experimental treatment is compared with a control in pre-specified patient subpopulations. In such settings, adaptive enrichment designs allow the enrolled population to be modified at an interim analysis, with subpopulations selected according to preplanned rules. Since these interim decisions are data-dependent, valid statistical inference must account for them. We focus on constructing confidence intervals for the treatment effect in the selected population. Confidence interval methods that ignore the possibility of population modification may fail to achieve the desired coverage probability. We propose a new approach that constructs confidence intervals with exact nominal coverage conditional on the interim decision. Importantly, our method applies to a broad class of adaptive enrichment designs, rather than a single specific design. Our method involves deriving the distribution of the naive estimator of the treatment effect in the selected population conditional on the interim decision and inverting uniformly most accurate unbiased tests to obtain the confidence interval. We provide an efficient computational procedure and show through extensive simulations that the resulting confidence intervals satisfy the theoretical coverage guarantees.

stat.ME↗

arXiv2026

How to interpret hazard ratios

The hazard ratio, typically estimated using Cox's famous proportional hazards model, is the most common effect measure used to describe the association or effect of a covariate on a time-to-event outcome. In recent years the hazard ratio has been argued by some to lack a causal interpretation, even in randomised trials, and even if the proportional hazards assumption holds. This is concerning, not least due to the ubiquity of hazard ratios in analyses of time-to-event data. We review these criticisms, describe how we think hazard ratios should be interpreted, and argue that they retain a valid causal interpretation. Nevertheless, alternative measures may be preferable to describe effects of exposures or treatments on time-to-event outcomes.

stat.ME↗

arXiv2024

Robust modestly weighted log-rank tests

The introduction of checkpoint inhibitors in immuno-oncology has raised questions about the suitability of the log-rank test as the default primary analysis method in confirmatory studies, particularly when survival curves exhibit non-proportional hazards. The log-rank test, while effective in controlling false positive rates, may lose power in scenarios where survival curves remain similar for extended periods before diverging. To address this, various weighted versions of the log-rank test have been proposed, including the MaxCombo test, which combines multiple weighted log-rank statistics to enhance power across a range of alternative hypotheses. Despite its potential, the MaxCombo test has seen limited adoption, possibly owing to its proneness to produce counterintuitive results in situations where the hazard functions on the two arms cross. In response, the modestly weighted log-rank test was developed to provide a balanced approach, giving greater weight to later event times while avoiding undue influence from early detrimental effects. However, this test also faces limitations, particularly if the possibility of early separation of survival curves cannot be ruled out a priori. We propose a novel test statistic that integrates the strengths of the standard log-rank test, the modestly weighted log-rank test, and the MaxCombo test. By considering the maximum of the standard log-rank statistic and a modestly weighted log-rank statistic, the new test aims to maintain power under delayed effect scenarios while minimizing power loss, relative to the log-rank test, in worst-case scenarios. Simulation studies and a case study demonstrate the efficiency and robustness of this approach, highlighting its potential as a robust alternative for primary analysis in immuno-oncology trials.

stat.ME↗

arXiv2024

Efficiency of nonparametric superiority tests based on restricted mean survival time versus the log-rank test under proportional hazards

Background: For RCTs with time-to-event endpoints, proportional hazard (PH) models are typically used to estimate treatment effects and logrank tests are commonly used for hypothesis testing. There is growing support for replacing this approach with a model-free estimand and assumption-lean analysis method. One alternative is to base the analysis on the difference in restricted mean survival time (RMST) at a specific time, a single-number summary measure that can be defined without any restrictive assumptions on the outcome model. In a simple setting without covariates, an assumption-lean analysis can be achieved using nonparametric methods such as Kaplan Meier estimation. The main advantage of moving to a model-free summary measure and assumption-lean analysis is that the validity and interpretation of conclusions do not depend on the PH assumption. The potential disadvantage is that the nonparametric analysis may lose efficiency under PH. There is disagreement in recent literature on this issue. Methods: Asymptotic results and simulations are used to compare the efficiency of a log-rank test against a nonparametric analysis of the difference in RMST in a superiority trial under PH. Previous studies have separately examined the effect of event rates and the censoring distribution on relative efficiency. This investigation clarifies conflicting results from earlier research by exploring the joint effect of event rate and censoring distribution together. Several illustrative examples are provided. Results: In scenarios with high event rates and/or substantial censoring across a large proportion of the study window, and when both methods make use of the same amount of data, relative efficiency is close to unity. However, in cases with low event rates but when censoring is concentrated at the end of the study window, the PH analysis has a considerable efficiency advantage.

stat.ME↗

arXiv2024

Treatment-control comparisons in platform trials including non-concurrent controls

Shared controls in platform trials comprise concurrent and non-concurrent controls. For a given experimental arm, non-concurrent controls refer to data from patients allocated to the control arm before the arm enters the trial. The use of non-concurrent controls in the analysis is attractive because it may increase the trial's power of testing treatment differences while decreasing the sample size. However, since arms are added sequentially in the trial, randomization occurs at different times, which can introduce bias in the estimates due to time trends. In this article, we present methods to incorporate non-concurrent control data in treatment-control comparisons, allowing for time trends. We focus mainly on frequentist approaches that model the time trend and Bayesian strategies that limit the borrowing level depending on the heterogeneity between concurrent and non-concurrent controls. We examine the impact of time trends, overlap between experimental treatment arms and entry times of arms in the trial on the operating characteristics of treatment effect estimators for each method under different patterns for the time trends. We argue under which conditions the methods lead to type 1 error control and discuss the gain in power compared to trials only using concurrent controls by means of a simulation study in which methods are compared.

stat.ME↗

arXiv2023

Visualizing hypothesis tests in survival analysis under anticipated delayed effects

What can be considered an appropriate statistical method for the primary analysis of a randomized clinical trial (RCT) with a time-to-event endpoint when we anticipate non-proportional hazards owing to a delayed effect? This question has been the subject of much recent debate. The standard approach is a log-rank test and/or a Cox proportional hazards model. Alternative methods have been explored in the statistical literature, such as weighted log-rank tests and tests based on the Restricted Mean Survival Time (RMST). While weighted log-rank tests can achieve high power compared to the standard log-rank test, some choices of weights may lead to type-I error inflation under particular conditions. In addition, they are not linked to an unambiguous estimand. Arguably, therefore, they are difficult to intepret. Test statistics based on the RMST, on the other hand, allow one to investigate the average difference between two survival curves up to a pre-specified time point $τ$ -- an unambiguous estimand. However, by emphasizing differences prior to $τ$, such test statistics may not fully capture the benefit of a new treatment in terms of long-term survival. In this article, we introduce a graphical approach for direct comparison of weighted log-rank tests and tests based on the RMST. This new perspective allows a more informed choice of the analysis method, going beyond power and type I error comparison.

stat.ME↗

arXiv2023

NCC: An R-package for analysis and simulation of platform trials with non-concurrent controls

Platform trials evaluate the efficacy of multiple treatments, allowing for late entry of the experimental arms and enabling efficiency gains by sharing controls. The power of individual treatment-control comparisons in such trials can be improved by utilizing non-concurrent controls (NCC) in the analysis. We present the R-package NCC for the design and analysis of platform trials using non-concurrent controls. NCC allows for simulating platform trials and evaluating the properties of analysis methods that make use of non-concurrent controls in a variety of settings. We describe the main NCC functions and show how to use the package to simulate and analyse platform trials by means of specific examples.

stat.ME↗

arXiv2022

On model-based time trend adjustments in platform trials with non-concurrent controls

Platform trials can evaluate the efficacy of several treatments compared to a control. The number of treatments is not fixed, as arms may be added or removed as the trial progresses. Platform trials are more efficient than independent parallel-group trials because of using shared control groups. For arms entering the trial later, not all patients in the control group are randomised concurrently. The control group is then divided into concurrent and non-concurrent controls. Using non-concurrent controls (NCC) can improve the trial's efficiency, but can introduce bias due to time trends. We focus on a platform trial with two treatment arms and a common control arm. Assuming that the second treatment arm is added later, we assess the robustness of model-based approaches to adjust for time trends when using NCC. We consider approaches where time trends are modeled as linear or as a step function, with steps at times where arms enter or leave the trial. For trials with continuous or binary outcomes, we investigate the type 1 error (t1e) rate and power of testing the efficacy of the newly added arm under a range of scenarios. In addition to scenarios where time trends are equal across arms, we investigate settings with trends that are different or not additive in the model scale. A step function model fitted on data from all arms gives increased power while controlling the t1e, as long as the time trends are equal for the different arms and additive on the model scale. This holds even if the trend's shape deviates from a step function if block randomisation is used. But if trends differ between arms or are not additive on the model scale, t1e control may be lost. The efficiency gained by using step function models to incorporate NCC can outweigh potential biases. However, the specifics of the trial, plausibility of different time trends, and robustness of results should be considered

stat.ME↗

arXiv2022

Stratified modestly-weighted log-rank tests in settings with an anticipated delayed separation of survival curves

Delayed separation of survival curves is a common occurrence in confirmatory studies in immuno-oncology. Many novel statistical methods that aim to efficiently capture potential long-term survival improvements have been proposed in recent years. However, the vast majority do not consider stratification, which is a major limitation considering that most (if not all) large confirmatory studies currently employ a stratified primary analysis. In this article, we combine recently proposed weighted log-rank tests that have been designed to work well under a delayed separation of survival curves, with stratification by a baseline variable. The aim is to increase the efficiency of the test when the stratifying variable is highly prognostic for survival. As there are many potential ways to combine the two techniques, we compare several possibilities in an extensive simulation study. We also apply the techniques retrospectively to two recent randomized clinical trials.

stat.ME↗

arXiv2021

Design and Analysis of group-sequential clinical trials based on a modestly-weighted log-rank test in anticipation of a delayed separation of survival curves: A practical guidance

A common feature of many recent trials evaluating the effects of immunotherapy on survival is that non-proportional hazards can be anticipated at the design stage. This raises the possibility to use a statistical method tailored towards testing the purported long-term benefit, rather than applying the more standard log-rank test and/or Cox model. Many such proposals have been made in recent years, but there remains a lack of practical guidance on implementation, particularly in the context of group-sequential designs. In this article, we aim to fill this gap. We discuss how the POPLAR trial, which compared immunotherapy versus chemotherapy in non-small-cell lung cancer, might have been re-designed to be more robust to the presence of a delayed effect. We then provide step-by-step instructions on how to analyse a hypothetical realisation of the trial, based on this new design. Basic theory on weighted log-rank tests and group-sequential methods is covered, and an accompanying R package (including vignette) is provided.

stat.AP↗

arXiv2020

Non-proportional hazards in immuno-oncology: is an old perspective needed?

A fundamental concept in two-arm non-parametric survival analysis is the comparison of observed versus expected numbers of events on one of the treatment arms (the choice of which arm is arbitrary), where the expectation is taken assuming that the true survival curves in the two arms are identical. This concept is at the heart of the counting-process theory that provides a rigorous basis for methods such as the log-rank test. It is natural, therefore, to maintain this perspective when extending the log-rank test to deal with non-proportional hazards, for example by considering a weighted sum of the "observed - expected" terms, where larger weights are given to time periods where the hazard ratio is expected to favour the experimental treatment. In doing so, however, one may stumble across some rather subtle issues, related to the difficulty in ascribing a causal interpretation to hazard ratios, that may lead to strange conclusions. An alternative approach is to view non-parametric survival comparisons as permutation tests. With this perspective, one can easily improve on the efficiency of the log-rank test, whilst thoroughly controlling the false positive rate. In particular, for the field of immuno-oncology, where researchers often anticipate a delayed treatment effect, sample sizes could be substantially reduced without loss of power.

stat.AP↗

arXiv2018

Modestly Weighted Logrank Tests

We propose a new class of weighted logrank tests (WLRT) that control the risk of concluding that a new drug is more efficacious than standard of care, when, in fact, it is uniformly inferior. Perhaps surprisingly, this risk is not controlled for WLRT in general. Tests from this new class can be constructed to have high power under a delayed-onset treatment effect scenario, as well as being almost as efficient as the standard logrank test under proportional hazards.

stat.AP↗

arXiv2014

Adaptive Survival Trials

Mid-study design modifications are becoming increasingly accepted in confirmatory clinical trials, so long as appropriate methods are applied such that error rates are controlled. It is therefore unfortunate that the important case of time-to-event endpoints is not easily handled by the standard theory. We analyze current methods that allow design modifications to be based on the full interim data, i.e., not only the observed event times but also secondary endpoint and safety data from patients who are yet to have an event. We show that the final test statistic may ignore a substantial subset of the observed event times. Since it is the data corresponding to the earliest recruited patients that is ignored, this neglect becomes egregious when there is specific interest in learning about long-term survival. An alternative test incorporating all event times is proposed, where a conservative assumption is made in order to guarantee type I error control. We examine the properties of our proposed approach using the example of a clinical trial comparing two cancer therapies.

stat.AP↗