Online Control of the Family-wise Error Rate (FWER) for Multi-stage Platform Trials

arXiv:2611.06530 · stat.ME · Submitted 2026-11-04 · Read on arXiv

Listen

Radio episode about this paper

Transcript

Introduction to the show: ident: Statistics. The day in statistics: methods, theory and the inference under them.

Tom: We're looking at the paper: "Online Control of the Family-wise Error Rate (FWER) for Multi-stage Platform Trials".

Jane: Here's the gist of the paper. In platform trials, multiple intervention arms can enter or leave the study at any time, which brings the challenge of how to control the type I error rate In a platform trial, multiple intervention arms can enter or leave the study at any time, which brings the challenge of how to control the type I error rate.

Tom: The ADDIS-Spending (ADaptive DIScarding Alpha-spending) procedure controls the FWER in an online setting Problem Definition and Motivation Platform trials present a challenge regarding error rate control because multiple hypotheses are tested sequentially over time, and the total number of hypotheses is unknown The paper addresses the lack of a method that controls the FWER and allows new arms to be added at any time without prior information While standard two-arm trials focus on a single intervention, platform trials allow multiple treatments to enter at any time point, introducing multiplicity problems since the total number of arms is not known initially.

Jane: The family-wise error rate (FWER) control is usually the standard requirement in a confirmatory clinical trial.

Tom: Proposed Method Development The authors propose the group sequential ADDIS-Spending algorithm to enable interim analyses for each hypothesis in platform trials This method updates the error level of an individual hypothesis at all (interim) analyses whenever new decisions on other hypotheses are made.

Title and authors: Tom: The title itself, "Online Control of the Family-wise Error Rate (FWER) for Multi-stage Platform Trials," tells you right away it's about managing multiple treatments over time without knowing the whole set beforehand.

Jane: And they are led by researchers from Regensburg and Cambridge, who are clearly deep in the statistical weeds of how to handle sequential testing under these complex conditions.

Lu: What’s interesting is that they point out that while other papers have looked at false discovery rate control or multiplicity issues with shared control arms, this paper specifically aims for FWER control while allowing new arms to be added without prior information.

Meng: So, the core problem they are solving isn't just about testing more hypotheses; it’s about maintaining a valid error rate when the structure of the trial itself is changing dynamically.

Lalam: It sounds like they are addressing a gap where previous methods either controlled the false discovery rate or only looked at specific types of shared control, but not this general online environment where arms are added unpredictably.

The paper's summary: Tom: The authors summarize that platform trials naturally create an online multiple testing procedure because the total number of hypotheses is unknown and they are tested sequentially over time.

Jane: They emphasize that family-wise error rate control is the standard requirement for confirmatory clinical trials, so this work is important because it provides a method to achieve that under these dynamic conditions.

Lu: The proposed ADDIS-Spending procedure works by updating the error level of an individual hypothesis at every interim analysis whenever a decision about another hypothesis is made, which prevents that error rate from getting inflated at later stages.

Meng: It seems they are using a specific rule involving a conflict set to decide which hypotheses are being tested and which ones are being discarded or kept for later testing.

Lalam: So, the summary points toward an adaptive approach where the testing level for one hypothesis changes based on what has happened with others in the trial timeline.

The paper's improvements: Tom: The main improvement they suggest is moving away from static alpha spending procedures, which usually have a fixed sequence like a gamma sequence that sums to one, because that just doesn't work well when new arms are entering.

Jane: They introduce the ADDIS-Spending procedure which improves power by using the fraction of non-null hypotheses and specifically exploiting those null hypotheses that turn out to be strictly non-significant.

Lu: The paper claims that this group sequential approach is more powerful than the usual alpha spending because it constrains the alpha cost to specific hypotheses using parameters like lambda and tau, which are different from a simple fixed sequence.

Meng: I'm interested in the claim about power gains when there's a high proportion of non-null hypotheses; that suggests this method is particularly useful when you have many potential treatments to test.

Lalam: And they suggest that there should exist specific values for lambda and tau where the power of this ADDIS-Spending procedure actually beats the performance of a standard alpha spending procedure with the same sequence.

Conclusion: Tom: So, to wrap up, this paper presents the group sequential ADDIS-Spending algorithm as a flexible way to manage multiplicity in online multi-stage platform trials while ensuring FWER control.

Jane: Essentially, it’s an improvement over static methods because it adapts the error level of each hypothesis based on decisions made about other hypotheses at earlier time points.

Lu: The implication is that you can have more efficient interim analyses and better power when you are adding treatments sequentially, even when the total number of arms remains unknown initially.

Meng: From a deployment standpoint, this means you could potentially run longer or test more arms while still keeping the error rate under control without needing perfect upfront planning for every single arm.

Lalam: This whole paper on "Online Control of the Family-wise Error Rate (FWER) for Multi-stage Platform Trials" shows that adaptive methods can provide a more robust way to handle sequential hypothesis testing in these complex trial designs.

Jiangyue Yao, David S. Robertson, Thomas Jaki

Faculty of Informatics and Data Science, University of Regensburg · MRC Biostatistics Unit, University of Cambridge

stat.ME

Submitted: 2026-11-04

Updated: 2026-11-04

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 91/100

The gist: In platform trials, multiple intervention arms can enter or leave the study at any time, which brings the challenge of how to control the type I error rate In a platform trial…

Key concepts

Platform Trials
These are clinical trials where multiple intervention arms can enter or leave the study at any point during the trial. This introduces complexity because the total number of possible treatments is unknown initially, making standard error control difficult.
Family-wise Error Rate (FWER)
This is a statistical measure that controls the probability of making at least one Type I error (false positive) across all multiple hypotheses tested simultaneously. Controlling FWER ensures that the overall risk of falsely claiming a treatment effect is kept low.
Group Sequential ADDIS-Spending
This is an adaptive method where the testing level (alpha) for each hypothesis changes dynamically during interim analyses. It uses a conflict set to manage hypotheses without decisions, ensuring that the error rate remains controlled even as new arms are added or data accumulates.

Terminology

Summary

In platform trials, multiple intervention arms can enter or leave the study at any time, which brings the challenge of how to control the type I error rate In a platform trial, multiple intervention arms can enter or leave the study at any time, which brings the challenge of how to control the type I error rate. p. 0

Where the paper says it

Online Control of the Family-wise Error Rate (FWER) for Multi-stage Platform Trials Jiangyue Yao1 , David S. Robertson2 , and Thomas Jaki1,2 1Faculty of Informatics and Data Science, University of Regensburg, Regensburg, Germany 2MRC Biostatistics Unit, University of Cambridge, Cambridge, UK November 5, 2026 Abstract In platform trials, multiple intervention arms can enter or leave the study at any time, which brings the challenge of how to control the type I error rate.

The ADDIS-Spending (ADaptive DIScarding Alpha-spending) procedure controls the FWER in an online setting

Problem Definition and Motivation

Platform trials present a challenge regarding error rate control because multiple hypotheses are tested sequentially over time, and the total number of hypotheses is unknown The paper addresses the lack of a method that controls the FWER and allows new arms to be added at any time without prior information While standard two-arm trials focus on a single intervention, platform trials allow multiple treatments to enter at any time point, introducing multiplicity problems since the total number of arms is not known initially. The family-wise error rate (FWER) control is usually the standard requirement in a confirmatory clinical trial. p. 1

Where the paper says it

[9] investigated the possible multiplicity issues of shared control for platform trials. The family-wise error rate (FWER) control is usually the standard requirement in a confirmatory clinical trial.

Proposed Method Development

The authors propose the group sequential ADDIS-Spending algorithm to enable interim analyses for each hypothesis in platform trials This method updates the error level of an individual hypothesis at all (interim) analyses whenever new decisions on other hypotheses are made. The procedure involves defining a conflict set to address the issue where hypotheses without decisions at the current analysis are considered conflicting The proposed group sequential ADDIS-Spending procedure is defined by the rule that the testing level αi,j is predictable under the filtration F−Nt(i,j) and satisfies X i∈S∖C αdi τ − λ ≡ X i∈N αdi τ − λ (Sdi − Cdi) ≤ α for all Hi. p. 5

Where the paper says it

testing level αi,j is predictable under the filtration F−Nt(i,j) (defined above) and X i∈S\C αdi τ − λ ≡ X i∈N αdi τ − λ (Sdi − Cdi ) ≤ α for all Hi, 0 ≤ λ < τ ≤ 1 and αi,j < τ . Any sequence of testing level αi,j that satisfies the conditions is referred to as a group sequential ADDIS-Spending procedure.

Simulation Results and Performance

Simulation results show that the proposed method achieves valid FWER control and can offer power gains at a comparable sample size The performance of the group sequential ADDIS-Spending procedure depends on trial characteristics, especially the maximum number of comparisons that can be admitted to a platform, the percentage and the effect sizes of non-null hypotheses. The proposed approach has a better performance on power when the proportion of non-null hypotheses is high and the maximum number of allowed hypotheses is large

Comparative Analysis

The simulation study compares GS-ADDIS with classic Bonferroni correction, group sequential Bonferroni correction, and a pre-planned approach proposed by Greenstreet et al. [15] The GS-ADDIS procedure is more powerful than the usual alpha-spending procedure because it constrains the alpha cost to particular hypotheses with parameters λ and τ The stability of power of our procedure with respect to the maximum number of arms is particularly important for a platform trial, where the number of hypotheses is unknown in the beginning. p. 5

Where the paper says it

Then k(i, j) ≥ r(i) which implies γk(i,j) ≤ γr(i) . The group sequential ADDIS-Spending approach is more powerful than the usual alpha-spending procedure in the sense that it constrains the alpha cost to particular hypotheses with parameters λ and τ .

In conclusion, the proposed extension of the ADDIS-Spending algorithm provides a flexible and efficient approach to multiplicity adjustment in online multi-stage platform trials. p. 3
Where the paper says it

2 Proposed group sequential ADDIS-Spending approach Unlike the usual MAMS design where all hypotheses will be tested at the same calendar time, there is no fixed order of analysis times in platform trials. Both the sequence of hypotheses and the sequence of their interim analyses need to be considered in a multi-stage platform trial.

References

[1] Janet Woodcock and Lisa M LaVange. Master protocols to study multiple therapies, multiple diseases, or both. p. 0

Where the paper says it

1 Introduction Platform trials, as well as basket and umbrella trials, have recently attracted increasing interest as innovative and efficient designs due to their efficiency and flexibility. They all fall under the concept of master protocols, in which multiple questions concerning the same disease or biomarker are addressed within a unified study [1].

New England Journal of Medicine, 377(1):62–70, 2017

[2] Sharon-Lise T Normand. The RECOVERY platform. p. 0

Where the paper says it

Online Control of the Family-wise Error Rate (FWER) for Multi-stage Platform Trials Jiangyue Yao1 , David S. Robertson2 , and Thomas Jaki1,2 1Faculty of Informatics and Data Science, University of Regensburg, Regensburg, Germany 2MRC Biostatistics Unit, University of Cambridge, Cambridge, UK November 5, 2026 Abstract In platform trials, multiple intervention arms can enter or leave the study at any time, which brings the challenge of how to control the type I error rate.

New England Journal of Medicine, 384 (8):757–758, 2021

[3] James Wason, Dominic Magirr, Martin Law, and Thomas Jaki. Some recommendations for multi-arm multi-stage trials. p. 0

Where the paper says it

In this paper, we propose an approach to address the multiplicity issue in platform trials. While standard two-arm randomized clinical trials focus on a single intervention, multi-arm-multistage (MAMS) design allows the comparison of multiple treatments with a shared control.

Statistical Methods in Medical Research, 25(2):716–727, 2016

[4] Dominic Magirr, Thomas Jaki, and John Whitehead. A generalized Dunnett test for multi-arm multi-stage clinical studies with treatment selection. p. 0

Where the paper says it

In this paper, we propose an approach to address the multiplicity issue in platform trials. While standard two-arm randomized clinical trials focus on a single intervention, multi-arm-multistage (MAMS) design allows the comparison of multiple treatments with a shared control.

Biometrika, 99(2):494–501, 2012

[5] Thomas Jaki. Multi-arm clinical trials with treatment selection: what can be gained and at what price? p. 0

Where the paper says it

In this paper, we propose an approach to address the multiplicity issue in platform trials. While standard two-arm randomized clinical trials focus on a single intervention, multi-arm-multistage (MAMS) design allows the comparison of multiple treatments with a shared control.

Clinical Investigation, 5(4):393–399, 2015

[6] Nigel Stallard, Susan Todd, Deepak Parashar, Peter K Kimani, and Lindsay A Renfro. On the need to adjust for multiplicity in confirmatory clinical trials with master protocols. p. 0

Where the paper says it

They all fall under the concept of master protocols, in which multiple questions concerning the same disease or biomarker are addressed within a unified study [1]. Platform trials have especially come to the forefront of clinical research and contributed significantly to the development of treatments for COVID-19 [2].

Annals of Oncology, 30(4):506–509, 2019

[7] Xinping Cui, Emily Ouyang, Yi Liu, Jingjing Schneider, Hong Tian, Bushi Wang, and Jason C Hsu. Statistical principles for platform trials. p. 1

Where the paper says it

have discussed the benefits and challenges of conducting a platform trial, fewer have focused on the problem of controlling the overall type I error rate of the trial. [7] introduce some statistical considerations of error rate metrics for platform trials, where they argue that FDR control does not comply with current regulatory policy.

arXiv preprint arXiv:2302. p. 0
Where the paper says it

This design improves efficiency and flexibility in identifying new treatments, but it also introduces multiplicity problems, since the total number of arms is not known when the trial starts [6]. While many papers 1 arXiv:2611.

12728, 2023

[8] Xiaofei Bai, Qiqi Deng, and Dacheng Liu. Multiplicity issues for platform trials with a shared control arm. p. 1

Where the paper says it

[8] and Nguyen et al. [9] investigated the possible multiplicity issues of shared control for platform trials.

Journal of Biopharmaceutical Statistics, 30(6):1077–1090, 2020

[9] Quynh Nguyen, Katharina Hees, and Benjamin Hofner. Adaptive platform trials: the impact of common controls on type one error and power. p. 0

Where the paper says it

The ADDIS-Spending (ADaptive DIScarding Alpha-spending) procedure controls the FWER in an online setting. We propose the group sequential ADDIS-Spending algorithm to enable interim analyses for each hypothesis in platform trials.

Journal of Biopharmaceutical Statistics, 34 (5):719–736, 2024

[10] David S Robertson, James MS Wason, Franz König, Martin Posch, and Thomas Jaki. Online error rate control for platform trials. p. 0

Where the paper says it

Online Control of the Family-wise Error Rate (FWER) for Multi-stage Platform Trials Jiangyue Yao1 , David S. Robertson2 , and Thomas Jaki1,2 1Faculty of Informatics and Data Science, University of Regensburg, Regensburg, Germany 2MRC Biostatistics Unit, University of Cambridge, Cambridge, UK November 5, 2026 Abstract In platform trials, multiple intervention arms can enter or leave the study at any time, which brings the challenge of how to control the type I error rate.

Statistics in Medicine, 42(14):2475–2495, 2023

[11] David S Robertson and James Wason. Online control of the false discovery rate in biomedical research. p. 0

Where the paper says it

Interim analyses in a multistage design offer the opportunity of early stopping for success or futility in clinical trials by testing the same hypothesis at different time points. Some methods targeting false-discovery rate (FDR) or family-wise error rate (FWER) control for multi-stage platform trials have been proposed.

arXiv preprint arXiv:1809. p. 0
Where the paper says it

This design improves efficiency and flexibility in identifying new treatments, but it also introduces multiplicity problems, since the total number of arms is not known when the trial starts [6]. While many papers 1 arXiv:2611.

07292, 2018

[12] Sonja Zehetmayer, Martin Posch, and Franz König. Online control of the false discovery rate in group-sequential platform trials. p. 0

Where the paper says it

Interim analyses in a multistage design offer the opportunity of early stopping for success or futility in clinical trials by testing the same hypothesis at different time points. Some methods targeting false-discovery rate (FDR) or family-wise error rate (FWER) control for multi-stage platform trials have been proposed.

Statistical Methods in Medical Research, 31(12):2470–2485, 2022

[13] Adel Javanmard and Andrea Montanari. Online rules for control of false discovery rate and false discovery exceedance. p. 0

Where the paper says it

Interim analyses in a multistage design offer the opportunity of early stopping for success or futility in clinical trials by testing the same hypothesis at different time points. Some methods targeting false-discovery rate (FDR) or family-wise error rate (FWER) control for multi-stage platform trials have been proposed.

The Annals of Statistics, 46(2):526–554, 2018

[14] Jinjin Tian and Aaditya Ramdas. Online control of the familywise error rate. p. 0

Where the paper says it

Online Control of the Family-wise Error Rate (FWER) for Multi-stage Platform Trials Jiangyue Yao1 , David S. Robertson2 , and Thomas Jaki1,2 1Faculty of Informatics and Data Science, University of Regensburg, Regensburg, Germany 2MRC Biostatistics Unit, University of Cambridge, Cambridge, UK November 5, 2026 Abstract In platform trials, multiple intervention arms can enter or leave the study at any time, which brings the challenge of how to control the type I error rate.

Statistical Methods in Medical Research, 30(4):976–993, 2021

[15] Peter Greenstreet, Thomas Jaki, Alun Bedding, and Pavel Mozgunov. A preplanned multi-stage platform trial for discovering multiple superior treatments with control of FWER and power. p. 1

Where the paper says it

[10] and Robertson and Wason [11] discussed how to control the online error rate for platform trials but without considering multiple stages. It would be worthwhile to establish a method controlling FWER for a multi-stage platform trial.

Biometrical Journal, 67(1):e70025, 2025

[16] Thomas Burnett, Franz König, and Thomas Jaki. Adding experimental treatment arms to multi-arm multi-stage platform trials in progress. p. 0

Where the paper says it

Online Control of the Family-wise Error Rate (FWER) for Multi-stage Platform Trials Jiangyue Yao1 , David S. Robertson2 , and Thomas Jaki1,2 1Faculty of Informatics and Data Science, University of Regensburg, Regensburg, Germany 2MRC Biostatistics Unit, University of Cambridge, Cambridge, UK November 5, 2026 Abstract In platform trials, multiple intervention arms can enter or leave the study at any time, which brings the challenge of how to control the type I error rate.

Statistics in Medicine, 43(18):3447–3462, 2024

[17] Tijana Zrnic, Aaditya Ramdas, and Michael I Jordan. Asynchronous online testing of multiple hypotheses. p. 0

Where the paper says it

Robertson2 , and Thomas Jaki1,2 1Faculty of Informatics and Data Science, University of Regensburg, Regensburg, Germany 2MRC Biostatistics Unit, University of Cambridge, Cambridge, UK November 5, 2026 Abstract In platform trials, multiple intervention arms can enter or leave the study at any time, which brings the challenge of how to control the type I error rate. Online testing procedures, where a stream of hypotheses are tested over time and the total number of hypotheses is unknown, might be a viable choice for error rate control in platform trials.

Journal of Machine Learning Research, 22(33):1–39, 2021

[18] John Whitehead. The design and analysis of sequential clinical trials. p. 3

Where the paper says it

To address this issue, we proposed a group sequential approach and will introduce it in the next subsection. 2 Proposed group sequential ADDIS-Spending approach Unlike the usual MAMS design where all hypotheses will be tested at the same calendar time, there is no fixed order of analysis times in platform trials.

John Wiley & Sons, 2nd edition, 1997

[19] Vincent Jankovic, Lasse Fischer, and Werner Brannath. Asymptotic online FWER control for dependent test statistics. p. 5

Where the paper says it

When compared to other alpha-spending procedures, it is claimed that there must exist τ and λ such that the power of ADDIS-Spending procedure is greater than that of the alpha-spending procedure (with the same choice of γ sequence) [14]. However, notice that such an advantage depends on the multivariate distribution of the test statistics.

arXiv preprint arXiv:2401. p. 0
Where the paper says it

This design improves efficiency and flexibility in identifying new treatments, but it also introduces multiplicity problems, since the total number of arms is not known when the trial starts [6]. While many papers 1 arXiv:2611.

09559, 2024

Pruning redundant reasoning tokens

The paper introduces the group sequential ADDIS-Spending algorithm to manage multiplicity in online multi-stage platform trials while ensuring FWER control The initial thought for alpha spending involves an infinite non-negative sequence where γi is proportional to a function of the index i, such as γi ∝ i − 1.6 The ADDIS-Spending procedure improves power by utilizing the fraction of non-null hypotheses and exploiting strictly non-significant null hypotheses The group sequential approach is introduced to address issues arising from repeated interim analyses in multi-stage designs where new information becomes available at each stage. p. 2

Where the paper says it

The ADDIS-Spending procedure is a powerful method for online testing; however, it is not suitable for platform trials that involve not only multiple arms but also multiple stages. Problems arise during repeated interim analyses in multi-stage designs, where new information becomes available at each stage, leading to reallocation of the significance level.

How the training signal is built

The group sequential ADDIS-Spending procedure requires the testing level αi,j to be predictable under the filtration F−Nt(i,j) and satisfies X i∈S∖C αdi τ − λ ≡ X i∈N αdi τ − λ (Sdi − Cdi) ≤ α for all Hi The suggested significance level is given by the formula αi,j = α(τ − λ)γk(i,j), where k(i, j) indexes the γ value which utilizes all information until time t = t(i, j) The local boundaries for each stage are derived using group sequential design principles based on cumulative estimated treatment effects and current total sample size.

Where the gains come from

Simulation results indicate that the proposed method performs well across a variety of scenarios, though performance naturally depends on trial characteristics The group sequential ADDIS-Spending approach is more powerful than the usual alpha-spending procedure in that it constrains the alpha cost to particular hypotheses with parameters λ and τ It is claimed that there must exist τ and λ such that the power of ADDIS-Spending procedure is greater than that of the alpha-spending procedure (with the same choice of γ sequence). The GS-ADDIS procedure shows particular benefit when the proportion of non-null hypotheses is high The stability of power of our procedure with respect to the maximum number of arms is particularly important for a platform trial, where the number of hypotheses is unknown in the beginning. p. 1

Where the paper says it

In a platform trial, typically only a few treatments are included in the beginning and additional arms will be added later. This setting also reflects an online multiple testing procedure, where hypotheses are tested sequentially over time and the total number of hypotheses remains unknown.

The role of interim analyses is also noteworthy; they can substantially lower the sample size and improve trial efficiency. p. 2
Where the paper says it

However, the overall power of such an alpha-spending procedure is not convincing since the later hypotheses will be allocated to a relatively lower testing level. On that basis Tian and Ramdas [14] proposed the ADDIS-Spending procedure to improve the power performance.

Appendices

A Proof of FWER control provides the proof of Theorem 1 in Section 2.2 by construction of the stagewise-ordered p-values p′di, this is conditionally uniformly distributed under the null, and hence the conditional probability of rejecting hypothesis Hi at decision time di is strictly controlled at testing level αi,j The proof shows that P(V > 0) ≤ E[V] = X i∈H0 E h E 1≤X i∈H0 E [αdi] = X i∈H0 E αdi τ − λ (Sdi − Cdi) The example of updating the alpha level illustrates how the proposed procedure executes when hypotheses are ordered sequentially across stages The tables illustrate all potential boundaries for different hypotheses under various boundary shapes and stopping rules

Simulation Results of FWER control

Tables 12 and 13 outline the FWERs discussed in Section 3.1 when using parameters τ = 1 and λ = 0. p. 1

Where the paper says it

Let {γi}∞ i=1 be an infinite non-negative sequence that sums to 1 (P∞ i=1 γi = 1); the alpha level of each hypothesis can be then written as αi = αγi . If we assume a maximum number of n hypotheses in online testing, the alpha level for each hypothesis Hi would be αi = α n for i ≤ n, and αi = 0 for i > n using Bonferroni correction.

Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.

Improvements for AI systems

  1. Adaptive alpha allocation in online testing allows for updating error levels of individual hypotheses whenever new decisions on other hypotheses are made to ensure FWER control in multi-stage platform trials. This enables a more flexible and efficient approach compared to static alpha spending procedures like the usual alpha-spending procedure.

  2. The group sequential ADDIS-Spending procedure can be implemented to incorporate interim analyses for each hypothesis by updating alpha allocations across (interim) analyses whenever formal testing decisions are made. This addresses the issue where alpha level could have already been used up at earlier stages and inflate the error rate.

  3. The proposed method allows new arms to be added at any time without prior information by adapting the testing boundaries based on a conflict set of hypotheses that have not yet reached a decision, as described in A group sequential ADDIS-Spending procedure is defined by the following rule: the testing level alpha i,j is predictable under the filtration F−Nt(i,j).

  4. The system can utilize stage-wise ordering p-values to compute an adjusted significance level alpha i,j for hypothesis Hi at testing time point t(i, j), which is a group sequential ADDIS-Spending procedure ensuring that X i∈S∖C αdi τ − λ ≡ X i∈N αdi τ − λ (Sdi − Cdi) ≤ α for all Hi.

  5. The AI system can perform simulations across various platform trial scenarios, including half-overlapping and completely overlapping arms, to evaluate the performance of the GS-ADDIS procedure against classic Bonferroni and pre-planned approaches under conditions where the independence assumption does not hold.

Abstract

In platform trials, multiple intervention arms can enter or leave the study at any time, which brings the challenge of how to control the type I error rate. Online testing procedures, where a stream of hypotheses are tested over time and the total number of hypotheses is unknown, might be a viable choice for error rate control in platform trials. Interim analyses in a multi-stage design offer the opportunity of early stopping for success or futility in clinical trials by testing the same hypothesis at different time points. Some methods targeting false-discovery rate (FDR) or family-wise error rate (FWER) control for multi-stage platform trials have been proposed. However, a method which controls the FWER and allows new arms to be added at any time without prior information is yet to be established. The ADDIS-Spending (ADaptive DIScarding Alpha-spending) procedure controls the FWER in an online setting. We propose the group sequential ADDIS-Spending algorithm to enable interim analyses for each hypothesis in platform trials. The developed method can update the error level of an individual hypothesis at all (interim) analyses whenever new decisions on other hypotheses are made. We compare alternative approaches with our method in a simulation study. The proposed method yields competitive results in the conducted simulations.

Sources

Related papers