Group Sequential Clinical Trials Research Articles

Discrete Boundaries in Group Sequential Clinical Trials

Methods have been devised for constructing discrete group sequential boundaries in interim monitoring of clinical trials by using type I error spending rate functions. This paper explores the behaviour of the existing type I error spending rate functions under the two more frequent patterns of group sequential analyses: early and late testing. Their properties are then used in the choice of the appropriate function for various situations. The design of the Scandinavian simvastatin survival study (SSSS) is used as an example for illustration. The results reveal that, in late interim analysis like the SSSS, the O'Brien-Fleming-type boundary is recommended to avoid unnecessary early stoppage of the trial. However, other less extreme type I error spending rate functions can be a better substitute for the O'Brien-Fleming-type boundary as they lead to less stringent early boundaries. In practice, the type I error spending rate function should be chosen before the collection of data to prevent data-directed selection bias and should not be changed during the course of the trial. The use function approach is especially flexible in enabling researchers to modify the frequency and timing of their interim analyses while the trial is in progress. Indeed, this is the main reason for and contribution of the type I error spending function approach.

An introduction to the use of interim data analyses in clinical trials

During a controlled clinical trial, data accumulate that contain information on the relative efficacy of the two treatments, yet these data often are not inspected or analyzed until the planned sample size has been reached and the trial has been terminated. This type of fixed-sample-size trial design, with a single terminal data analysis, has the ethical disadvantage that more patients than are necessary to obtain a reliable result may be randomized to the less efficacious treatment, whichever one that turns out to be. Thus, it often is desirable to schedule one or more analyses of the data, to be conducted before planned termination, to see if a reliable conclusion may be drawn from the data and the trial terminated early. Such analyses are called interim analyses. When interim analyses occur after each of severe relatively large groups of patients, the trial is called a group-sequential trial. Interim data analyses must be planned in advance to avoid increasing the risk of committing a Type I error and to achieve adequate power. This article introduces the statistical issues involved in the planning of interim data analyses and the design of group-sequential clinical trials.

Exact repeated confidence intervals for Bernoulli parameters in a group sequential clinical trial

This paper presents methods for constructing exact repeated confidence intervals (RCIs) for the success probability, p, of a single Bernoulli treatment and for the difference of success probabilities, Δ = p 1 − p 2, of two independent Bernoulli treatments in the context of a group sequential clinical trial. These RCIs calculated1at each interim analysis are useful for evaluating the data in light of all the information available rather than relying on rigid stopping criteria used by repeated significance tests. Extensions to construction of RCIs for the relative risk ϱ = p 1 p 2 and odds ratio ψ = p 1(1 − p 2) p 2(1 − p 1) are indicated.

Sample size determination for group sequential clinical trials with immediate response.

The use function approach to group sequential methods has been explored previously. Any group sequential design requires specifying the frequency and times of repeated analyses, but only the use function approach allows deviations from those specified in the design, without affecting the type I error in the analysis. This paper illustrates how the use function provides a simple and flexible design procedure and how the initially projected maximum sample size can be calculated for randomized clinical trials in which responses are known relatively soon after patient entry and for which there is an early stopping rule built into the study protocol. Also the consequence of using the proposed design procedure is investigated in terms of the operating characteristics of the subsequent group sequential analyses.

On the Choice of Times for Data Analysis in Group Sequential Clinical Trials

Planned interim analysis of randomized clinical trials has been implemented for over a decade. While the initial proposal advocated analyzing after equal numbers of patients were evaluated, a later modification by Lan and DeMets (1983, Biometrika 70, 659-663) allowed for more flexible boundaries. Rather than fixing the times of analysis at equal numbers of patients, they fixed the rate at which overall alpha was used up according to a use function alpha * (t) on t in with alpha * (0) = 0 and alpha * (1) = alpha. Here we consider how flexible Lan and DeMets' procedure is. We show that the choice of alpha * (t) for a particular trial affects the permissible analysis times if other desirable properties of the sequence of nominal significance levels are to hold. To overcome the difficulties posed by patterns of late analysis, piecewise linear convex use functions are proposed.

Study Duration for Clinical Trials with Survival Response and Early Stopping Rule

A comparative clinical trial with built-in sequential stopping rules allows earlier-than-scheduled stopping, should there be a significant indication of treatment difference. In a clinical trial where the major outcome is time (survival time or response) to a certain event such as failure, the design of the study should determine how long one needs to accrue patients and follow through until there is a sufficient number of events observed during the entire study duration. This paper proposes a unified design procedure for group sequential clinical trials with survival response. The time to event is assumed to be exponentially distributed, but the arguments extend naturally to the proportional hazards model after suitable transformation on the time scale. An example from the Eastern Cooperative Oncology Group (ECOG) is given to illustrate how this procedure can be implemented. The same example is used to explore the overall operating characteristics and the robustness of the proposed group sequential design.

Analyses of group sequential clinical trials

In the first part of this article the methodology of group sequential plans is reviewed. After introducing the basic definition of such plans the main properties are shown. At the end of this section three different plans (Pocock, O'Brien-Fleming, Koepcke) are compared. In the second part of the article some unresolved issues and recent developments in the application of group sequential methods to long-term controlled clinical trials are discussed. These include deviation from the assumptions, life table methods, multiple-arm clinical trials, multiple outcome measures, and confidence intervals.

Design of Group Sequential Clinical Trials with Multiple Endpoints

Abstract Group sequential testing for randomized clinical trials designed with multiple endpoints is considered. Previously computed tables for single endpoints are still useful, with a change in interpretation of certain parameters. The advantage in setting sample size based on multiple endpoints is that the sample size required when a trial is designed using more than one endpoint is smaller than the sample size based on any single endpoint, if the two calculations are undertaken using matching significance levels and powers. An example is provided, and possible extensions of the work are discussed.

Design of Group Sequential Clinical Trials with Multiple Endpoints

Design of Group Sequential Clinical Trials with Multiple Endpoints

Discussion of the Paper by Jennison and Turnbull

Discussion of the Paper by Jennison and Turnbull

Exact confidence intervals following a group sequential trial: A comparison of methods

With interest in designing group sequential clinical trials increasing, methodology for analysing the data arising in such trials is needed. We discuss several approaches to constructing confidence intervals for the parameter of interest in a group sequential trial after the trial has ended. We compare these methods to one already in the literature and point out interesting differences.

Design of group sequential clinical trials with survival endpoints: Kyungmann Kim Department of Biostatistics, Harvard School of Public Health, Boston, Massachussetts(S01)

Design of group sequential clinical trials with survival endpoints: Kyungmann Kim Department of Biostatistics, Harvard School of Public Health, Boston, Massachussetts(S01)

Repeated confidence intervals for group sequential clinical trials

We describe methods for the construction of valid confidence intervals for parameters of interest at repeated times during the course of a clinical trial with two treatments. This approach gives the investigator a way to monitor the trial and allows greater flexibility than methods that have rigid statistical stopping rules. Two situations are considered. In the first, responses are available immediately and normally distributed, the parameter of interest being the difference in means. In the second, observations are survival times and a proportional hazards model is assumed. Here the parameter of interest is the hazard ratio. Group sequential tests can be based on repeated confidence interval methods and these are compared to other multiple testing procedures that have been proposed.

Correction: Group Sequential Clinical Trials with Triangular Continuation Regions

Correction: Group Sequential Clinical Trials with Triangular Continuation Regions

Group Sequential Clinical Trials with Triangular Continuation Regions

In this paper a new class of group sequential procedures for clinical trials is introduced, and the use of these procedures is illustrated by reference to a recently completed comparative study. In a group sequential trial the decision to stop or to continue is made at regular intervals throughout the trial, but not as frequently as after every patient response. This more practical formulation retains most of the advantages of sequential analysis, particularly the economy in sample size. Comparisons are made with group sequential designs derived from the repeated significance test.