Group Sequential Design (GSD)

Overview

In summary:

  • Group sequential trials were originally developed to evaluate one active treatment compared to a comparator allowing for early trial stopping; 
  • Based on accumulating outcome data from a group of recruited patients, the trial is stopped early when there is sufficient evidence to reach a conclusion;
  • The criteria for stopping a trial early (e.g., of efficacy, futility, or harm) and decision rules at interim analyses  as well as when decisions are made should be pre-specified;
  • Group sequential trials are useful when there is a strong need for quick clinical decisions and policy-making and there are opportunities to do so (e.g., outcome data quickly observed for interim analyses);
  • By stopping the trial early, the design can offer ethical, economic and time-saving benefits;
  • Group sequential methods have recently been extended to cover other adaptive designs (e.g., multi-arm multi-stage, MAMS designs).
 TO ADD VIDEO HERE 

Motivation

The conduct of clinical trials is generally time-consuming and expensive. Researchers test new treatments on patients during the trial to gather evidence of their benefits and harms. Most of these new treatments do not translate into clinical practice mainly because they are ineffective and/or unsafe 1, 2, 3, 4, 5. As a result, researchers spend most research resources and time studying new treatments for several years that would never be used to treat patients. A small proportion of new treatments that are effective and safe often take several years to develop and test before they are adopted in practice to benefit patients. 

It is, therefore, sensible to amend or even stop the trial if the data already collected is sufficient to answer the original  research questions. There are situations where a new treatment shows little or no benefit on the basis of the outcomes of patients already recruited (interim data), and that continuing the trial as originally planned is unlikely to show clinical benefits. This is known as  futility, meaning that there is little value of continuing if the conclusions are already clearly negative. On the contrary, interim data might show overwhelming evidence of benefit that is unlikely to be overturned even if researchers progress to completion. This also reduces the need to recruit more patients than necessary to address research questions. The incentives to reach early conclusions fit in with the interests of several key stakeholders such as patients, funders, regulators, researchers, and medical practitioners. This is even more appealing in emergency and pandemic situations (e.g., Ebola and COVID-19) to facilitate quick clinical decisions and policy-making. A group sequential design (GSD) offers this opportunity to stop a trial early when the opportunity arises while maintaining statistical rigour and the validity of results. 

In context

A GSD is the most commonly used adaptive design in routine practice 6, 11.  A review of trials 6 that used a GSD published in leading medical journals found that the majority 68% (46/68) stopped early: predominantly for futility, 61% (28/46), and efficacy, 22% (10/46). Other early stopping reasons, which are not unique to GSDs, were poor recruitment, financial, safety, and external information. 

References

1. Smietana et al. Trends in clinical success rates. Nat Rev Drug Discov. 2016;15(6):379–80.
2. Wong et al. Estimation of clinical trial success rates and related parameters. Biostatistics. 2018;20(2):1–14.
3. BIO. Clinical development success rates 2006-2015. Online Report. 2016.
4. Hussain et al. Clinical trial success rates of anti-obesity agents: the importance of combination therapies. Obes Rev. 2015;16(9):707–14.
5. Grignolo et al. Phase III trial failures : costly, but preventable. Appl Clin Trials. 2016;25(8).
6. Stevely et al. An investigation of the shortcomings of the CONSORT 2010 statement for the reporting of group sequential randomised controlled trials: A methodological systematic review. PLoS One. 2015;10(11):e0141104.