Sample Size Re-estimation (SSR)

Conduct

 PANDA users may also wish to read more on general data management considerations for adaptive trials.

Measures to minimise operational bias

It is essential to ensure that the research team members who are blinded to the treatment allocation will remain so during the trial. Where possible, researchers involved in the running of the trial should be ‘masked’ to minimise their ability to guess correctly or work out the emerging treatment effect. Specific details of the method to estimate parameters may be documented in a restricted trial document when the trial is ongoing and only disclosed when the trial finishes. The roles and responsibilities of those involved in the SSR and decision-making process should be clear, including the information they should have access to. In some cases, especially for unblinded SSR, personnel external to the trial may be required to perform the SSR and recommend adaptation decisions (e.g. the involvement of an independent statistician and data monitoring committee). 

Thus, researchers should put in place and document clear processes and procedures that encompass:

  • the method for (re-)estimating design parameters; 
  • the personnel to undertake the SSR; 
  • the nature of data to be transferred to those undertaking the SSR and by whom; 
  • what results of the SSR will be communicated how and to whom; 
  • who will recommend or make a decision on the course of action after SSR and; 
  • what should be communicated to the research team.