Quanticate Blog

Conducting Randomization in Clinical Trials [VIDEO]

Written by Statistical Consultancy Team | Fri, Dec 04, 2015

 

When running a clinical trial the industry standard is a double-blind placebo‑controlled parallel group trial. This is because it is the best way to ensure that the characteristics of subjects in each treatment group are the same, whilst ensuring the investigators cannot anticipate the treatment of a subject.

An integral part of the study, ensuring the success of the trial and the integrity of the results is the randomization of subjects into the trial. Randomization ensures that every patient entering a trial has an equal chance of receiving the active treatment. This minimizes potential bias that could be introduced into the clinical trial.

There are several factors that need to be taken into account when conducting randomization in clinical trials. All current documentation (e.g. protocol, protocol amendments) for the trial should be provided by the client. It is important to establish the requirements for the randomization at an early stage and develop a specification document to detail the format that the randomization will take. This allows the client to see the format that the randomization will take without unblinding them to the final randomization schedule.

The following must be clearly defined, documented and agreed with the client prior to the randomization being produced:

  • Proportion of subjects to be allocated to each treatment group
  • Stratification factors to be considered
  • Block size to use
  • Number of centres
  • How the seed will be defined
  • What outputs are required (e.g. schedule, randomization envelopes)
  • The format any outputs will take

Regular communication with the client in developing the specification document will help ensure the correct information is collected. Then the randomization information is sent to the correct places to ensure there is no accidental unblinding.

The team producing the randomization must not be involved in the reporting of the study as they will be unblinded to the treatment allocation and could cause bias in the results. We would advise where possible staff at different sites should be utilised. A dummy randomization may be produced to show the client the exact format and layout that the final outputs will take – these must clearly be labelled as draft to ensure there is no confusion with the final randomization.

Print randomization envelopes may also be produced to unblind study personnel in an emergency to the treatment allocation of a particular subject, for example, a subject may be experiencing a serious adverse event (SAE), and it is critical to find out what treatment they have received. Randomization envelopes allow unblinding of an individual subject, without having to unblind the entire study which could cost the sponsor significant amounts of money and time invested in the clinical trial.

Once a randomization has been produced and sent to the relevant staff as detailed in the specification document it is important to store the randomization correctly to ensure that no study staff can accidentally be unblinded to the treatment allocations. We recommend that there should be no information stored on computer networks. All randomization information should be stored in a secure facility which has limited access only for certain unblinded staff members. For example we use fireproof cabinets and controlled access with records of which statisticians are allowed access to the keys to these secure facilities.

Randomizations conducted correctly help to ensure the success of the trial and the integrity of the results. If the randomization of a clinical trial was to be performed incorrectly it could result in introducing bias into the study results and potentially the use of incorrect methodology. This could require additional analyses to gain regulatory approval for the drug.