In the randomized clinical trial (RCT), the process of deciding the randomization method and implementing is critically important. Unfortunately, it is not unheard of for problems to arise. In an article (Downs et al 2010 1), it is noted that as well as initial errors of trial design, problems can arise from errors with programming of the randomization or even human error during the course of the trial. Maintaining the rigor of the RCT relies on robust and reliable randomization with no errors. If treatment allocation is inadequately concealed then overestimation of treatment effect can occur, and the ‘randomized’ control trial becomes effectively ‘non-randomized’ – putting the entire study at risk (2).
Randomization lists are often created by copying SAS code from one project to the next and adjusting the parameters accordingly. However more robust approaches are available. Built on a foundation of statistical and simulation expertise, Cytel’s FlexRandomizer software tool is used extensively by pharma companies, CROs and other organizations to provide secure, centralized randomization for clinical trials.
For a number of clients, Cytel also acts as an Independent Randomization Center. Under this model, Cytel statisticians work to create the randomization specifications and generate the lists (or in the case of some adaptive trials, the randomization algorithms) using Cytel’s FlexRandomizer, for all the clinical trials conducted by the sponsor. These validated lists are then usually provided to the client’s preferred Interactive Voice or Web Response System (IVRS or IWRS) vendor. On occasion, when the study is being conducted without an Interactive Response Technology (IRT), the list may be used to create sealed randomization envelopes for use in an emergency to unblind an individual subject.
Using this Independent Randomization Center model in combination with a preferred IVRS or IWRS technology provider to manage the randomization implementation and clinical trial supply allows pharmaceutical companies to derive key benefits:
- Most critically, the process is always led by an experienced statistician who produces the randomization specifications and oversees the production of the randomization list. This provides the reassurance of validation, and the chance of an error is reduced..
- Pharmaceutical companies may be creating a large volume of randomizations over the course of a year. The task lends itself well to the creation of a standard, streamlined approach- improving efficiency and again reducing the potential for error.
- Often, there is a tight turnaround required for the creation of the randomization list. Using a familiar team within a CRO provides flexibility and access to qualified resources to provide assurance the task will be conducted to a high level of quality.
It is useful to note that for studies where Cytel is performing the data management via an EDC system that has an integrated IWR module, Cytel is experienced building both the eCRF and the IWR forms and assigning an independent unblinded team at a separate geographic location to manage the randomization.
By leveraging both statistical capabilities and the technology solutions which are now available, it is possible to ensure randomizations are conducted securely, efficiently and correctly thereby putting the study on a solid statistical foundation.
With thanks to Caroline Morgan and Michael Kutch of Cytel.
References
1) Downs, M., Tucker, K., Christ-Schmidt, H. and Wittes, J. (2010). Some practical problems in implementing randomization. Clinical Trials, 7(3), pp.235-245.
2) Eliminating bias in randomized controlled trials: importance of allocation concealment and masking.
Viera AJ Bangdiwala SI.Fam Med. 2007 Feb;39(2):132-7
Liked this article? Join our global community of biopharma innovators and sign up for Cytel blog updates direct to your inbox.
