• Jean Raymond, Tim E Darsaut, and David J Roy.
• Department of Radiology, Centre hospitalier de l'Université de Montréal, Montreal, Quebec, Canada. Electronic address: jean.raymond@umontreal.ca.
• World Neurosurg. 2017 Feb 1; 98: 403-410.

ObjectiveRandomized allocation of treatment options is not well accepted within the clinical community. Some methods of implementation may be received more favorably than others. Prerandomization may be an acceptable means to facilitate recruitment in some clinical trials.MethodsWe first compare randomization and prerandomization using illustrative neurovascular trials. We review some problems with conventional trials, Zelen's prerandomization as an alternative method, and the ethical issues that have surrounded prerandomization since its inception in a historic trial. Conventional and Zelen's randomization are then compared with other means to provide and verify care in the context of clinical uncertainty.ResultsThe major problem with conventional randomization is that consent is requested for 2 management options, one of which the patient will not receive. The problem with prerandomization is that treatment is allocated before the patient has consented to trial participation. Prerandomization may trade recruitment difficulties for excessive crossovers. However, other ways to practice under uncertainty and verify patient outcomes, such as case series and registries, are more ethically and scientifically problematic.ConclusionsUntil the ethical functions of randomized allocation of selected treatment options in the care of patients are recognized by the neurovascular community, Zelen's prerandomization may help recruitment into difficult trials and contribute a means to provide best possible care in the presence of uncertainty.Copyright © 2016 Elsevier Inc. All rights reserved.

18 keywords...

hide…