Randomization allows assigning patients to a particular treatment by chance, balancing known and unknown confounders equally to the study arms. In these cases, the investigators can interpret the differences in outcomes caused by the intervention and not by chance. For example, patients with comorbidities are likely to have even worse outcomes than patients without. If all patients with diabetes and other patients with complications were moved to the conservative arm, it would unbalance the study and make the results from the operative arm appear even better than they are.
Effective randomization depends on adequate eligibility criteria for patients. It should be underscored that after designating a patient either to the experimental group or to the control group, they must remain in it until the end of the study, except when there are reasons justifying their removal.
It is important not only to say that the study was randomized but also to describe the randomization used to verify if it is adequate. The off-site randomization center reduces any in uence from the researchers. Any system that can be guessed (such as using patient dates of birth) or in uenced by an investigator who selects which enve- lope to pick to choose patients for their study, is not real randomization. When the envelope is the method used, it is recommended to provide a tamper-proof envelope to prevent having the treatment or diagnosis group chosen on purpose.
Randomization allows one to: 1) provide for compa- rable study groups; 2) balance known and unknown con- founders equally to each study arm; 3) remove investigator bias and maintain internal validity; and 4) interpret the differences in outcomes caused by the intervention and not by chance. The questions that need to be answered by the reviewers are: 1) are patients allocated randomly to trial groups; 2) is the randomization described and detailed enough to verify its adequacy; and 3) can you think of any way in which someone could get around the randomiza- tion process described in the article?