Sequence generation

Item 21a: Who will generate the random allocation sequence and the method used.

Example

“To ensure fairness, an independent statistician uses SPSS.25 software to generate randomized allocations for eligible patients. . . The statistician has no involvement in evaluating or executing the experiment" [319].

​

“One of the leading investigators (TMS) will generate the allocation sequences using a random number generator in Excel" [320].

​

“A statistician, not involved in the analysis of the trial results, will prepare the randomisation schedule. The randomisation schedule will be created using computer-generated random numbers before the first participant has been recruited, in a one-to-one ratio" [321]. 

​

Explanation

Who generated the random allocation sequence is important for two main reasons. First, someone, or some group, should take responsibility for this critical trial function. Second, providing information on the generator might help readers to evaluate whether anyone else had access to the allocation sequence during implementation (Item 21b).

Participants should be assigned to comparison groups in the trial on the basis of a chance (random) process characterised by unpredictability. Successful randomisation in practice depends on two interrelated aspects: adequate generation of an unpredictable allocation sequence and concealment of that sequence until assignment occurs. A key issue is whether the schedule is known or predictable by the people involved in allocating participants to the comparison groups. The treatment allocation system should thus be set up so that the person enroling participants does not know in advance which group assignment the next person will receive, a process termed allocation concealment. Proper allocation concealment shields knowledge of forthcoming assignments, whereas proper random sequences prevent correct anticipation of future assignments based on knowledge of past assignments.

Examples of adequate random sequence generation methods include the use of a computerised random number generator or a random number table. If the random sequence will be computer generated, we recommend that the software used be specified. Randomisation decreases selection bias in allocation, helps to facilitate blinding after allocation, and enables the use of probability theory to test whether any difference in outcome between trial groups reflects chance [322]. 

​

​Use of terms such as “randomisation” without further elaboration in the protocol is not sufficient to describe the allocation process, as these terms have been used inappropriately to describe non-random, deterministic allocation methods such as alternation or allocation by date of birth [323]. In general, these non-random allocation methods introduce selection bias and biased estimates of an intervention’s effect size [324-329]. Bias presumably arises from the inability to adequately conceal (Item 22) these more predictable, non-random sequence generation systems.

Three-quarters of randomised trial protocols approved by a research ethics committee in Denmark in 1994–1995, and a US cooperative cancer research group in 1968–2006, did not describe the method of sequence generation [330, 331]. More recent studies reported improved reporting. Reviews of two samples of 108 and 292 trial protocols from 2016 found that 75% and 61% respectively addressed the method of sequence generation [9, 10]. 

​

Summary of key elements to address

• Who will generate the allocation sequence

• Method of sequence generation (e.g., computerised random number generator)

• Any software used