Jump to comment:

• The "technique" of allocation concealment?
  Vance W Berger
  Published on: 22 March 2016

• Published on: 22 March 2016

  The "technique" of allocation concealment?

  • Vance W Berger, Biostatistician

  Dear Editor,

  In the recent piece “Evidence-Based Child Health 2: Understanding Randomized Controlled Trials”, Akobeng demonstrated a fundamental misunderstanding which has caused some of the advice given to be misleading and incorrect. Moreover, this misunderstanding has been adequately covered in the literature, and so a more diligent search would have prevented the error. Instead, your readers are now being told...

  Show More

  Dear Editor,

  In the recent piece “Evidence-Based Child Health 2: Understanding Randomized Controlled Trials”, Akobeng demonstrated a fundamental misunderstanding which has caused some of the advice given to be misleading and incorrect. Moreover, this misunderstanding has been adequately covered in the literature, and so a more diligent search would have prevented the error. Instead, your readers are now being told that “allocation concealment is a technique” which, unlike masking, is always possible.

  In fact allocation concealment is not a technique. Rather, it is a state of affairs that can be quantified based on the extent to which future allocations can be predicted.[1,2] There are techniques that can be used to help allocation concealment along, but this is hardly the same as “doing the allocation concealment” thing. Some of these techniques, such as using sealed envelopes, may in fact always be possible, but, as Akobeng points out, this does not always result in the concealment that was desired.

  To complicate matters, Akobeng’s misunderstanding means that your readers are now being told that to ensure allocation concealment, all one needs to do is ensure that no investigators have access to the allocation sequence prior to the execution of randomization. In fact, nothing could be further from the truth, as the block randomization that Akobeng implicitly endorses serves to undermine allocation concealment by allowing for future allocations to be predicted based on past ones, at least in unmasked trials.[2,3] Indeed, a strong case could be made that the primary threat to allocation concealment is prediction, and not direct observation, of future allocations. With this in mind, it becomes quite clear that perfect allocation concealment requires any combination of 1) perfect masking (and how can this ever be demonstrated?); 2) the elimination of patterns in the allocation sequence such as would result from completely unrestricted randomization (but how often is this used in practice?); and 3) simultaneous as opposed to sequential randomization (how often are the patients all present at once?).[2,4] Short of these steps, allocation simply cannot be said to be possible all the time.

  To help ensure allocation concealment in randomized clinical trials, researchers can do more than simply wave a magic wand and claim to have attained allocation concealment nirvana. Specifically, options available aside from the obvious (conceal the allocations to the extent possible and mask the trial to the extent possible) include using less restricted randomization procedures such as the maximal procedure [2,3], and testing for allocation concealment after the trial is completed.[2,5] Researchers might also do well to avoid the trap that Akobeng stepped into, and think through what allocation concealment really means, and how to get there, rather than simply repeating back what was heard or read in other misguided presentations on the subject.

  References

  1. Berger, VW, “Is Allocation Concealment a Binary Phenomenon?”, Medical Journal of Australia 183, 3, 165.

  2. Berger, VW (2005), “Selection Bias and Covariate Imbalances in Randomized Clinical Trials”, John Wiley & Sons, Chichester.

  3. Berger, VW, Ivanova, A, Deloria-Knoll, M (2003). “Minimizing Predictability while Retaining Balance through the Use of Less Restrictive Randomization Procedures”, Statistics in Medicine 22, 19, 3017-3028.

  4. Berger, VW, Christophi, CA (2003). “Randomization Technique, Allocation Concealment, Masking, and Susceptibility of Trials to Selection Bias”, JMASM 2, 1, 80-86.

  5. Berger, VW, Exner, DV (1999). “Detecting Selection Bias in Randomized Clinical Trials”, Controlled Clinical Trials 20, 319-327.

  Show Less

  Conflict of Interest:
  None declared.

  • Back to top