Investigators Androgen Receptor Antagonist supplier must then contact this person to enrol a new participant in the study and be informed of the next allocation. For an example, see the trial of exercise with incorporated breathing techniques for people with cystic fibrosis by Reix and colleagues (2012). Independant assistance with randomisation can be purchased from

commercial randomisation services. Such services can offer 24-hour-a-day randomisation, which may be beneficial if participants need to be randomised at unpredictable hours, such as within two hours of an injury or upon admission to intensive care. Note that the method of generating the random allocation list is distinct from the method of concealment of allocation. It Small Molecule Compound Library is also important to recognise that the method of allocation concealment is distinct from blinding. A trial may blind participants, therapists, and assessors, but still fail to conceal the allocation list (eg, Saunders 1995). Even if a trial cannot be blinded, the allocation list should still be concealed for the reasons discussed above. Blocked randomisation can allow partial loss of concealment of the allocation list. A blocked randomisation list is comprised of blocks of allocations that maintain reasonable balance of the group sizes throughout recruitment. For example, a trial intended

to randomise 60 participants may use a list made up of 10 blocks of six allocations, with three treatment and three control allocations old randomly ordered within each block. This ensures that group sizes will be similar even if the trial stops recruiting early. A potential problem with blocking is that it can threaten concealment. If the trial is not blinded the enrolling investigators may recognise that the allocations occur in balanced blocks of six. Once the allocations to one group are used up within a block, the remaining allocations in that block can be predicted with certainty. This allows the enrolling investigator to know the upcoming allocation for a potentially large proportion of participants, exposing the

trial to the same problems described earlier. Fortunately, this is easily solved by randomly varying the size of the blocks. The exact size of blocks should not be made public in trial protocols or registers prior to completion of the trial. Concealed allocation is not mentioned in the published reports of many trials of physiotherapy interventions (Moseley et al 2011). This is disappointing because concealed allocation is easy to implement and quick to describe in the published report. In 2011, only 20% of all trials on the Physiotherapy Evidence Database (PEDro; www.pedro.org.au) reported having concealed allocation (Moseley et al 2011). However, it is encouraging that this percentage has been increasing since shortly after the issue was first described in the literature (Chalmers et al 1986).