The following diagram for a crossover trial does not use the crossing slanted lines sometimes used to represent crossover trials, because the order of the blocks is sufficient to represent the design of the trial. Slanted lines are used only to represent the branch point at randomization, when a subject is assigned to a sequence of treatments. As in most crossover trials, the arms are distinguished by the order of treatments, with the same treatments present in each arm. Note that even though all 3 arms of this trial end with the same block (i.e., the block for the follow-up element), the diagram does not show the arms converging into one block. Also note that the same block (the "rest" element) occurs twice within each arm. Elements are conceived of as "reusable" and can appear in more than 1 arm, in more than 1 epoch, and more than once in an arm.
The next diagram for this crossover trial shows the prospective view of the trial; it identifies the epoch and arms of the trial, and gives each a name. As for most crossover studies, the objectives of the trial will be addressed by comparisons between the arms and by within-subject comparisons between treatments. Because the design depends on differentiating the periods during which the subject receives the 3 different treatments, there are 3 different treatment epochs. The fact that the rest periods are identified as separate epochs suggests that these also play an important part in the design of the trial; they are probably designed to allow subjects to return to “baseline,” with data collected to show that this occurred. Note that epochs are not considered reusable; each epoch has a different name, even though all the treatment epochs are similar and both the rest epochs are similar. As with the first example trial, there is a one-to-one relationship between the epochs of the trial and the elements in each arm.
The next diagram shows the retrospective view of the trial.
The last diagram for this trial shows the trial from the viewpoint of blinded participants. As in the simple parallel trial in Example Trial 1, blinded participants see only 1 sequence of elements; during the treatment epochs they do not know which of the treatment elements a subject is in.
The following table illustrates the trial design matrix for this crossover example trial. It corresponds closely to the preceding retrospective diagram.
Trial Design Matrix
| Screen | First Treatment | First Rest | Second Treatment | Second Rest | Third Treatment | Follow-up | |
|---|---|---|---|---|---|---|---|
| P-5-10 | Screen | Usual Tobacco Product | Rest | Tobacco Product A | Rest | Tobacco Product B | Follow-up |
| 5-P-10 | Screen | Tobacco Product A | Rest | Usual Tobacco Product | Rest | Tobacco Product B | Follow-up |
| 5-10-P | Screen | Tobacco Product B | Rest | Tobacco Product A | Rest | Usual Tobacco Product | Follow-up |
It is straightforward to produce the TA dataset for this crossover trial from the diagram showing arms and epochs, or from the trial design matrix.
Overview
Content Tools