Читать книгу Randomised Clinical Trials - David Machin - Страница 88

The choice of statistical design will reflect the (major) hypotheses under test and the options for design will need to be discussed thoroughly amongst the design team and with other colleagues as appropriate. If not already included elsewhere, a schema of the trial is a useful addition to this section.

The important features here will include the number of interventions under test and the type of design, for example, parallel group or cross‐over trial with a superiority or non‐inferiority objective. If a parallel group design, then the allocation ratio should be specified and if this implies unequal numbers in the intervention groups, then a brief rationale for this should be included if not explained elsewhere. As appropriate, the standard (or control) intervention should be clearly identified. Further, if the design is to be stratified for one or more major prognostic factors this should be indicated. In addition, reference to the degree of blinding should be made with specific mention of ‘open’ trial if there are no masking mechanisms included. Importantly it should be made very clear that the trial is randomised and when assessments are to be made.