Randomization

8. Sequence generation This section should describe the randomization method; for example if simple (pure) randomization was used, if blocked randomization was used and, if so, what block size was chosen, and if prognostic factors were taken into account, for example by stratification or minimization (see Chapter 4). It is not sufficient to state that 'computerized' randomization was used - the underlying method must be stated - but you should state if randomization lists were prepared in advance, or whether a computer program was used to produce allocations interactively. Also state here the randomization ratio.

9. Allocation concealment Concealment of the allocation is essential to ensure that those responsible for entering a patient into a trial will not know in advance what the patient will be allocated, and therefore be in a position to decide whether or not to enter them on that basis. Where randomization is effected by calling a central randomization service or other independent group, or where 'dynamic' randomization is used, it maybe clear that concealment is in place; where pre-prepared lists are used it would be necessary to state how they were prepared and stored until needed.

10. Implementation This point clarifies further, if necessary, who was responsible for preparing the randomization schedule, who was responsible for inviting patients to join the trial and discussing it with them and who ascertained the treatment allocation (i.e. made the telephone call or opened the envelope). It is important to distinguish preparation and performance. For example, for sealed envelope randomization, an independent individual should prepare the randomization list and the accompanying sealed envelopes which, in due course, a participating clinician will open. If the preparation or performance of randomization was carried out by someone involved in delivering the intervention this may be questionable.

11. Blinding (masking) Describe whether or not interventions or assessments were blinded to patients, clinicians, assessors or analysts. Also, describe steps taken to ensure any blinding was maintained, and any problems which you were aware of.

12. Statistical methods Describe the methods of analysis used in general for specific outcome measures, or specific types of data, but remember that you may also need to indicate the methods used when describing the results (for example whether parametric or non-parametric methods were used for a given outcome). Explain here how you dealt with complications such as missing data. When discussing additional analyses, mention the use (or not) of adjusted p-values for multiplicity of comparisons if appropriate.

0 0

Post a comment