Domain 1: Randomization process
1. Was the allocation sequence random?
If a random component was used in the sequence generation process, reviewers should answer this question with “Yes/PY”. For example, computer-generated random numbers, random number tables, throwing a dice, or coin tossing.
2. Was the allocation sequence concealed until participants were enrolled and assigned to interventions?
Answer “Yes/PY” if there was a remote or centrally administered method of allocation, where the allocation process is organized by an external unit, organization, or researcher that is independent of the enrollment personnel. Answer “Yes/PY” if sealed envelopes were used.
3. Is the number of randomized participants balanced between groups?
Imbalance in the number of randomized participants can be examined through the proportion assigned to one group and its 99% confidence intervals. If the intervals do not cover the expected proportion (e.g. 50% in a 1:1 allocation), there is suspicion of baseline imbalance. The calculations needed for this item can also be performed at metapsy.org/rob/assistant/.
4. Is baseline clinical severity balanced between groups?
Imbalance between groups in baseline severity is examined by the standardized mean difference between groups at baseline and its 99% confidence intervals. The SMD > 0.2 (with 99% CI lower limit > 0) will be suggestive of baseline imbalance. We focus on baseline severity only because it is consistently reported across trials, can be examined objectively, and it is a known predictor of the outcome. P-values and confidence intervals resulting from multiple significance testing (e.g., due to multiple clinical severity measures) can be adjusted using Bonferroni corrections. Depending on the clinical field and mental health disorder, there might be key prognostic factors that should be added to this examination. In that case, the judgment of the reviewers can overwrite the algorithm and the proposed items and further variables can be examined for baseline imbalances. The calculations needed for this item can also be performed at metapsy.org/rob/assistant/.
