Table 1.

Credibility criteria for credible subgroup analyses.

Subgroup analysis credibility criteria	Description (from Saragiotto et al., 2016)
Is the subgroup variable a characteristic measured at baseline?	Subgroup variables measured after randomisation might be influenced by the tested interventions. The apparent difference of treatment effect between subgroups can be explained by the intervention, or by differing prognostic characteristics in subgroups that appear after randomisation.
Was the subgroup variable a stratification factor at randomisation?	Credibility of subgroup difference would be increased if a subgroup variable was also used for stratification at randomisation (i.e. stratified randomisation).
Was the hypothesis specified a priori?	A subgroup analysis might be clearly planned before to test a hypothesis. This must be mentioned on the study protocol (registered or published) or primary trial, when appropriate. Post-hoc analyses are more susceptible to bias as well as spurious results and they should be viewed as hypothesis generating rather than hypothesis testing.
Was the subgroup analysis one of a small number of subgroup analyses tested (≤5)?	The greater the number of hypotheses tested, the greater the number of interactions that will be discovered by chance, that is, the more likely it is to make a type I error (reject one of the null hypotheses even if all are actually true). A more appropriate analysis would account for the number of subgroups.
Was the test of interaction significant (interaction p < 0.05)?	Statistical tests of significance must be used to assess the likelihood that a given interaction might have arisen due to chance alone (the lower a P value is, the less likely it is that the interaction can be explained by chance).
Was the significant interaction effect independent, if there were multiple significant interactions?	When testing multiple hypotheses in a single study, the analyses might yield more than one apparently significant interaction. These significant interactions might, however, be associated with each other, and thus explained by a common factor.
Was the direction of the subgroup effect correctly pre-specified?	A subgroup effect consistent with the pre-specified direction will increase the credibility of a subgroup analysis. Failure to specify the direction or even getting the wrong direction weakens the case for a real underlying subgroup effect
Was the subgroup effect consistent with evidence from previous studies?	A hypothesis concerning differential response in a subgroup of patients may be generated by examination of data from a single study. The interaction becomes far more credible if it is also found in other similar studies. The extent to which a comprehensive scientific overview of the relevant literature finds an interaction to be consistently present is probably the best single index as to whether it should be believed. In other words, the replication of an interaction in independent, unbiased studies provides strong support for its believability.
Was the subgroup effect consistent across related outcomes?	The subgroup effect is more likely to be real if its effect manifest across all closely related outcomes. Studies must determine whether the subgroup effect existed among related outcomes.
Was there indirect evidence to support the apparent subgroups effect (biological rationale, laboratory tests, animal studies)?	We are generally more ready to believe a hypothesised interaction if indirect evidence makes the interaction more plausible. That is, to the extent that a hypothesis is consistent with our current understanding of the biologic mechanisms of disease, we are more likely to believe it. Such understanding comes from three types of indirect evidence: (i) from studies of different populations (including animal studies); (ii) from observations of interactions for similar interventions; and (iii) from results of studies of other related outcomes.