Subgroup analysis was done to look for the effect of age subgroups (adult: mean age ≥18 years; paediatric: mean age <18 years). Separate random-effects meta-analysis models were generated to estimate pooled prevalence estimates within these subgroups. Only those comorbidities that had at least 5 studies within the subgroups were meta-analysed. Subgroup effect is indicated using the R2 (indicates heritability explained by the moderator), I2 (residual heterogeneity of the meta-analysis after considering the moderator), and the Cochrane's QM statistic with its p-value. We also carried out meta-regression analysis with AAA, AOO, mean Y-BOCS/CY-BOCS total score (a measure of illness severity), and the percentage of male gender. We used the QM and corresponding p-value to test significance of the moderator.
