Table 2.

Certainty analysis using GRADE

Certainty analysis
Outcome	No. of studies	Study design	Risk of bias	Inconsistency	Indirectness	Imprecision	Publication bias	Certainty
Anti-depressant effect	5	Randomised controlled trial	Not serious	Seriousa	Not serious	Seriousb	Strongly suspectedc	⊕⊝⊝⊝ Very Low
Anti-anxiety effect	3	Randomised controlled trial	Not serious	Seriousa	Not serious	Seriousb	Strongly suspectedc	⊕⊝⊝⊝ Very Low

^aDown-graded one level for inconsistency: I² value represents high heterogeneity (>50%)

^bDown-graded for imprecision due to lower sample size for standardised mean difference (< 100/group)

^cDown-graded one level for publication bias: Funnel plot was not created because of low study numbers (<10); however, leave one out sensitivity analysis changes the significance