Table 2.

GRADE Summary of Findings for the effects of exercise interventions on health-related quality of life (HR-QoL) in adults with major depressive disorder.

Outcome	N of studies	Participants	Effect estimate (SMD, 95% CI)	Certainty of evidence (GRADE)
HR-QoL total	8	637	0.31 [0.16, 0.46]	Moderate ⊕⊕⊕◯
HR-QoL total, last follow-up	4	408	0.41 [0.13, 0.70]	Moderate ⊕⊕⊕◯
Overall HR-QoL domain	3	388	0.25 [0.05, 0.45]	Moderate ⊕⊕⊕◯
Physical HR-QoL domain	4	193	0.50 [0.10, 0.91]	Low ⊕⊕◯◯
Psychological HR-QoL domain	4	138	0.58 [0.24, 0.92]	Moderate ⊕⊕⊕◯
Social HR-QoL domain	2	70	0.11 [−0.34, 0.57]	Low ⊕⊕◯◯
Emotional HR-QoL domain	2	123	0.51 [0.11, 0.91]	Low ⊕⊕◯◯

a. Risk of bias: downgraded one level for all outcomes because all included studies had some concerns in the RoB 2 assessment, particularly in the measurement of outcomes domain due to reliance on self-reported HR-QoL and lack of blinding.

b. Inconsistency: not downgraded for the main outcomes (post-intervention and follow-up HR-QoL), as heterogeneity was low to moderate (I² = 0% and 34.42%, respectively). Domain-level analyses also showed no substantial inconsistency.

c. Imprecision: downgraded for physical, social, and emotional domains due to small sample sizes, limited number of studies, and/or wide confidence intervals. The social domain was further downgraded because the confidence interval included the null effect.

d. Indirectness: not downgraded, as the included studies directly addressed the target population (adults with MDD), intervention (exercise-based), comparator, and outcome (HR-QoL).

e. Publication bias: not downgraded, as funnel plots were symmetric and Egger’s and Begg’s tests did not indicate significant publication bias for the main outcomes.