Table 3.
Quality of evidence included RCTs by Grading of Recommendations Assessment, Development, and Evaluation (GRADE).
| Outcomes | Included RCTs (patients) | Relative effect (95% CI) | Quality assessment | Quality of evidence | |||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Risk of bias | Inconsistency | Indirectness | Imprecision | Publication bias | Large effect | Dose response | All plausible confounding | ||||
| MMSE | 10 (592) | MD 1.22 (0.78, 1.66) |
Serious① | Serious② | Not serious | Not serious | Undetected | Undetected | Undetected | Undetected | Low |
| MoCA | 7 (339) | MD 1.22 (0.47, 1.97) |
Serious① | Serious② | Not serious | Not serious | Undetected | Undetected | Undetected | Undetected | Low |
| CMS | 2 (62) | MD 6.05 (-1.33, 13.44) |
Serious① | Not serious | Not serious | Serious③ | Undetected | Undetected | Undetected | Undetected | Low |
| ADL | 2 (120) | MD 3.19 (-4.53, 10.90) |
Serious① | Serious② | Not serious | Serious③ | Undetected | Undetected | Undetected | Undetected | Critically Low |
① Most information is from the moderate risk studies, and there are major limitations; ② The size and direction of the effect size, the overlap of the confidence interval (CI) is small, the P-value of the heterogeneity test is small, and the combined results of I2 value are large; ③ The sample is insufficient.