Table 2.
GRADE evidence profile of Tai Chi plus conventional rehabilitation therapy vs. conventional rehabilitation therapy.
| Quality assessment | No of patients | Effect | Quality | Importance | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| No of studies | Design | Risk of bias | Inconsistency | Indirectness | Imprecision | Other considerations | Tai Chi plus conventional rehabilitation therapy | Control | Relative (95% CI) | Absolute | ||
| BBS (follow-up 2–24 weeks; measured with: the Berg Balance Scale; Better indicated by lower values) | ||||||||||||
| 7 | Randomized trials | Seriousa | No serious inconsistency | No serious indirectness | Seriousb | Reporting biasc | 166 | 162 | – | Not pooled | ⊕OOO VERY LOW |
CRITICAL |
| BI (follow-up 2–8 weeks; measured with: the modified Barthel Index; Better indicated by lower values) | ||||||||||||
| 6 | Randomized trials | Seriousa | No serious inconsistency | No serious indirectness | Seriousb | Reporting biasc | 165 | 160 | – | Not pooled | ⊕OOO VERY LOW |
CRITICAL |
| FMA (follow-up 4-8 weeks; measured with: the Fugl-Meyer Assessment; Better indicated by lower values) | ||||||||||||
| 2 | Randomized trials | Seriousa | No serious inconsistency | No serious indirectness | Seriousb | Reporting biasc | 51 | 49 | – | MD 4.49 higher (1.92 to 7.06 higher) | ⊕OOO VERY LOW |
CRITICAL |
| FMA-UE (follow-up 4-6 weeks; measured with: the Fugl-Meyer Assessment of upper extremity; Better indicated by lower values) | ||||||||||||
| 2 | Randomized trials | Seriousa | No serious inconsistency | No serious indirectness | Seriousb | Reporting biasc | 56 | 51 | – | MD 8.27 higher (4.69 to 11.84 higher) | ⊕OOO VERY LOW |
CRITICAL |
| FMA-LE (follow-up 4-6 weeks; measured with: the Fugl-Meyer Assessment of lower extremity; Better indicated by lower values) | ||||||||||||
| 3 | Randomized trials | Seriousa | No serious inconsistency | No serious indirectness | Seriousb | Reporting biasc | 85 | 81 | – | MD 2.75 higher (0.95 to 4.56 higher) | ⊕OOO VERY LOW |
CRITICAL |
| Holden (follow-up 4-8 weeks; measured with: Holden walking function score; Better indicated by lower values) | ||||||||||||
| 3 | Randomized trials | Seriousa | No serious inconsistency | No serious indirectness | Seriousb | Reporting biasc | 94 | 92 | – | MD 0.61 higher (0.38 to 0.85 higher) | ⊕OOO VERY LOW |
CRITICAL |
| TUGT (follow-up 6-24 weeks; measured with: To Up-and Go time; Better indicated by lower values) | ||||||||||||
| 4 | Randomized trials | Seriousa | No serious inconsistency | No serious indirectness | Seriousb | Reporting biasc | 100 | 96 | – | MD 2.59 higher (1.76 to 3.43 higher) | ⊕OOO VERY LOW |
CRITICAL |
a
Downgraded one level because of risk of bias: none of the studies applied double-blind design, most of whom also ignored to blind assessor, some didn't report the method used to generate the random allocation sequence or to achieve allocation concealment.
b
Downgraded one level because of imprecision: the total number of the participants is < 400.
c
Downgraded one level due to a possible publication bias: none had been enrolled on a platform for clinical trials, and no researcher agreed to share any more results except of those in their literatures, though some of them told us exactly the unpublished data existed.