Table 3 GRADE table for guiding evaluation of quality of evidence and strength of the recommendation
Quality assessment | No of patients | Effect | Quality | Importance | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
No of studies | Design | Risk of bias | Inconsistency | Indirectness | Imprecision | Other considerations | NASG | Standard care | Relative (95% CI) | Absolute | ||
Maternal deaths - Non-Randomised clinical trials | ||||||||||||
5 | observational studies | no serious risk of bias | no serious inconsistency | no serious indirectness | no serious imprecision | none | 46/1274 (3.6%) | 72/1056 (6.8%) | RR 0.52 (0.36 to 0.77) | 33 fewer per 1000 (from 16 fewer to 44 fewer) | ⊕⊕ΟΟ LOW | CRITICAL |
2.3% | 11 fewer per 1000 (from 5 fewer to 15 fewer) | |||||||||||
Maternal deaths - Cluster Randomized Trials | ||||||||||||
1 | randomised trial | no serious risk of bias | no serious inconsistency | no serious indirectness | Serious1 | none | 4/405 (1%) | 11/475 (2.3%) | RR 0.43 (0.14 to 1.33) | 13 fewer per 1000 (from 20 fewer to 8 more) | ⊕⊕⊕Ο MODERATE | CRITICAL |
2.3% | 13 fewer per 1000 (from 20 fewer to 8 more) | |||||||||||
Severe Maternal Outcome (Severe Morbidity + Deaths) | ||||||||||||
4 | observational studies | no serious risk of bias | no serious inconsistency | no serious indirectness | no serious imprecision | none | 17/1167 (1.5%) | 44/1055 (4.2%) | RR 0.33 (0.19 to 0.57) | 28 fewer per 1000 (from 18 fewer to 34 fewer) | ⊕⊕ΟΟ LOW | IMPORTANT |
3.2% | 21 fewer per 1000 (from 14 fewer to 26 fewer) | |||||||||||
Severe Maternal Outcome (Severe Morbidity + Deaths) - Before and After Studies | ||||||||||||
3 | observational studies | no serious risk of bias | no serious inconsistency | no serious indirectness | no serious imprecision | none | 13/764 (1.7%) | 32/590 (5.4%) | RR 0.31 (0.17 to 0.59) | 37 fewer per 1000 (from 22 fewer to 45 fewer) | ⊕⊕ΟΟ LOW | IMPORTANT |
4.7% | 32 fewer per 1000 (from 19 fewer to 39 fewer) | |||||||||||
Severe Maternal Outcome (Severe Morbidity + Deaths) - Cluster Randomized Trials | ||||||||||||
1 | randomised trials | no serious risk of bias | no serious inconsistency | no serious indirectness | Serious1 | none | 4/403 (0.99%) | 12/465 (2.6%) | RR 0.38 (0.13 to 1.18) | 16 fewer per 1000 (from 22 fewer to 5 more) | ⊕⊕⊕Ο MODERATE | IMPORTANT |
2.6% | 16 fewer per 1000 (from 23 fewer to 5 more) | |||||||||||
Blood transfusion (ever) | ||||||||||||
5 | observational studies | no serious risk of bias | no serious inconsistency | no serious indirectness | no serious imprecision | none | 970/1467 (66.1%) | 791/1333 (59.3%) | RR 1.02 (0.92 to 1.12) | 12 more per 1000 (from 47 fewer to 71 more) | ⊕⊕ΟΟ LOW | CRITICAL |
71.1% | 14 more per 1000 (from 57 fewer to 85 more) | |||||||||||
Blood transfusion (ever) - Before and After Studies | ||||||||||||
4 | observational studies | no serious risk of bias | no serious inconsistency | no serious indirectness | no serious imprecision | none | 803/1069 (75.1%) | 623/898 (69.4%) | RR 0.99 (0.97 to 1.02) | 7 fewer per 1000 (from 21 fewer to 14 more) | ⊕⊕ΟΟ LOW | IMPORTANT |
75.9% | 8 fewer per 1000 (from 23 fewer to 15 more) | |||||||||||
Blood transfusion (ever) - Cluster Randomized Trials | ||||||||||||
1 | randomised trials | no serious risk of bias | no serious inconsistency | no serious indirectness | no serious imprecision | none | 167/398 (42%) | 168/435 (38.6%) | RR 1.09 (0.92 to 1.28) | 35 more per 1000 (from 31 fewer to 108 more) | ⊕⊕⊕⊕ HIGH | IMPORTANT |
38.6% | 35 more per 1000 (from 31 fewer to 108 more) | |||||||||||