|
Methodology of meta-analysis |
| Protocol development |
| Definition of population, intervention and outcomes |
| Specification of data sources and selection criteria |
| Search strategy |
| Adequate comprehensiveness |
| Sources and techniques |
| Language restrictions |
| Study selection |
| A priori criteria |
| Justification of exclusions |
| Methodologic quality assessment |
| Basis for inclusion, explanation of heterogeneity or sensitivity |
| analyses |
| Minimization of bias |
| Quality scoring system |
| Data extraction |
| Specification of data to be extracted |
| Avoidance of bias |
| Inclusion of all relevant data |
| Analysis |
| Appropriate model (fixed versus random effects) |
| Effect size (eg relative risk, odds ratio, risk difference) |
| Precision (confidence intervals) |
| Evaluation of heterogeneity |
| Subgroup analyses |
| Sensitivity analyses |
Sibbald Critical Care 2000 4(Suppl 2):S8 doi:10.1186/cc970 |