|
Literature Search:
|
| • Specification of efforts to search for direct and indirect evidence of relevance |
| • Presentation of search terms and the full search strategy (or strategies if separate searches undertakenfor each comparison) for one electronic database |
| • Provision of information regarding involvement of primary literature searching versus use of existingsystematic reviews as a means of including studies |
| • If existing reviews were used, specification of how these were located and description of theirrelated inclusion criteria |
|
Specification of Eligibility Criteria (and Planned Network Structure):
|
| • Specification of PICOS eligibility criteria for the review, including specification of all treatmentsincluded in the planned meta-analysis |
| • Specification of how related but different implementations of the same agent (e.g. varied doses ofpharmacologic treatments) are to be handled with associated rat ionale (i.e. address ‘lumping and splitting’of interventions) |
|
Specification of Assumptions for Network Meta-Analysis:
|
| • Specification of the assumptions of homogeneity, similarity and consistency (or related terminologyused, e.g. transitivity, exchangeability) |
| • Specification of efforts taken by reviewers to evaluate the assumption appropriateness |
| • Specification of what information is being provided to readers to allow them to consider the validityof the assumptions |
|
Statistical Approach to Analysis:
|
| • Specification of details of the approach to statistical analysis taken: hierarchical model, adjustedindirect approach or meta-regression, Frequentist or Bayesian framework, fixed or random effectsmodel, homogeneous or heterogeneous between-trial variance structure |
| • Specification of methods used to assess the degree of statistical heterogeneity and the potential forpublication bias within the treatment network |
| • Specification of methods used to evaluate for the presence of disagreement between direct and indirectevidence in the treatment network |
| • Description of statistical methods used to address clinical and methodologic homogeneity in theanalyses (e.g. subgroups, meta-regression including adjustments for baseline risk and the impactof risk of bias variations) |
