11 Results of the systematic review

The quality of the systematic review depends both on the quality of the individual studies and the aggregate characteristics of these studies. If the aggregate results are missing studies, contain predominantly poorly conducted studies, or are highly heterogeneous then this will likely warrant lower confidence in the results.

Checklist Questions

Do all inclusions & exclusions of trials make sense?
Are you aware of any relevant studies that were not identified/included in this review?
Did reviewers adequately assess individual trials for risk of bias?
Was each component reported separately, or summarized with a composite quality score?
Are there any differences between studies that should preclude meta-analysis?

Risk of bias within trials (internal validity): Did reviewers adequately assess for (& report) risk of bias?

Risk of bias should be evaluated by using a tool that is specific to RCTs. The Cochrane risk of bias tool (version 1 (Higgins JPT et al. 2011) or 2 (Sterne JAC et al. 2019)) evaluates the risk of individual trial biases and offers the most transparent assessment of trial internal validity (see NERDCAT-RCT for more information regarding internal validity). ROBIS-I (Sterne JA et al.) is a similar tool available for appraising risk of bias in observational trials.

Quality Scores

“Quality scores” such as the Jadad score are more closely related to reporting quality than methodological issues, and lead to wide variability in conclusions on “quality” based on the score used. In particular, the Jadad score is considered obsolete and is a poor measure of risk of bias.

Methodological & clinical heterogeneity: Is it appropriate to perform a meta-analysis?

  • Methodological heterogeneity: Are there methodological differences (e.g. risk of bias) between studies?
  • Clinical heterogeneity: Are there any differences in clinical characteristics between the individual trials (i.e. any component of PICO) that preclude pooling the trials together in a meta-analysis?
  • Is the impact of any of these characteristics tested in a subgroup analysis or meta-regression?

Testing possible sources of heterogeneity may identify causes for statistical heterogeneity identified in the meta-analysis (e.g. the intervention may only appear beneficial in trials at high risk of bias, but not in those at low risk).

See NERDCAT-RCT to learn more on how to appraise validity of subgroup effects.

definition

License

Icon for the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License

NERDCAT Copyright © 2022 by Ricky Turgeon is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License, except where otherwise noted.

Share This Book