Objectives: The objective of this study was to investigate whether systematic reviews of mediation studies identify limitations in reporting quality and methodological conduct. Study Design and Setting: We conducted an overview of systematic reviews. We searched four databases (MEDLINE, PsycINFO, Cochrane Database of Systematic Reviews, and PubMed) to identify systematic reviews of studies that used mediation analysis to investigate mechanisms of health care interventions or exposures in clinical populations between 2007 and 2017. Two reviewers independently screened titles and abstracts. Summary data on the characteristics, reporting quality, and methodological conduct of the studies included in the systematic reviews were extracted independently by two reviewers. The protocol was prospectively registered on PROSPERO (CRD42017059834). Results: Fifty-four systematic reviews were included, representing 11 health care fields, 26 health conditions, and 2008 mediation studies. Eighteen of fifty-four systematic reviews (33%) explicitly stated that the reporting of primary studies was suboptimal. Of these, 14/18 (78%) reviews noted incomplete reporting of effect sizes and precision estimates from mediation analyses. Twenty-nine of fifty-four systematic reviews (54%) identified limitations in the methodological conduct of primary studies. Conclusion: The reporting and methodological conduct of studies investigating mechanisms in health care seems to be suboptimal. Guidance is needed to improve the quality, completeness, and transparency of mediation studies.
Bibliographical note© 2019, Elsevier. Licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International http://creativecommons.org/licenses/by-nc-nd/4.0/
Funding: Medical Research Council (grant no. APP1126767); National Institute for Health Research (NIHR) Collaboration for Leadership in Applied Health Research and Care Oxford at Oxford Health NHS Foundation Trust. National Institute of Health Research OUH, NHS Biomedical Research Centre.
- Causal inference
- Health care
- Mediation analysis
- Quality of reporting