Methodological quality and reporting standards in systematic reviews with meta-analysis of physical activity studies : a report from the Strengthening the Evidence in Exercise Sciences Initiative (SEES Initiative)

Abstract

Background: Several resources have been developed (e.g., reporting guidelines) to promote high-standard practices in health research. However, there was no continuous and systematic assessment of recommended practices in published systematic reviews with meta-analysis (SRMAs), which increases the usability of the available resources. Therefore, we aimed to assess the methodological and reporting standards in SRMAs of physical activity studies. This report presents the main results of the SEES Initiative in 2019. Methods: Our approach is based on a prospective systematic review methodology to implement post-publication surveillance of research practices in exercise sciences. Briefy, during the year 2019, pre-specifed searches were conducted monthly (PubMed/MEDLINE) in journals from the exercise sciences (n=9) and medicine (n=5). The assessments were independently conducted by two authors, based on 36 items/practices derived from established statements/tools (PRISMA, AMSTAR 2, ROBIS). To be eligible, SRMAs should summarize studies that had, at least, one arm consisting of physical activity interventions/exposures and one health or behavioral outcome. Results: Out of 1028 studies assessed for eligibility, 103 SRMAs were included. The minimum adherence was 13/36 items, whereas only one SRMA adhered to all items. Some highly contemplated items included identifcation of title as SRMA (97.1%) and descriptions of the main outcome in the abstract (95.1%) and risk of bias (RoB) assessment (95.1%). Some poorly contemplated items included publicly available protocol (4.9%), discussion of the results in light of RoB in studies included (32.0%), and data sharing statements (35.9%). Conclusion: In summary, there is a suboptimal adherence to recommended practices on methodological quality and reporting standards in the SRMAs of physical activity intervention/exposure evaluated from the selected journals in 2019, which likely reduce the reproducibility and usefulness of these studies. This incipient evidence from our frst 12 months of post-publication surveillance should serve as a call for attention and action for multiple stakeholders (e.g., authors, reviewers, editors, funders, academic institutions) in this important health research feld.