The methodological quality of dose-response meta-analyses needed substantial improvement: a cross-sectional survey and proposed recommendations.

To investigate methodological quality of published dose-response meta-analysis (DRMA) and explore study characteristics associated with the quality. We searched three databases for published DRMAs and used a modified AMSTAR (15 items) checklist to assess the methodological quality. We summarized the compliance of those DRMAs to the AMSTAR items and used multivariable regression analysis to explore the association between prespecified study characteristics with the overall methodological quality. We identified 529 DRMAs. Of the methodological quality items, six were well complied (80% or more) and six poorly complied (30% or fewer) by these DRMAs. The median score was nine points [first and third quartile: 7, 10] and only 64/529 had score over 10 points. Regression analysis suggested that studies with more authors (β = 0.19; 95% confidence interval [CI]: 0.05, 0.33), published more recently (β = 0.29; 95% CI: 0.21, 0.36), with financial support (β = 0.41; 95% CI: 0.13, 0.70), conducted by authors from European (other regions vs. European, β = -0.68; 95% CI: -1.05, -0.31) were associated with better methodological quality. The methodological quality of published DRMAs was suboptimal. Substantial efforts are warranted to improve the quality, including developing methodology guideline, involving more methodological trained authors, and so forth.

Informations de copyright

Copyright © 2018 Elsevier Inc. All rights reserved.