Article type
Year
Abstract
Background:
An increasing number of dose-response meta-analyses (DRMA) have been published in the past several years, but the scientific quality of these meta-analyses has not been investigated.
Objectives:
In this study, we conducted a cross-sectional survey of DRMAs published during the past seven years. We aimed to overview the reporting as well as methodological quality of these DRMAs, and to investigate the potential factors associated with the quality.
Methods:
We searched MEDLINE (Ovid), Embase (Ovid), and Wiley Online Library for published DRMAs from 1 January 2011 to 31 July 2017. Two experienced authors screened the literature and extracted the basline characters independently. The modified AMSTAR (A MeaSurement Tool to Assess systematic Reviews) and PRISMA checklist were used to assess the methodological and reporting quality of these DRMAs. The total methodology score and reporting score were then fitted in a weighted linear regression model against six predefined variables (number of authors, year of publication, time to complete the DRAM, methodologist involved, region of first author, and funding status). We collected all data in Excel 2013 (Microsoft, Washington) and used STATA 14.0 (STATA, College Station, TX) for analysis, with a P value of < 0.05 denoting significance.
Results:
Initially, we obtained 7061 records from the primary literature search, and finally included 529 DRMAs in this article. About 66.16% of the DRMAs were conducted by Asian authors. Generally, the methodological quality was poor, with poor compliance for two-fifths (6/15) of the quality items, but the reporting quality was relatively high with poor compliance for only three items (3/26). In our regression analysis, we found positive associations with the methdological and (or) reporting quality when more authors were involved (P < 0.01) and any methodologist participated (P < 0.01). We observed further significant improvement of the methodological and reporting quality as time progressed (P < 0.01).
Conclusions:
Based on current evidence, while the methodological quality of published DRMAs has been poor, the reporting quality has been relatively high. Both the methodological quality and reporting quality of DRMAs have improved over the time. The more authors that are involved and any methodologist participation may benefit the scientific quality of DRMAs.
Patient or healthcare consumer involvement:
None.
An increasing number of dose-response meta-analyses (DRMA) have been published in the past several years, but the scientific quality of these meta-analyses has not been investigated.
Objectives:
In this study, we conducted a cross-sectional survey of DRMAs published during the past seven years. We aimed to overview the reporting as well as methodological quality of these DRMAs, and to investigate the potential factors associated with the quality.
Methods:
We searched MEDLINE (Ovid), Embase (Ovid), and Wiley Online Library for published DRMAs from 1 January 2011 to 31 July 2017. Two experienced authors screened the literature and extracted the basline characters independently. The modified AMSTAR (A MeaSurement Tool to Assess systematic Reviews) and PRISMA checklist were used to assess the methodological and reporting quality of these DRMAs. The total methodology score and reporting score were then fitted in a weighted linear regression model against six predefined variables (number of authors, year of publication, time to complete the DRAM, methodologist involved, region of first author, and funding status). We collected all data in Excel 2013 (Microsoft, Washington) and used STATA 14.0 (STATA, College Station, TX) for analysis, with a P value of < 0.05 denoting significance.
Results:
Initially, we obtained 7061 records from the primary literature search, and finally included 529 DRMAs in this article. About 66.16% of the DRMAs were conducted by Asian authors. Generally, the methodological quality was poor, with poor compliance for two-fifths (6/15) of the quality items, but the reporting quality was relatively high with poor compliance for only three items (3/26). In our regression analysis, we found positive associations with the methdological and (or) reporting quality when more authors were involved (P < 0.01) and any methodologist participated (P < 0.01). We observed further significant improvement of the methodological and reporting quality as time progressed (P < 0.01).
Conclusions:
Based on current evidence, while the methodological quality of published DRMAs has been poor, the reporting quality has been relatively high. Both the methodological quality and reporting quality of DRMAs have improved over the time. The more authors that are involved and any methodologist participation may benefit the scientific quality of DRMAs.
Patient or healthcare consumer involvement:
None.