Article type
Year
Abstract
Background: The PRISMA, AMSTAR and GRADE are used to assess the quality of report, methodology and evidence in systematic reviews (SRs), but up to now there was no literature had ever studied the correlation of the quality of report, methodology and evidence in SRs.
Objectives: To assess the correlation of PRISMA, AMSTAR and GRADE in SRs.
Methods: We selected the SRs about complementary and alternative medicine to respiratory diseases as the sample. Chinese Biomedical Literature database (CBM) and Cochrane Database of Systematic Reviews (CDSR) were searched up to Sept 2012. Each SR was independently identified and evaluated by two reviewers, and discussed with the third member when disagreement appeared.
Results: (1)79 SRs were included. (2) The most inadequately reported item in PRISMA was ‘structured summary’ (96%), followed by ‘objective’ (95%) and ‘search’ (86%). (3) 91% and 85% SRs had never provided the information of the conflict of interests and the list of studies (included and excluded) in AMSTAR. (4) A majority of the SRs had a ‘low’ or ‘very low’ quality of evidence in GRADE. (5) the correlation of PRISMA, AMSTA and GRADE was showed in Table 1.
Conclusions: The PRISMA and AMSTAR obviously had a positive correlation (p = 0), but there was no correlation between PRISMA and GRADE, as well as AMSTAR and GRADE.
Objectives: To assess the correlation of PRISMA, AMSTAR and GRADE in SRs.
Methods: We selected the SRs about complementary and alternative medicine to respiratory diseases as the sample. Chinese Biomedical Literature database (CBM) and Cochrane Database of Systematic Reviews (CDSR) were searched up to Sept 2012. Each SR was independently identified and evaluated by two reviewers, and discussed with the third member when disagreement appeared.
Results: (1)79 SRs were included. (2) The most inadequately reported item in PRISMA was ‘structured summary’ (96%), followed by ‘objective’ (95%) and ‘search’ (86%). (3) 91% and 85% SRs had never provided the information of the conflict of interests and the list of studies (included and excluded) in AMSTAR. (4) A majority of the SRs had a ‘low’ or ‘very low’ quality of evidence in GRADE. (5) the correlation of PRISMA, AMSTA and GRADE was showed in Table 1.
Conclusions: The PRISMA and AMSTAR obviously had a positive correlation (p = 0), but there was no correlation between PRISMA and GRADE, as well as AMSTAR and GRADE.
Images
