Article type
Year
Abstract
Background:
As systematic reviews (SRs) continue to increase, more and more SRs are updated to include new evidence. However, empirical research suggests that there are a large number of unnecessary, misleading, and conflicting SRs, in part because of inappropriate methodological design, conduct, or reporting. And it is still unclear whether the quality of the SR is improved during the update process.
Objectives:
To compare the methodological and reporting quality of updated systematic reviews (SRs) and original SRs.
Methods:
We included 30 pairs of non-Cochrane updated and original SRs, identified from a search of PubMed and Embase.com databases. We used AMSTAR-2 to assess methodological quality and PRISMA for reporting quality. Stratified analyses were conducted to compare the differences between updated SRs and original SRs and explore the factors that might affect the degree of quality change.
Results:
Of the 60 non-Cochrane SRs, only 2 (3.3%) were of low quality, the remaining 58 (96.7%) were of critical low quality. There were no statistically significant differences in methodological quality between the updated SRs and the original SRs, although the compliance rates of 8 items of updated SRs were higher than that of original SRs. Updated SRs showed an improvement on 15 PRISMA items, but no items with statistically significant differences. The differences in fully reported AMSTAR-2 and PRISMA items between original SRs and updated SRs were also not statistically significant after adjusting for multiple review characteristics.
Conclusions:
The methodological and reporting quality of updated SRs were not improved compared with the original SRs, although the quality could be further improved for both updated SRs and original SRs. Further research should focus on developing a methodological or a reporting quality tool that is specifically applicable to updated SRs.
Patient or healthcare consumer involvement: No.
As systematic reviews (SRs) continue to increase, more and more SRs are updated to include new evidence. However, empirical research suggests that there are a large number of unnecessary, misleading, and conflicting SRs, in part because of inappropriate methodological design, conduct, or reporting. And it is still unclear whether the quality of the SR is improved during the update process.
Objectives:
To compare the methodological and reporting quality of updated systematic reviews (SRs) and original SRs.
Methods:
We included 30 pairs of non-Cochrane updated and original SRs, identified from a search of PubMed and Embase.com databases. We used AMSTAR-2 to assess methodological quality and PRISMA for reporting quality. Stratified analyses were conducted to compare the differences between updated SRs and original SRs and explore the factors that might affect the degree of quality change.
Results:
Of the 60 non-Cochrane SRs, only 2 (3.3%) were of low quality, the remaining 58 (96.7%) were of critical low quality. There were no statistically significant differences in methodological quality between the updated SRs and the original SRs, although the compliance rates of 8 items of updated SRs were higher than that of original SRs. Updated SRs showed an improvement on 15 PRISMA items, but no items with statistically significant differences. The differences in fully reported AMSTAR-2 and PRISMA items between original SRs and updated SRs were also not statistically significant after adjusting for multiple review characteristics.
Conclusions:
The methodological and reporting quality of updated SRs were not improved compared with the original SRs, although the quality could be further improved for both updated SRs and original SRs. Further research should focus on developing a methodological or a reporting quality tool that is specifically applicable to updated SRs.
Patient or healthcare consumer involvement: No.