Article type
Abstract
"Objective: This study aimed to analyze the characteristics of the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) used in Cochrane systematic reviews (CSRs) on public health.
Methods: The Cochrane Library was searched in October 2021 to identify CSRs on public health. The CSRs were rated using the GRADE approach and assessed by the AMSTAR-2 tool. Finally, descriptive analyses of included CSRs were performed for the epidemiological characteristics of GRADE.
Result: 54 CSRs published between 2011 and 2021 were included in this study, most of which focused on nutrition (31%), health systems (15%), occupational health (14%), and environmental health (12%). Besides, most of the CSRs were randomized controlled trials (83%). The methodological quality of the CSRs was either high (89%) or moderate (11%). A total of 619 outcomes were respectively rated as being of high (4%), moderate (16%), low (38%), and very low (42%) evidence quality by GRADE. Among the five downgrade factors of GRADE, the most prevalent were risk of bias (68%), imprecision (34%), and indirectness (16%). However, only the magnitude of effect size was prevalent among the upgrade factors.
Conclusion: On public health, the usage of GRADE approach need to be strengthened in systematic reviews, and the risk of bias, imprecision, and indirectness are the main factors impacting the evidence quality."
Methods: The Cochrane Library was searched in October 2021 to identify CSRs on public health. The CSRs were rated using the GRADE approach and assessed by the AMSTAR-2 tool. Finally, descriptive analyses of included CSRs were performed for the epidemiological characteristics of GRADE.
Result: 54 CSRs published between 2011 and 2021 were included in this study, most of which focused on nutrition (31%), health systems (15%), occupational health (14%), and environmental health (12%). Besides, most of the CSRs were randomized controlled trials (83%). The methodological quality of the CSRs was either high (89%) or moderate (11%). A total of 619 outcomes were respectively rated as being of high (4%), moderate (16%), low (38%), and very low (42%) evidence quality by GRADE. Among the five downgrade factors of GRADE, the most prevalent were risk of bias (68%), imprecision (34%), and indirectness (16%). However, only the magnitude of effect size was prevalent among the upgrade factors.
Conclusion: On public health, the usage of GRADE approach need to be strengthened in systematic reviews, and the risk of bias, imprecision, and indirectness are the main factors impacting the evidence quality."