Methodological use and report of AMSTAR 2: a cross-sectional study

This study aimed to analyze the reporting and methodology approach used for the methodological assessment of systematic reviews (SRs) using AMSTAR 2.

We conducted a cross-sectional meta-research study. We included full reports of scholarly articles that applied AMSTAR 2 to assess SRs or other evidence syntheses. Eligible studies were published between January 1, 2018, and December 1, 2023. We analyzed data items regarding reporting and methodology approach used for the methodological assessment of SRs using AMSTAR 2. We also evaluated temporal trends (2018-2023) in seven AMSTAR 2-related reporting practices using Spearman and Kendall correlation tests, complemented by linear regression.

We analyzed 444 studies applying AMSTAR 2. In 78%, multiple authors conducted the assessments, with 91% performed independently. The recommended rating system was used in 84% of studies, and critical domains in 82%. A final confidence rating was reported in 76% of studies, but only 30% included a rationale. Domain-level judgments were presented in 72%, with justifications in only 3%. Most of the analyzed reporting practices showed significant increasing trends over time, with linear slopes ranging from 1 to 20 additional studies per year adopting recommended AMSTAR 2 procedures.

Authors should clearly define key domains and explain how overall confidence ratings are determined. Both overall ratings and domain-level judgments, with justifications, should be reported. Such practices enhance transparency and support efficient replication and updating of evidence syntheses. Although adherence to AMSTAR 2 guidance has increased over time, major gaps in justification and transparency remain.