Adherence of systematic reviews to Cochrane RoB2 guidance was frequently poor: a meta epidemiological study

• Published in: Journal of clinical epidemiology 2022-12, Vol.152, p.47-55
• Main Authors: Minozzi, Silvia, Gonzalez-Lorenzo, Marien, Cinquini, Michela, Berardinelli, Daniela, Cagnazzo, Celeste, Ciardullo, Stefano, De Nardi, Paola, Gammone, Mariarosaria, Iovino, Paolo, Lando, Alex, Rissone, Marco, Simeone, Giovanni, Stracuzzi, Marta, Venezia, Giovanna, Moja, Lorenzo, Costantino, Giorgio, Cianciulli, Angelo, Cinnirella, Andrea, Grosso, Francesca, Luceri, Francesco, Venuti, Giuseppe, Vultaggio, Stefania, Zambarbieri, Emiliano
• Format: Article
• Language: English
• Subjects: Adequacy; Adherence; Bias; Epidemiologic Studies; Epidemiology; Humans; Impact analysis; Impact factors; Medicine; Meta-epidemiologic methods; Methodological quality; Quality assessment; Quartiles; Randomized controlled trials; Regression analysis; Research Design; Research Report; Reviews; Risk of bias; Software; Systematic review; Systematic reviews; Systematic Reviews as Topic
• ISSN: 0895-4356; 1878-5921
• DOI: 10.1016/j.jclinepi.2022.09.003

To assess whether the use of the revised Cochrane risk of bias tool for randomized trials (RoB2) in systematic reviews (SRs) adheres to RoB2 guidance. We searched MEDLINE, Embase, Cochrane Library from 2019 to May 2021 to identify SRs using RoB2. We analyzed methods and results sections to see whether risk of bias was assessed at outcome measure level and applied to primary outcomes of the SR as per RoB2 guidance. The relation between SR characteristics and adequacy of RoB2 use was examined by logistic regression analysis. Two hundred-eight SRs were included. We could assess adherence in 137 SRs as 12 declared using RoB2 but actually used RoB1 and 59 did not report the number of primary outcomes. The tool usage was adherent in 69.3% SRs. Considering SRs with multiple primary outcomes, adherence dropped to 28.8%. We found a positive association between RoB2 guidance adherence and the methodological quality of the reviews assessed by AMSTAR2 (p-for-trend 0.007). Multivariable regression analysis suggested journal impact factor [first quartile vs. other quartiles] was associated with RoB2 adherence (OR 0.34; 95% CI: 0.16-0.72). Many SRs did not adhere to RoB2 guidance as they applied the tool at the study level rather than at the outcome measure level. Lack of adherence was more likely among low and very low quality reviews.