Mapping risk of bias criteria in systematic reviews of in vitro endodontic studies: an umbrella review
摘要
The risk of bias (RoB) assessment is a critical aspect of systematic reviews, yet a lack of standardization in RoB criteria persists in the endodontic literature. This study aimed to evaluate the RoB tools used in systematic reviews (SRs) of in vitro endodontic studies to synthesize the most commonly employed criteria.
MethodsA comprehensive search of five databases (PubMed, Embase, Web of Science, Scopus, and Cochrane Library) was conducted, identifying SRs with RoB evaluation. Data were extracted independently by two reviewers, focusing on the criteria used for RoB assessment. All SRs were categorized by main objective and the evaluation test. The study protocol was registered in the Open Science Framework (https://osf.io/azd8u/).
ResultsOf the 10,409 records initially identified, 142 systematic reviews (SRs) published between January 2018 and September 2024 met the inclusion criteria. These SRs employed 36 distinct RoB criteria, with the most common being blinding (40.9%), sample size (34.5%), adequate statistical analysis (31%), randomization (30.3%), and standardized sample preparation and handling (21.8%).
DiscussionInconsistencies in the application of RoB criteria were evident, with many SRs employing tools adapted from clinical studies that were not fully suitable for the unique methodological requirements of in vitro research. Although some SRs attempted to adapt criteria from other studies, these adaptations often fell short of addressing the specific needs of in vitro studies. The absence of standardized tools compromises the reliability of current SRs and limits their comparability with future research.
ConclusionsThe observed inconsistencies in RoB assessments across SRs of in vitro endodontic studies highlight the critical need for a standardized tool specifically designed for this context. Such a tool should account for the methodological nuances of in vitro research, including the relevance of criteria like blinding and sample size calculation, to enhance the reliability, reproducibility, and overall relevance of future SRs.