Use of GRADE in Dentistry Systematic Reviews: A Methodological Study

Article type
Authors
Ibrahim S1, Azab M2, Li A3, Chu J3, Wang C4, Modi K4, Abumustafa Y5, Hong K6, Park S7, Yan Y4, Miroshnychenko A1, Brignardello-Petersen R1
1Department of Health Research Methods, Evidence, and Impact, McMaster University
2Temerty Faculty of Medicine, University of Toronto
3Faculty of Science, McMaster University
4Faculty of Health Sciences, McMaster University
5Faculty of Medicine, University of Ottawa
6 Department of Kinesiology, Faculty of Science, McMaster University
7School of Pharmacy, University of Waterloo
Abstract
Background: Systematic reviews (SRs) synthesize evidence from all available studies. Trustworthy SRs should assess the certainty of the evidence through approaches, such as the Grading of Recommendations, Assessment, Development and Evaluation (GRADE). However, uptake of GRADE in dentistry is limited.

Objectives: To evaluate the frequency of GRADE use in dentistry SRs and the relationship between GRADE use and SR methodological quality. Among SRs that use GRADE, to evaluate whether GRADE is used appropriately and the frequency of the levels of certainty according to GRADE.

Methods: We searched Ovid MEDLINE for dentistry SRs of randomized controlled trials published between January 1, 2016, and September 23, 2021. Pairs of reviewers independently conducted screening and data extraction. We determined the frequency of GRADE use, determined the frequency of the levels of certainty for the main outcome, and assessed whether GRADE was used appropriately using pre-specified criteria. We assessed whether SRs using GRADE are more likely to have better methodological quality according to two aspects of the Risk of Bias Assessment Tool for Systematic Reviews (ROBIS).

Results: We included 200 SRs as per our target sample size. Thirty-six percent of these used GRADE and of those, 51% and 30% used GRADE appropriately at the review and outcome level, respectively. Among SRs using GRADE, most reported low (34%) or moderate (31%) certainty evidence with few high certainty ratings (7%). SRs using GRADE may be more likely to search for grey literature (OR = 3.03; 95% CI [1.66-5.53]) and minimize errors in both screening and data extraction (OR = 2.81; 95% CI [1.36-5.81]) compared with those not using GRADE.

Conclusions: Most SRs in our sample did not use GRADE, and many of those using GRADE did not use it appropriately. Using GRADE may be associated with higher methodological quality. SR authors and consumers should be aware of these limitations when interpreting SR conclusions. Improving the quality and trustworthiness of future SRs impacts patients as SRs are considered to be at the top of the hierarchy of evidence-based medical research.

Patient, public, and/or healthcare consumer involvement: Patients were not involved.