Reporting the GRADE approach by systematic reviews of interventions in pediatric dentistry: methodological study

Article type
Authors
Martins-Pfeifer C1, Alvarenga-Brant R1, Notaro S1, Stefani C2, Pereira A1, Povoa-Santos L1, Souza-Oliveira A1, Campos J1, De Luca Canto G3
1Universidade Federal De Minas Gerais, Belo Horizonte, Minas Gerais, Brazil
2Universidade de Brasília, Brasília, DF, Brazil
3Universidade Federal de Santa Catarina, Florianópolis, Santa Catariana, Brazil
Abstract
"Background: Assessing the certainty of the evidence in systematic reviews (SRs) of intervention can provide a more consistent judgment of the evidence, communicating results to support better-informed choices in health care.
Objective: To analyze the use of the GRADE approach to assess the certainty of evidence by SRs of interventions in pediatric dentistry.
Methods: The inclusion criteria were SRs of randomized clinical trials (RCTs) and non-randomized studies of interventions (NRSIs) in pediatric dentistry that reported the certainty of the evidence through the GRADE approach. The search was limited to two years (from 2020 to 2022). Paired independent reviewers screened the studies, extracted data, and appraised the methodological quality using the AMSTAR 2 tool. The certainty of the evidence was extracted for each outcome and comparison. A descriptive analysis was carried out.
Results: Twenty-four of 87 SRs of intervention in pediatric dentistry reported the certainty of the evidence through the GRADE approach (28%). The methodological quality of the 24 SRs was high (16.7%), moderate (12.5%), low (37.5%) and critically low (33.3%). Twenty studies were included for GRADE analysis, reporting 125 evidence outcomes: 112 generated from RCTs and 13 from NRSIs. The certainty of the evidence for outcomes generated from RCTs and NRSIs was (respectively) very low (40.2% and 84.6%), low (33.1% and 7.7%), moderate (17.8% and 7.7%) and high (9.8% and 0.0%). The main reasons to downgrade the certainty were (for RCTs and NRSIs, respectively) risk of bias (68.8% and 84.6%), imprecision (67.8% and 100.0%), inconsistency (18.8% and 0.0%), indirectness (17.8% and 0.0%), and publication bias (7.1% and 0.0%). Only six outcomes reported a superior effect of the intervention compared to the comparison with high certainty: virtual reality glasses for pain and anxiety control, rotatory instruments for root canal, and articaine for pain control during anesthesia.
Conclusion: The certainty of the evidence using the GRADE approach of SRs of intervention in pediatric dentistry was mainly very low and low. The main reasons for downgrading the certainty of evidence were risk of bias and imprecision.
Protocol: PROSPERO #CRD42022365443.
Support: PRPq/ UFMG and FAPEMIG #APQ-00323-17."