Article type
Year
Abstract
Background: Evidence from systematic reviews (SR) addressing different comparisons for the same treatment could provide effect estimates of different quality. Overviews of SR incorporating GRADE to rate the quality of evidence would provide readers with summaries of evidence and classification of its quality. The accuracy and reproducibility of the quality evaluation process entails some challenges for overview authors.
Objectives: To describe the reliability aspects in the evidence synthesis process applying the GRADE framework, in the development of an overview of SR in the field of respiratory care in neonates.
Methods: The selection process was completed independently by two authors. A third author assessed the methodological quality of eligible SR using the AMSTAR. Two authors independently summarized the evidence for each comparisons/ outcomes in GRADE profiles. Consensus about the importance of outcomes for patients was reached ad hoc. Results from trials published after the SRs' search dates were added to the profiles. A third author compared the quality ratings provided by authors for each outcome and checked the accuracy of the overall process. Measurement of inter-rater agreement was done using the Cohen’s kappa coefficient. Reasons given by authors to rating down the quality of evidence for each outcome were also compared and observed agreement presented.
Results: The presentation will describe the above mentioned methods implemented in an overview assessing the effect of different ventilation strategies in preventing bronchopulmonary dysplasia and other neonatal outcomes in preterm infants with respiratory distress syndrome. A total of six SR addressing eight different comparisons between ventilation modes and strategies were included, and seven outcomes assessed. Only one additional trial was found and incorporated in the meta-analyses for each outcome.
Conclusions: The example illustrates some challenges in rating the quality of evidence from the GRADE framework in the context of developing overviews of SR, and the need of integrating reliable consensus methods among reviewers.
Objectives: To describe the reliability aspects in the evidence synthesis process applying the GRADE framework, in the development of an overview of SR in the field of respiratory care in neonates.
Methods: The selection process was completed independently by two authors. A third author assessed the methodological quality of eligible SR using the AMSTAR. Two authors independently summarized the evidence for each comparisons/ outcomes in GRADE profiles. Consensus about the importance of outcomes for patients was reached ad hoc. Results from trials published after the SRs' search dates were added to the profiles. A third author compared the quality ratings provided by authors for each outcome and checked the accuracy of the overall process. Measurement of inter-rater agreement was done using the Cohen’s kappa coefficient. Reasons given by authors to rating down the quality of evidence for each outcome were also compared and observed agreement presented.
Results: The presentation will describe the above mentioned methods implemented in an overview assessing the effect of different ventilation strategies in preventing bronchopulmonary dysplasia and other neonatal outcomes in preterm infants with respiratory distress syndrome. A total of six SR addressing eight different comparisons between ventilation modes and strategies were included, and seven outcomes assessed. Only one additional trial was found and incorporated in the meta-analyses for each outcome.
Conclusions: The example illustrates some challenges in rating the quality of evidence from the GRADE framework in the context of developing overviews of SR, and the need of integrating reliable consensus methods among reviewers.