Impact of including conference abstracts in systematic reviews: analysis from a living systematic review and network meta-analysis of COVID-19

Date & Time
Monday, September 4, 2023, 12:05 PM - 12:25 PM
Location Name
Churchill
Session Type
Oral presentation
Category
Bias
Oral session
Research integrity, transparency and fraud
Authors
Oliveros M1, Ibrahim S2, Zeraatkar D2, Zaror C3, Bravo G2, Brignardello-Petersen R2
1HEI McMaster University, Canada
2Department of Health Research Methods, Evidence, and Impact, McMaster University, Canada
3Department of Pediatric Dentistry and Orthodontics, Faculty of Dentistry, Chile
Description

Background: Summarizing all available evidence while decreasing bias raises important methodological challenges for systematic reviews (SRs). Grey literature in SRs is important for minimizing the risk of publication bias but is often inadequately reported, which may reduce the credibility of the evidence synthesis.
Objectives: To evaluate the extent of consistency between conference abstracts (CAs) and their final publications and the impact of including results from CAs on meta-analytic effect estimates and the certainty of the evidence (CoE) in our living systematic review and network meta-analysis (SRNMA) for COVID-19.
Methods: We will conduct a retrospective review of all CAs included in our living SRNMA of drug treatments for COVID-19 until January 2023. For all CAs included in our living SRNMA with a subsequent final publication, two reviewers, working independently, will compare the results and risk of bias assessments of the CA and its final publication. To assess the impact of the inclusion of CAs on effect estimates and the CoE, we will include the outcomes most frequently reported in the CAs reporting data for interventions with existing recommendations according to the January 2023 World Health Organization guidelines. We will conduct two analyses, one including the CAs and one excluding them. We will compare the results of these analyses to evaluate the magnitude of the change in the pooled estimate and identify any changes to the CoE.
Results: This work is in progress. We will describe the nature and frequency of discrepancies between CAs and their final publications. We will present how effect estimates and the CoE changed with and without the CAs. We will consider the direction of effect to be different, whether the effect changes from beneficial to harmful or vice versa, considering the thresholds of effectiveness established in the SRNMA. Conclusion: The results of this analysis will highlight the impact of including CAs in a living SRNMA, which can inform eligibility decisions for future evidence syntheses. Understanding the influence of including CAs in network meta-analyses can assist decision-makers with interpreting the results of evidence syntheses with CAs.
Patient, public and/or healthcare consumer involvement: Patient partners were not involved in this study.