Article type
Year
Abstract
Background: We conducted a systematic review of randomized controlled trials (RCTs) on the comparative effectiveness and safety of drugs for acute migraine headache in the emergency department. Safety is poorly reported and often the same adverse effects (AEs) are not reported for similar interventions, so traditional pair-wise meta-analysis of data is not possible. We proposed an alternative comprehensive safety report for these treatments.
Objective: To describe a novel approach to presenting AE results in the absence of traditional pair-wise meta-analyses.
Methods: A priori AEs were grouped into categories (e.g., vomiting, sedation). After mapping the interventions and AE categories, a traditional pair-wise meta-analysis was not possible. Therefore, we analyzed AEs for individual arms of trials. When an intervention had more than one study reporting on any AE, the results were pooled using a standard inverse variance random effects meta-analysis. The risks (i.e., incidence) for each AE category are presented as a summary estimate and 95% confidence interval.
Results: From a comprehensive search, 39 RCTs investigating 9 drug classes reported specific AEs. Summary tables present the risk of each AE category for specific active agents and placebo. Graphs for each AE category were created using Excel. The horizontal axis shows the active agents or placebo that resulted in an AE; the vertical axis shows the risk of an AE associated with each agent. Specific examples of tables and graphs for AE categories will be presented.
Conclusion: A traditional pair-wise meta-analysis of AEs was not possible since multiple RCTs testing the same drugs failed to report common AEs. Instead, we present a summary of AEs by treatment arm that provides an overall picture of which interventions had a high risk of specific AEs.
Objective: To describe a novel approach to presenting AE results in the absence of traditional pair-wise meta-analyses.
Methods: A priori AEs were grouped into categories (e.g., vomiting, sedation). After mapping the interventions and AE categories, a traditional pair-wise meta-analysis was not possible. Therefore, we analyzed AEs for individual arms of trials. When an intervention had more than one study reporting on any AE, the results were pooled using a standard inverse variance random effects meta-analysis. The risks (i.e., incidence) for each AE category are presented as a summary estimate and 95% confidence interval.
Results: From a comprehensive search, 39 RCTs investigating 9 drug classes reported specific AEs. Summary tables present the risk of each AE category for specific active agents and placebo. Graphs for each AE category were created using Excel. The horizontal axis shows the active agents or placebo that resulted in an AE; the vertical axis shows the risk of an AE associated with each agent. Specific examples of tables and graphs for AE categories will be presented.
Conclusion: A traditional pair-wise meta-analysis of AEs was not possible since multiple RCTs testing the same drugs failed to report common AEs. Instead, we present a summary of AEs by treatment arm that provides an overall picture of which interventions had a high risk of specific AEs.