Article type
Year
Abstract
Background: Systematic reviews (SRs) of harms can provide valuable information to help describe adverse events (AEs) but they are hampered by lack of standardized methods to report these events. A systematic review undertaken by our team has identified major gaps in reporting in systematic reviews of AEs, necessitating the need for a new guideline, The PRISMA Harms Extension.
Objectives: To modify PRISMA to include items specific to harms reporting.
Methods: A modified Delphi process was used. An initial checklist of potential items to be reported in a SR of AEs was developed and refined by experts. The first phase of the Delphi process was online, such that 40 potential items were sent to experts in SRs, guideline development, clinical trials, statistics and epidemiology, who assigned relevance of each item on a 1–10 Likert scale in two rounds. After the second round, items voted 8 or higher were kept. Items voted 5 or less were removed, and those that were indeterminate were carried forward for the next Delphi phase: the in-person consensus meeting. At this meeting, an invited group of relevant experts discussed and decided the final list of checklist items that should be kept in the guideline.
Results: The online Delphi originated significant agreement among participants. Out of the 40 items scored twice by 72 participants, one item was voted ‘excluded’ and 7 items received indeterminate votes. The consensus meeting had 25 worldwide experts in guideline development and systematic reviews. After two full day discussions on the relevance of items, it was decided that 5 items should be mandatory and 14 items should be considered recommended when reporting harms in systematic reviews.
Conclusions: The ultimate goal of this guideline development is to improve quality of reporting in systematic reviews, so that both benefits and harms are discussed.
Objectives: To modify PRISMA to include items specific to harms reporting.
Methods: A modified Delphi process was used. An initial checklist of potential items to be reported in a SR of AEs was developed and refined by experts. The first phase of the Delphi process was online, such that 40 potential items were sent to experts in SRs, guideline development, clinical trials, statistics and epidemiology, who assigned relevance of each item on a 1–10 Likert scale in two rounds. After the second round, items voted 8 or higher were kept. Items voted 5 or less were removed, and those that were indeterminate were carried forward for the next Delphi phase: the in-person consensus meeting. At this meeting, an invited group of relevant experts discussed and decided the final list of checklist items that should be kept in the guideline.
Results: The online Delphi originated significant agreement among participants. Out of the 40 items scored twice by 72 participants, one item was voted ‘excluded’ and 7 items received indeterminate votes. The consensus meeting had 25 worldwide experts in guideline development and systematic reviews. After two full day discussions on the relevance of items, it was decided that 5 items should be mandatory and 14 items should be considered recommended when reporting harms in systematic reviews.
Conclusions: The ultimate goal of this guideline development is to improve quality of reporting in systematic reviews, so that both benefits and harms are discussed.