Completeness of reporting of rapid reviews of primary studies according to journal publication status using PRISMA and PRISMA for Abstracts




Poster session 2


Monday 17 September 2018 - 12:30 to 14:00

All authors in correct order:

Stevens A1, Garritty C1, Hersi M2, Hartling L3, Stewart L4, Thavorn K5, Tricco A6, Welch V7, Moher D5
1 Cochrane Rapid Reviews Methods Group and Ottawa Hospital Research Institute, Canada
2 Ottawa Hospital Research Institute, Canada
3 University of Alberta, Canada
4 University of York, Canada
5 Ottawa Hospital Research Institute and University of Ottawa, Canada
6 St. Michael's Hospital and University of Toronto, Canada
7 Campbell Collaboration, Canada
Presenting author and contact person

Presenting author:

Adrienne Stevens

Contact person:

Abstract text
Background: Rapid reviews have emerged to address pressing decision-making situations and follow a similar process to systematic reviews. However, the process of conduct of rapid reviews can differ in important ways, such as use of methodological shortcuts, inclusion of secondary evidence, and development in close contact with the end user. As with any other type of research report, complete transparency about this process and the results are essential. Little empirical evidence exists on the completeness of reporting of rapid reviews.

Objectives: To evaluate the completeness of reporting of rapid reviews, according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) and PRISMA for Abstracts guidelines. Reporting will be compared according to whether rapid reviews are journal-published or not.

Methods: Cross-sectional, comparative study guided by a protocol. Bibliographic databases, gray literature sources, contact with 148 rapid review-producing organizations, and consultation within our team were used to locate rapid reviews. 'Rapid review’ was defined a priori. Duplicate review was used for study selection, and journal-published rapid reviews were screened first. Non-journal-published rapid reviews were organized by clusters and sampled proportionately, for feasibility. Extraction was verified by a second person. Primary analysis was a mean summed score (mean difference and 95% confidence intervals [CI]) and secondarily by-item (risk ratio and 99% CI) for the publication years 2014 and 2016. We also collected additional reporting items beyond that of PRISMA and PRISMA for Abstracts.
Results: We excluded rapid reviews including secondary evidence (approximately 40%) as they would have challenged the face validity of one-third of PRISMA items. A total of 91 rapid reviews of primary studies were included (47 for 2014 and 44 for 2016). Analysis by mean summed score and by-item will be provided at the time of the presentation.

Conclusions: This investigation will contribute to the understanding of reporting characteristics of an international sample of rapid reviews of primary studies.

Patient or healthcare consumer involvement: This study is the first phase of developing the PRISMA reporting guideline for rapid reviews of primary studies, and consumer involvement is planned in the next phase.

Relevance to patients and consumers: 

Use of research for decision-making in patient care and for the healthcare system relies on completely and transparently reported research. Rapid reviews are no exception, and any shortcuts used to develop them (relative to systematic reviews) and their findings need to be clearly stated. This will ensure the essential information is provided when being considered by patients and other decision-makers.