Skip to main content

Languages

No RCT is an island: how often do RCTs cite relevant existing Cochrane reviews?

Date and Location

Session: 

P3.020
Presenting author and contact person

Presenting author

Philippa Middleton

Contact person

Philippa Middleton
Abstract text
Background: Randomised controlled trials (RCTs) need to place themselves in the context of other studies, ideally through making reference to any existing systematic review in completed RCT publications. Objectives: To assess the extent of RCTs citing relevant Cochrane reviews in existence before publication of the RCT and the nature and purpose of these cites. Methods: All Cochrane reviews including > 50 RCTs, updated at least once, as at January 2013 were assessed for trials included since the previous version(s) of the review. These trials were then checked to see if and how they had cited the previous review version. Results: Of the 100 Cochrane reviews estimated to fulfil the above criteria, results are presently available for 40 reviews. Just under half (19/40) included at least one RCT which cited previous versions of the Cochrane review – median 2 (range 1 to 28) per review. On a per RCT basis, nearly 77% of eligible trials (258/334) failed to cite a previous Cochrane review where an update of the review now includes that trial. When RCTs did cite a Cochrane review, this was generally a statement of review results. Six trial reports described research gaps and three mentioned that they had used the Cochrane review to help design their trial. Full results will be available for the Colloquium. Conclusions: Most new RCTs included in updates of Cochrane reviews should be able to cite the previous version of the review, yet less than a quarter do so. This may be due to lack of knowledge that a Cochrane review exists or a preference for citing individual RCTs addressing similar questions. Either way, failure to use a relevant Cochrane review at design stage may decrease the risk of the RCT addressing the most pertinent questions.