Using the Cochrane Register of Studies to identify the evidence gaps in Cochrane systematic reviews

Article type
Authors
Showell M1, Jani D1, Farquhar C1, Jordan V1
1University of Auckland
Abstract
Background: Cochrane aims to produce the 'gold standard of evidence'. The Cochrane Gynaecology and Fertility Group (CGF) produces systematic reviews (SRs), within this subject area, using trials from a database of over 17,500 randomised control trials (RCTs). We believe many of these RCTs are not currently adding to the evidence base, as there are gaps in the topics covered by the existing SRs.

Objectives: The main aim of this project is to identify gaps in evidence by identifying the RCTs, published in the area of gynaecology, that have not been used in Cochrane SRs.

Methods: We conducted an audit of all gynaecology trials in the CGF specialised register, published from 2010 to 2013. The search function in the Cochrane Register of Studies (CRS web) could identify quickly which trials had been ‘included’, ‘excluded’ or simply ‘not used’ in SRs across the Collaboration. We then classified the ‘not used trials’ into their population/condition and intervention. From this point we looked at the existing SRs in CGF to assess whether they in fact could be incorporated. The remaining ‘not used trials’ were grouped into potential review topics.

Results: Our database search found 1514 trials published between 2010 and 2013. From these, we excluded 112 as they did not meet inclusion criteria; 289 were already used in Cochrane SRs and 1198 were unused. At this preliminary stage, we have classified 567 unused publications published from 2010 to 2011. Of these, 63 trials could fit into existing reviews, and 68 trials had no existing SR topic. Menopause was the most commonly unused population (42%), and, within this group, the most common unused interventions were medical therapies (40.5%).

Conclusions: We found more trials than expected that are not currently summarised within existing Cochrane SRs. The largest number of unused trials we found were for the topic of menopause and medical interventions. By identifying the populations and interventions that are not currently reviewed systematically by Cochrane, we can now develop priority topics and thus provide better healthcare evidence.

Patient or healthcare consumer involvement: Consumer involvement is vital to the prioritisation of important topics from the unused trials, and, in phase two of this project consumers will be consulted on the priority on the topics identified above.