Audit of the use of clinical trial registries in Cochrane Intervention Reviews – preliminary results

ID: 

18932

Session: 

Short oral session 1: Improving conduct and reporting of evidence synthesis

Date: 

Wednesday 13 September 2017 - 11:00 to 12:30

Location: 

Meeting Room 1.43

All authors in correct order:

Berber S1, Tan-Koay A1, Opiyo N2, Glanville J3, Lasserson T2, Willson M1
1 Cochrane Breast Cancer Group, Australia
2 Cochrane Central Executive, UK
3 York Health Economics Consortium, UK
Presenting author and contact person

Presenting author:

Slavica Berber

Contact person:

Slavica Berber
Melina Willson
Abstract text
Background: Clinical trial registries were established with a general aim of reducing publication bias. Searching trial registries as part of the search process for Cochrane reviews has been mandatory since 2013. There is growing interest in how clinical trial registration status and trial registry data impacts on systematic review findings. The extent of trial registry information use in Cochrane reviews is not well established.

Objectives: To audit how clinical trial registries of various types and associated subject-specific registers are currently used in Cochrane intervention reviews.

Methods: We selected the 5 most recently published intervention reviews or review updates from 52 Cochrane Review Groups (search date: 1 February 2017). A single assessor extracted data from each review on the frequency of: i) trial registry searching; ii) reporting results of trial registry searches; and, iii) use of information obtained from trial registries in the review (e.g. for risk-of-bias or GRADE assessment, and planning review updates).

Results: Preliminary results from 60 reviews (12 CRGs) showed that 90% (54/60) included a search of at least one major registry or portal (ClinicalTrials.gov, WHO ICTRP). The number of records retrieved from trial registries was reported in 26% of reviews. Information obtained from clinical trial registries was used in 65% (39/60) of reviews - most frequently for risk-of-bias assessment with 20% (12/60) of reviews using trial registry data when judging reporting bias. In 18% (11/60) of reviews, the Discussion section included mentions of trial registry information in Potential Biases in the Review Process and Implications for Practice or Research. Twelve per cent (7/60) of reviews used the phrase 'No ongoing studies were identified' if trial registry searches did not retrieve any eligible ongoing studies.

Conclusions: The majority of audited reviews used trial registry data in some way. Based on these preliminary results, trial registries are being routinely searched in Cochrane reviews, but guidance should be established to assist review teams to make use of information obtained from searching these sources.