How do Cochrane Reviews use information from trial registers? Findings from a cross-sectional audit

Presentation video:




Oral session: Education and training (2)


Wednesday 23 October 2019 - 14:00 to 15:30


All authors in correct order:

Opiyo N1, Tan-Koay AG2, Berber S2, Dwan K1, Glanville J3, Lasserson T1, Willson M2
1 Editorial and Methods Department, Cochrane Central Executive, UK
2 Cochrane Breast Cancer Group, NHMRC Clinical Trials Centre, The University of Sydney, Australia
3 York Health Economics Consortium, University of York, UK
Presenting author and contact person

Presenting author:

Melina Willson

Contact person:

Abstract text
Background: trial registers fulfil an important function by providing a transparent scientific record of clinical trials. Their role has been widely acknowledged as being a mechanism for reducing publication bias and detecting reporting issues around study outcomes by study investigators. We conducted this study to assess the extent to which Cochrane authors are incorporating this important source of trial information into their reviews and to determine whether guidance is required for Cochrane authors.

Objectives: to assess current Cochrane Review practice in identifying and incorporating information from clinical trial registers.

Methods: we conducted a cross-sectional study to assess a sample of new or updated intervention reviews from all Cochrane Review Groups up to February 2017. Two authors independently extracted data from each review using a pre-tested audit questionnaire. We analyzed data relating to the frequency of reporting: 1) the source and search strategy; 2) results of trial register searches; and 3) the use of trial register information in the review.

Results: over 90% (236/260) of Cochrane Reviews reported searching a trial register (e.g. or registers via the World Health Organization’s International Clinical Trial Registry Platform (WHO ICTRP). In reviews that reported trial register searches, 39% (92/236) indicated the number of trial records retrieved and 56.7% (134/236) used information from the trial register record in the review. Trial record information was incorporated into the results (39.6%; 53/134), 'Risk of bias' assessments (53.7%; 72/134), discussion (24.6%, 33/134) and conclusion sections (25.4%, 34/134). In the discussion section, trial register information was used mostly to describe the overall completeness and applicability of the evidence (48.5%, 16/33), and potential biases in the review process (30.3%, 10/33). Table 1 provides examples of how trial register information was incorporated in the review.

Conclusions: the majority of audited reviews searched trial registers. Information identified from these sources was most commonly integrated into results, but there are opportunities to consider their relevance in formulating implications for research. Based on this audit, we suggest Cochrane authors may benefit from guidance on how to incorporate information from registers into Cochrane Reviews. This could be useful to inform decisions on planning additional primary research and updating systematic reviews.

Patient or healthcare consumer involvement: although patients or healthcare consumers were not involved in this audit, the findings of this work indicate that some improvements are needed in the uptake of trial register information in Cochrane Reviews. By using trial register information more effectively, systematic reviews will be able to provide a more complete picture of the current and emerging evidence to its consumers.