3.2.2.13 Developing a specialized register of RCTs

Note: The Cochrane Central Advisory Group (referred to below) was disbanded in October 2005.

A prerequisite for systematic reviews of the evidence relevant to the prevention or treatment of a particular health problem is to identify relevant randomized controlled trials (RCTs) as completely as possible, and to assemble them in a specialized register. One of the tasks of a Cochrane Review Group is to maintain and develop a specialized register containing all RCTs in their area of interest. This task is an essential part of the initial efforts to establish a Cochrane Review Group.

Some Cochrane Review Groups expect authors to conduct a handsearch of one journal that is likely to be important to their review as a contribution to the specialized register. Others undertake centrally a systematic handsearch of a core body of relevant journals. In either case, the searcher needs to identify all RCTs, not just those of particular interest to that Group. Someone in the Cochrane Review Group (usually the Managing Editor, a dedicated Trials Search Co-ordinator, or one of the editors) should agree to co-ordinate the search and register processes, and that person should work directly with the US Cochrane Center, which has responsibility for co-ordinating and supporting this activity. (Note: The USCC ceased to be responsible for co-ordinating CENTRAL activities in October 2005.)

Some Cochrane Review Groups may wish to include study designs other than RCTs in their specialized registers. In deciding to do so, they will need to develop inclusion criteria and pilot the application of these in practice. There has so far been relatively little experience in extending inclusion criteria beyond randomized controlled trials. Of the Cochrane Groups now considering other study designs, the Cochrane Effective Practice and Organisation of Care Group (EPOC) is probably most advanced, and advice from that source might help. Those who wish to base reviews on studies that have used methods other than randomization to control selection biases in comparing healthcare interventions should also consider contributing to the relevant Methods Groups exploring this.

A description of how the register was developed and is being maintained must be included in the Cochrane Review Group’s module. The following details should be available:

1. The inclusion criteria for the register, in particular the type of study included, e.g. randomized and controlled clinical trials only, or other comparative studies as well. There should be no language restriction.
2. The search strategy used to generate the register. This strategy might include:
   1. handsearching of relevant journals not being searched by other members of   The Cochrane Collaboration, with details of which volumes have been searched.
   2. handsearching of relevant conference proceedings.
   3. electronic searching (if possible using searches validated against handsearching) of electronic databases, such as MEDLINE and EMBASE. Details of the search terms used should be available. Many other electronic databases exist which could be searched, details of which are available from medical libraries.
   4. searching reference lists of studies identified.
   5. consulting existing trials’ registers.
3. Unpublished studies are often difficult to identify, but regular discussion with colleagues around the world can help identify them. Authors should also be encouraged to contact pharmaceutical companies where appropriate. Members of the Cochrane Review Group should encourage prospective registration of trials in their field.
4. How authors access the register to identify relevant studies (e.g. are authors sent references and, if so, how often?).
5. Data from the register were, until October 2005, forwarded to the US Cochrane Center for retagging as RCT in MEDLINE and contributing to The Cochrane Central Register of Controlled Trials (CENTRAL), a database of reports of RCTs that is published in The Cochrane Library.

The main purpose of CENTRAL is to establish the system for the flow of information of studies within The Cochrane Collaboration to ensure that each Cochrane Review Group is aware of all possible relevant studies that have been identified through the work of The Cochrane Collaboration. CENTRAL contains information that is simple and easily retrieved by Cochrane Review Groups to which it might be relevant. Each Cochrane Review Group must maintain its own specialized register, but CENTRAL provides an additional resource.

Various bibliographic software packages are available for storing references to trials, none of which are ideal for Cochrane specialized registers.