Guidelines for conducting a systematic review with the Cochrane Breast Cancer Group

review Author PACKAGE

Would you like to know more about conducting your own systematic review?
If you are interested in developing a systematic review you should firstly familiarise yourself with this package. The aim of the package is to provide information about developing a review, describe each step of the process, and to ensure that authors are aware of, and have access to, the range of resources available from The Cochrane Collaboration and the Cochrane Breast Cancer Group (CBCG).

What are systematic reviews?
Systematic reviews are a powerful tool for answering important clinical questions when a single definitive large-scale trial has not been conducted. They combine the results of all relevant high quality studies that address the same clinical question. A properly conducted systematic review defines the elements of the research question 'a priori' (before the review commences). These elements include the patient population, the intervention being evaluated, the intervention it is to be compared to, the outcomes and the methods to be used. These are documented and published in the review protocol.

A comprehensive search strategy is developed to identify studies which correspond to the items listed above for each question. All eligible studies that are identified are critically appraised and summarised. Statistical methods for combining the study results (meta-analysis) are used when appropriate.

THE ROLE OF THE BREAST CANCER GROUP

Preparing a Cochrane review is somewhat different to submitting work to a peer reviewed journal. The CBCG's role is to support review authors while the review is being developed particularly in relation to advice, training and editorial expertise to expediate and secure publication in The Cochrane Library (CLib). Like a peer reviewed journal however, each review should meet the quality and editorial standards determined by the CBCG, who ultimately reserve the right to make the final decision regarding publication.

The CBCG has developed a number of policies relating to the conduct of systematic reviews in breast cancer. These are outlined on this website and in the group's module which is published quarterly on the Cochrane Library. Authors are strongly advised to familiarise themselves with the specific policies of the CBCG and to prepare their protocols and reviews accordingly. It should be noted that policies are regularly revised and updated and authors are therefore encouraged to review these more than once during the process of review development. If you are ever unsure about anything relating to this process you should contact the CBCG coordinator at Cochrane@ctc.usyd.edu.au.

GENERAL POLICIES RELATING TO THE CONDUCT OF A SYSTEMATIC REVIEW WITH THE CBCG

1. All reviews require at least two review authors but it is that there is a review team which comprises a clinical and methodological/statistical expert. The team should also comprise a consumer advocate and advisor. Review groups are encouraged to include a consumer as a review author.

2. Individuals/Groups who are the first to nominate and register a specific question will usually take priority for that question.

3. Review groups are required to meet specific timeframes in relation to review production.

4. The Cochrane Breast Cancer Group reserves the right to withdraw protocols and reviews where authors fail to meet their obligations or to transfer these to alternative review groups.

5. Review groups are required to nominate one author who will take responsibility for the integrity of the data contained in the review. This is documented in the review.

6. Methods of conflict resolution will be applied where there is dispute or overlap in a review question.

These policies are outlined in full in the group's module which is published quarterly on The Cochrane Library.

Review Manager
Review Manager (RevMan) is the Cochrane Collaboration software for the production of systematic reviews. Authors are required to prepare and maintain their reviews using this program. Authors will be required to update their software as updates are released.

Assessment of methodological quality
1. Some measure of quality is required for all breast cancer reviews.
2. Several methods are available to assess or measure the risk of bias, validity or quality of a randomised controlled trial. The CBCG recommends that one of the following be used for each breast cancer review.
a) The Cochrane Collaboration tool for assessing Risk of Bias (please refer to Chapter 8 of The Cochrane Handbook for Systematic Reviews of Interventions). This handbook is available from http://209.211.250.105/resources/handbook/index.htm or the 'help' menu in RevMan. The risk of bias tool is available within RevMan software as an extension of the characteristics of included studies table. RevMan enables judgements to be made for each of the six domains. These can then be illustrated in two types of graphics.
b) An assessment tool based on the Schultz criteria (Schultz et al 1995)(http://www.ctc.usyd.edu.au/cochrane/conduct_review/CBCG%20quality%20assessment%20tool%20December%202006.pdf).
3. Authors are required to state in their protocol how quality will be assessed and any action to be taken based on quality. For example, will trials be excluded on the basis of their quality? Will a sensitivity analysis be undertaken? A sensitivity analysis will involve analyses where studies of poorer quality are included and then excluded, and any differences discussed.
4. Given the nature of the interventions used in the management of breast cancer it will be neither possible, nor practical, to expect blinding of the intervention or outcome assessment in the majority of reviews.
5. Wherever possible, two people should independently assess each potentially eligible trial for quality.
6. A third person should be used to resolve any discrepancies regarding quality.
7. For further guidance on the assessment of quality, authors should refer to The Cochrane Handbook for Systematic Reviews of Interventions.

Peer review
Each protocol and review is required to go through a system of peer review before it can be published on the CLib. The average length of time for peer review is 6-8 weeks but this may be lengthened depending on referee availability and the number of reviews already in peer review. Referees from clinical, statistical and consumer backgrounds are enlisted to provide comment on the work and these are fed back to authors for consideration and to guide amendments. Some referees may also wish to see the amendments made to protocols/reviews and this may also add to the length of time in peer review. The time required for comprehensive peer review may also impact on the publication schedule of protocols/reviews and authors should factor this into their time frames. The CBCG provides a list of publishing deadlines to assist with this. The CBCG also provides guidelines for dealing with peer review comments and guidelines for referees .

Plain Language Summaries (PLS)
Each review is required to have a PLS. This aims to summarise the review in a way which would be easily understood by lay persons accessing the information. PLS are made freely available on the internet, so will often be read as stand-alone documents. They should, therefore, contain a summary of the methodology, the main results and the objectives of the review.

Plain language summaries have two parts. The first part is a restatement of the review's title using plain language terms. The second part or body of the summary should be no more than 400 words in length and should include:

•  A statement about why the review is important: for example definition of and background to the health care problem, signs and symptoms, prevalence, description of the intervention and the rationale for its use.

•  The main findings of the review: this could include numerical summaries when the review has reported results in numerical form, but these should be given in general and easily understood forms. Results in the plain language summary should not be presented any differently from in the review (ie no new results should appear in the summary).Where possible an indication of the number of trials and participants on which the findings are based should be stated.

•  A comment on any adverse effects.

•  A brief comment on any limitations of the review (for example trials in very specific populations or poor methods of included trials).

The first draft of the plain language summary should be written by the review authors and be included in the review at the time of peer refereeing. This draft may be subject to alteration, and authors should anticipate one or more iterations.

Further information on the process of finalising plain language summaries is available in chapter 11section 11.9 from the Cochrane Handbook for Systematic Reviews of Interventions.

Updating
By undertaking to conduct a systematic review with the Breast Cancer Group, authors also agree to regularly update their review. The timeframe for this is negotiated between the contact author and the editorial base. If additional research has not become available since the last update of the review then authors will be asked to include a statement along these lines in the review.

Licence for Publishing
For each protocol and review, all authors must complete a 'Licence for Publication Form' and return this to the CBCG. Authors are advised to review the Cochrane Collaboration's policies regarding publication of Cochrane reviews in print journals (section 2.2 of the Cochrane manual).

Declaring Conflicts of Interest
All authors are required to complete a declaration of interest form at title registration and prior to publication or both the Cochrane protocol and review.