Methods used in Cochrane PaPaS reviews

[This page is due to be updated]

Search strategies

Our Information Specialist will contact all authors who have new registered titles within two to four weeks after registration to arrange drawing up your search strategy for your protocol (only one example of a search strategy is expected for the protocol and all for the full review). Our Information Specialist will also then check and comment on any search strategies submitted with the protocol in case any changes have been made. Review authors are asked to search MEDLINE, EMBASE, the Cochrane Central Register of Controlled Trials (CENTRAL) and any relevant specific databases such as PsycINFO orCINAHL. Our Information Specialist will run the searches for our authors on MEDLINE, EMBASE and CENTRAL, unfortunately due to funding we cannot always support the running of searches on other databases for our authors so authors  may be asked to search these other databases themselves. Authors are also expected to search clinical trials registries for ongoing trials.

The PaPaS review group's policy is to follow the standard methodology for carrying out systematic reviews as detailed in 'Cochrane Handbook for Systematic Reviews of Interventions'. A hypertext version of the Cochrane Handbook is accessible from the Help menu in Review Manager (RevMan); or through Cochrane's web site: cochrane.org. The Handbook is available to purchase as a book from Wiley-Blackwell, but is also available as individual chapters in PDF format through the Archie website (username and password required) - if you have problems finding this please contact the Managing Editor.

Access to specialised register by reviewers

The PaPaS review group's trials register of pain, palliative and supportive care trials comprises approximately 43,000 trials. Of these, 42,057 are indexed in the Cochrane Central Register of Controlled Trials (CENTRAL) in the Cochrane Library Issue 10, 2011.

Additional search strategies

The PaPaS review group's policy is to follow the standard methodology for carrying out systematic reviews as detailed in 'Cochrane Handbook for Systematic Reviews of Interventions'. A hypertext version of the Cochrane Handbook is accessible from the Help menu in Review Manager (RevMan); or through Cochrane's web site: cochrane.org. The Handbook is available to purchase as a book from Wiley-Blackwell, but is also available as individual chapters in PDF format through the Archie website (username and password required) - if you have problems finding this please contact the Managing Editor.

Study selection

The policy for 'Additional search strategies' described in the section above also applies to study selection and we advise review authors check the Cochrane Handbook for Systematic Reviews of Interventions for information on how to select studies for their review.

Assessment of methodological quality

The PaPaS review group encourages the use of the quality tool devised by Dr Alejandro Jadad and the Pain Research Unit, Oxford. See 'Instrument to measure the likelihood of bias in pain research reports' described in the "Additional Information" and also the "References" sections, both in the 'About' section on the Cochrane Library. We also encourage you to use both the new Risk of Bias table and Summary of Findings table available in RevMan 5.

In addition, PaPaS requires that quality is assessed independently by two review authors. The agreed quality assessment should be reported in reviews in the 'Characteristics of included studies' table. Disagreements between members of the review team regarding quality of included trials should be resolved through discussion. 

Data collection

PaPaS has developed generic forms for data collection, and review authors can obtain copies from the editorial office. The extraction of data from trial reports should be undertaken independently by at least two review authors. Differences in data interpretation between review authors' should be resolved through discussion. It is not a requirement that review authors should routinely seek unpublished studies, as experience has shown that contacting authors can be unproductive. However, it may be an appropriate course of action for some review topics, and review authors should discuss this with the editorial office.

Analysis

Data entry to RevMan may be undertaken by only one review author but must be checked independently by at least one other author. 

Statistical guidance is available from the group's statistical editor, Gavin Stewart, through the editorial office.

We would like review authors to include NNTs (number-needed-to-treat-to-benefit) where appropriate, in addition to the analyses available in RevMan. NNTs should be described in the 'Methods of the review', and the outcome included in the text of the 'Results' section of RevMan. An 'NNT calculation' spreadsheet (Excel 5.0) is available from the editorial office on request.

Reporting of reviews

Where possible, review authors should comment on the applicability of their findings, and include any information relevant to health-care decision making, such as adverse events and adverse effects, prescribing data and costs.