Personal tools
You are here: Home → IMSG → RAG → Minutes → RAG minutes, 15 January 2004
 

RAG minutes, 15 January 2004

Minutes approved: 2 December 2004

Present: Phil Alderson, Jacqueline Birks, Chris Cates, Mike Clarke (Convenor), Katharine Deane, Sonja Henderson, Dymphna Hermans, Julian Higgins, Monica Kjeldstrøm, Valeria Marinho, Heather Maxwell, Tess Moore, Rasmus Moustgaard, Simon Newton, Elizabeth Pienaar, and Lindsay Stead.

In attendance: Claire Allen (minutes), Andrew Cullis, Peter Friis, Paul Glasziou (for item 5), Pauline Howarth and Jacob Riis.

Not present: Clive Adams, Gerd Antes, Bill Macguire and Carrol Preston.

Contents:

  1. Welcome and apologies
  2. Approval of the agenda
  3. Approval of minutes and matters arising from the previous meeting (10 September 2003)
  4. Update on the IMS development
  5. Inclusion of Balance Sheets in Cochrane reviews
  6. Content of What's New section of Cochrane reviews
  7. Specification for RevMan 5
  8. Functions in each section of a Cochrane review
  9. Additional figures in Cochrane reviews
  10. Recalculation of analyses in Cochrane reviews after publication
  11. Notification of structural changes to the publisher
  12. Search strategies in a single-celled additional table
  13. Deletion and withdrawal of reviews
  14. New version of the Cochrane Reviewers Handbook
  15. Remit of the RevMan Advisory Group
  16. RAG representation for methodology and diagnostic reviews
  17. Any other business
  18. Date of next meeting

  1. Welcome and apologies

    2. Mike welcomed everyone to the meeting and everyone introduced themselves. Apologies had been received from Gerd Antes and Bill Macguire.

  2. Approval of the agenda

    3. The agenda was approved.

  3. Approval of minutes and matters from the previous meeting (10 September 2003)

    4. The minutes were approved with the following addition:

    Item 5.a.i - Date of last minor update: It was agreed to remove this date in the future version of RevMan

    Rasmus agreed to make this amendment and revise the status of the minutes on the IMSG web site to reflect this.
    Action: Rasmus

    Matters Arising

    1. Item 4b – Wishlist for RevMan:

      2. Julian reported that the Statistical Methods Group Convenors would like to implement an additional statistical method (‘inverse variance’) for analysis of dichotomous data and a new test for differences between sub-groups. This was supported by the RAG and Julian agreed to send the specification to Rasmus.
      Action: Julian

    2. Item 4c – List of journals

      3. Monica reported that she had asked Kay Dickersin to raise the issue of a list of journals for a standard pick list, with the Cochrane CENTRAL Advisory Group.

    3. Item 5a ii – MeSH terms in reviews:

      4. Monica advised that the IMS team will discuss the inclusion of a field for MeSH terms in a review with representatives of Wiley at their next meeting on 5 May 2004.
      Action: Monica

    4. Item 5f iii – Scales for graphs

      5. Julian reported that this had not yet been discussed by the Statistical Methods Group Convenors but that the Statistical Methods Group would be likely to approve the addition of more options.
      Action: Julian

    5. Item 5f ix – Additional headings

      6. Phil reported that the Handbook Advisory Group (HAG) had not discussed this issue at their meeting in Barcelona. He advised that the HAG would revisit all the headings in reviews to see if it is appropriate to include additional headings in any other section to help make the structure more logical. Monica agreed to let Phil know the timescale for including additional headings in RevMan 5.
      Action: Phil, Monica

    6. Item 6 – Dates relevant to Cochrane reviews in RevMan and CDSR

      7. Monica agreed to raise this issue with the ModMan Advisory Group (MAG).
      Action: Monica

    7. Item 8 – Search strategies in additional tables

      8. Phil reported that the HAG had discussed the inclusion of search strategies in additional tables. He reported that guidance is being prepared for Section 5 of the Cochrane Reviewers’ Handbook.

    8. Item 9 – Style guidance for levels of headings in Cochrane reviews:

      9. Mike advised that the revised guidance for headings had been approved by the PPG and Sonja reported that this had been added to the Style Guide.

  4. Update on the IMS development

    5. Monica explained that the development and implementation of the new IMS was on schedule, and most of the document handling module had been developed.

    Monica explained that documents would be checked in and out of the new IMS using RevMan 4.3, which will have the ability to communicate with the central server. When the document management system is complete, access control and the new interface for the Contact Database will be developed. This will be beta tested in April 2004. The timelines had been agreed with Wiley and Monica was keeping them informed of progress. She explained that RevMan 5, including all agreed structural changes would be released in 2006, after the general release of the new IMS (during 2005).

    Monica confirmed that the Steering Group had agreed that the Contact Database would be mandatory for Cochrane entities from July 2004. Monica had e-mailed CRGs about this. Jacob confirmed that he would be working with RGCs to help them to move data into the Contact Database to ensure that the details will be as up to date as possible. This would begin by importing the contact details from the current version of the Parent Database (ie issue 1, 2004 of The Cochrane Library) but it was recognised that further work would be needed to add information on people who had become contact reviewers since the submission date for that issue (November 2003). Rasmus offered to investigate if it would be possible to identify these people automatically in subsequent versions of the Parent Database. Entities would need to keep their members’ details up to date on the Contact Database after July 2004. Jacob explained that each reviewer will have one contact ID and one record within the Contact Database. The contact ID will become eight characters long and will be generated using a combination of the Collaborative Review Group code and the RevMan 4.2 Contact ID. Monica agreed to send documentation on this to RGCs.
    Action: Monica, Rasmus

    Monica reported that the new IMS will be beta tested first by two or three of the CRGs based in Copenhagen (who are not currently represented on MAG or RAG). The IMS team will meet them on a monthly basis to demonstrate developments to the system, test it, and help solve any problems encountered in the previous month. Data from these CRGs would be the first to move to the new IMS. Monica would also ask the Methodology Review Group if they wished to join the pilot. She had arranged regional IMS training days for all CRGs to demonstrate the new IMS, and to enable participants to return to their editorial base and plan for the future. The milestones document has been put on the IMS web site. CRGs are continuing to be provided with updated information through the IMS web site, and the dates of the IMS training days were announced to the relevant e-mail discussion lists. Monica agreed to update CRGs on progress on a monthly basis, but it was noted that RGCs wished to be responsible for dissemination of necessary information to their reviewers (given that most of the information would relate to the editorial bases, rather than reviewers).
    Action: Monica

    Monica reported that any major structural changes to reviews in RevMan 5 had to be presented to, and discussed by, the Steering Group at their next meeting on 29 February to 2 March 2004. She confirmed that it might be possible to delay approval of minor changes until July 2004.

  5. Inclusion of Balance Sheets in Cochrane reviews

    6. Mike introduced Paul Glasziou to the meeting and said that Paul had been asked to attend because of the need to prepare the specification for RevMan 5, even though the concept to be presented by Paul was still under discussion. Paul’s presentation was on the proposal from the Applicability and Recommendations Methods Group and the GRADE Collaboration (via Andy Oxman) to include a ‘balance sheet’ or ‘summary of findings’ in each review. This would help users to access the results of a review quickly, enabling clinicians to find and understand relevant information rapidly when making health care decisions. He explained that the abstract of a review does not necessarily contain all the information needed to do this.

    Paul also stressed that information about the safety of the intervention is not usually included in the abstract. He asked whether it is possible to include the summary of findings in the current version of RevMan, and what needed to be included in specification of RevMan 5 to enable this. Julian expressed concerns about reviews that include insufficient evidence and how the results of these would be displayed. He also had concerns about the use of NNTs in reviews and the dangers of automatic generation of ‘Summary of findings’ tables. Paul said that these concerns had been raised by others and would be discussed further during the Ottawa Colloquium. Julian agreed to discuss the elements of the summary of findings table with the Statistical Methods Group.
    Action: Julian

    Monica explained that a specific table for the ‘Summary of findings’ could be included in the structure for a future version of RevMan, but that currently, reviewers would need to create this in the additional tables section of RevMan. It was agreed that reviewers who wished to include such tables before the release of RevMan 5 should use this option. Paul reported that the Acute Respiratory Infections and Renal CRGs had agreed to pilot the new table. He suggested that the summary of findings table could eventually be displayed directly below the abstract so that it is easily visible to users but accepted that the placement of the table needed to be discussed at a later date. He explained that the second phase of the discussions on this would involve working with the IMS team and the Statistical Methods Group to develop an advanced proposal, investigating an expansion of the functions available within the table in the longer term. Rasmus explained that some changes to the tables in RevMan had already been agreed, making them more sophisticated, and allowing merging of cells and linking within cells. It was agreed that if Summary of findings tables were introduced to RevMan, there should be adequate guidance in the Cochrane Reviewers’ Handbook.
    Action: Paul, Julian, Monica

    Rasmus raised concerns that a complex table including graphics and buttons might be problematic for publishing, but Simon explained that it was dependent on what type of information is to be included in the table. Simon confirmed that Wiley would be able to display the summary if it is created as a normal table with columns and rows, but if it becomes more sophisticated, it could possibly become more difficult to display. He explained that if it appeared as a stand alone table displayed below the abstract, the publisher would need to be able to identify it separately from other additional tables. Andrew suggested that before the structure for the Summary of findings table was finalised, the IMS team and Wiley would need to make sure that the XML structure could cope with them. Discussion took place as to whether any of the elements of the summary of findings table could be automated, but it was agreed that some manual intervention was also needed.
    Action: Monica, Andrew

    Paul suggested that the ‘Summary of findings’ tables might eventually include information from both the pooled analysis, but also as projections for people with different prognoses. Julian stressed the need to consult the Statistical Methods Group about these proposals and Paul agreed that this would be done.

    Pauline explained that it was unlikely that Wiley would publish the table within the abstract. Paul agreed to consult with Co-ordinating Editors at their meeting in Ottawa. Monica and Rasmus agreed to plan the review format in RevMan 5 to allow one or more tables to be identifiable to the publishers as ‘Summary of findings’ tables. Monica confirmed that formatting and manipulating of tables will be easier in RevMan5, and discussion of where to display the tables would take place at a future RAG meeting.
    Action: Monica, Rasmus, Mike

    Monica advised that the agreement to include a separate Summary of findings table in Cochrane reviews would be a structural change, and therefore it should be included in the IMS report to the Steering Group.
    Action: Monica, Mike

  6. Content of What's New section of Cochrane reviews

    7. Mike advised that the Publishing Policy Group (PPG) had received an e-mail from Peter Gøtzsche suggesting that the PPG consider the name of the ‘What’s new’ section in RevMan, which is currently labelled ‘Most recent changes’ in The Cochrane Library, and that the PPG had passed this request to the RAG. Mike explained that one of the things to discuss was whether to overwrite the text of this section each time the review was updated, or to keep the old text within the section, thereby, creating a history of the review. Lindsay suggested that previous text from the ‘What’s new’ section should be archived in the published notes field. Chris suggested that the section could be used to inform the user of, for instance, any new studies, names of new studies, and whether the conclusions have changed. He also suggested that hyperlinks could be added for new studies. Sonja suggested that the history should not be contained in the ‘What’s new’ section, but that the section should be linked to the history section of the review. It was agreed that better guidance is needed for reviewers, so that they complete these sections correctly. Phil agreed to discuss this with the Handbook Advisory Group. Andrew noted that, in the future, it should be possible to trace the history of each review through its DOI (substantive updates only), but that users will be encouraged to use the most recent version of the review.

    Monica explained that the new IMS will have a built in history for Cochrane review stored in the tracking system, but this will not be published. After discussion, it was agreed that the ‘What’s new’ section should be used to inform the user of useful information about changes to the review. It was agreed that, in RevMan 5, it should be configured as a two column table, the first column for a date field, and the second for a text field. Phil agreed to take this to the HAG so that guidance could be provided in the Cochrane Reviewers’ Handbook. It was agreed that The Cochrane Library should label the section ‘What’s New’ rather than ‘Most recent changes’, and that it should appear in the Outline for a Cochrane review. Mike agreed to raise this with the PPG.
    Action: Phil, Mike

  7. Specification for RevMan 5

    1. Dates

      2. Mike reported that he had recently circulated a proposal on the storage of dates in reviews to CRGs and The Cochrane Library Users Group, based on the lengthy discussions at the RAG meeting on 10 September 2003. He asked for further comments now. Rasmus questioned the need for both the date a review had been assessed as up to date, and the date of the last search. The RAG had agreed previously that both dates should appear, and following discussion, this agreement was reaffirmed. Pauline had noticed that some full reviews had substantive update dates that preceded the conversion of the protocol into a review and asked if the date of this conversion should be automatically inserted as the date of last substantive update. Monica and Rasmus agreed to implement this.
      Action: Monica, Rasmus

      Tess reported that Hilda Bastian and Philippa Middleton had found that reviews were marked as substantive updates inconsistently. It was agreed that the guidance on the use of the date fields in RevMan needs to be improved.

      Following further discussion of how to store dates of revisions in Cochrane reviews, it was agreed that there should be a single ‘Date revised’, with the following tick box options: ‘Substantive update’, ‘Minor update’, ‘Comment added’, or ‘Response to comment added’ and these boxes would not be mutually exclusive. It was also agreed that two dates labelled ‘Date last search done’ and ‘Date last assessed as up to date’ should be added
      Action: Rasmus

      Simon confirmed that Wiley would check whether the tick box is checked in order to flag a review as, for example, “updated”.

    2. Reference and study ID lengths

      3. It had previously been suggested to expand the reference identifier to 50 characters and in the absence of any technical limitations, this was agreed.

    3. Study ID lengths:

      4. It had previously been suggested to expand the study identifier to 30 characters, and in the absence of any technical limitations, this was agreed.

    4. New statistical method

      5. Rasmus agreed to discuss the new inverse variance algorithm for dichotomous data with Brendan Gibson when he has arrived the specification from the Statistical Methods Group.
      Action: Rasmus

    5. Scales on graphs

      6. Julian said that he was unsure of what needed to be discussed in relation to scales on graphs but Chris explained that it is a display issue and that reviewers would like more options than those currently available. It was agreed that the ideal would be for reviewers to be able to define their own endpoints for the scale and Julian reported that the Statistical Methods Group would support this. Monica and Rasmus will liaise with Wiley to define what is publishable.
      Action: Monica, Rasmus

    6. Quality assessment

      7. Mike pointed out that this relates to whether allocation concealment should continue to be the only quality grading system for studies in reviews. Julian reported that several people had expressed a desire for more options. He agreed to discuss this with the Handbook Advisory Group and Monica agreed that it would be possible to consider any possible structural changes until July 2004. It was agreed that if any new systems were added, they should include an option for ‘not assessed’.
      Action: Monica, Julian

    7. Headings

      8. Phil explained that the Handbook Advisory Group had not discussed the issue of extra headings in the main text of reviews. Rasmus explained that the discussion should be based on whether the headings were to be optional or fixed. Phil pointed out that there could be two possibilities, firstly a pick list of optional headings, and secondly, including the headings but allowing reviewers to delete them. It was agreed that it would be helpful to have optional extra headings and Phil agreed to discuss these with the Handbook Advisory Group and to inform Rasmus of their recommendation.
      Action: Phil, Rasmus

    8. Erratum

      9. Monica explained that including an erratum section has been discussed by the Publishing Policy Group and the RAG agreed that the facility to have formal errata was important. Simon reported that MEDLINE dictates that an erratum is only noted within a MEDLINE record if it is published separately to the original document. Simon and Andrew agreed to discuss with MEDLINE whether it would be possible to replace erroneous records, rather than publishing an erratum. However, if this was not possible, it was agreed that a mechanism should be introduced within the new IMS to pass erratum to the publisher for inclusion in The Cochrane Library (and, hence, recording in MEDLINE).

      Action: Simon, Andrew

    9. Review Manager wish list – New items

      10. It was agreed to add the ability to include flowcharts via a simple drawing option within RevMan to the wishlist in keeping with what Rasmus reported as being possible at the RAG meeting on 12 September 2003.This item was ‘Approved’ and Rasmus agreed to add it to the wishlist.

      ID Decision made by the RAG
      193 Approved as part of item 8
      195 Approved
      197 Approved
      198 Approved
      199 Approved to three decimal places
      200 Approved as high priority for RevMan 4.2.4. It was agreed that the Contact Database should identify people with the same name, e-mail address, and ID.
      204 This request was replaced by the discussion in item 7a.
      205 Approved as a special reference type – Mike agreed to raise this issue with the CCAG representative attending the IMSG meeting. It is important to cite Internet references accurately.
      208 Approved, low priority
      209 Approved. The programme should start in the way that it was left when it was last used.
      210 Approved
      214 Approved
      215 This has already been dealt with by the new IMS in version control.
      216 Approved by updating the user-defined spell-check dictionary packaged with RevMan.
      217 This has already been dealt with in the new IMS.
      218 Rejected
      219 Approved to design the format so this will be possible.
      220 Approved as part of item 8
      228 Approved
      229 Approved
      230 Approved
      231 Wiley have agreed to investigate a mechanism to export citations for Cochrane reviews from The Cochrane Library so that these can be imported easily into, for example, RevMan.
      232 Approved. Examples should be one click away.
      233 Approved
      234 Approved
      236 The IMS team will investigate how the new IMS will deal with the examples that Andrew raised.

      Jacob raised an additional item relating to referencing results in the text of Cochrane reviews. This wish has previously been rejected, but he suggested that there could be an automatic mechanism whereby if the results were changed, this would automatically update the text, with a validation check to ensure the reviewer had changed the sentence if appropriate. He agreed to liaise with Julian about this.
      Action: Jacob, Julian

  8. Functions in each section of a Cochrane review

    9. Rasmus, Lindsay, Harriet MacLehose and Mike had worked on a grid indicating which functions should be available in which parts of Cochrane reviews. Mike reported that they had not been unanimous in their recommendations but that he had prepared a summary, which was included in the papers for this meeting. The RAG discussed each cell in this grid and agreed the following changes:

    Contributions – Add: Subscript, superscript, internal links and embedded tables.

    Sources of support - Add: Superscript and subscript.

    What’s new - Add: Lists and embedded tables.

    Keywords/MeSH terms - Remove: Bold, italics, underline, subscript, superscript and special symbols.

    Synopsis - Remove: Lists.

    Abstract - Remove: Lists.

    Study IDs - Add: Special symbols.

    Reference IDs - Add: Special symbols.

    Reference field TI - Same as title

    Reference field AU - Same as byline

    Other ref. Fields - Remove: Bold, italics, underline, subscript, superscript and special symbols.

    Other data tables - Add: Lists, Internal links and Internet links.

    Contact details - Remove: Subscript, superscript and internet links.

    It was agreed that another category should be added labelled ‘Synopsis title’ to reflect the decision under the relevant wish list request in item 7i and that the functions available for this should be identical to ‘Title of review’. Rasmus agreed to revise the grid with these changes for use in the specification for RevMan 5.
    Action: Rasmus

  9. Additional figures in Cochrane reviews.

    10. Mike explained that after the RAG and IMSG meetings in September 2003, the need for better guidance on additional figures was identified. Monica had written to the Convenors of the Handbook Advisory Group, RAG and MAG, and Simon had prepared a checklist as an aide memoir.

    The MAG had recommended that if CRGs accept additional figures, they should be of suitable quality and if not, the publisher has the ability to refuse to publish the additional figures. Rasmus explained that a pilot was continuing whereby additional figures can be sent to him by RGCs. He would require that the checklist should be completed before figures were sent to him, with an indication of any problematic areas. He suggested that the checklist could be simplified to make it easier to use for validation. Phil agreed to discuss including the checklist in the Cochrane Reviewers’ Handbook with the Handbook Advisory Group.
    Action: Phil, Simon

    The RevMan team and Simon agreed to modify the checklist in accordance with the discussions. It was agreed that validation checks should be included for additional figures and Rasmus agreed to implement this in RevMan 5.
    Action: The RevMan Team, Simon

    Rasmus explained that currently legal copyright is referred to on the permission for publication form, but it was agreed that copyright on an additional figure should be included with the figure and the checklist should be expanded to note that if copyright had been obtained. The RevMan team and Simon agreed to be jointly responsible for updating the checklist and to include it in the RevMan documentation. Phil agreed to make sure that the final version of the checklist is included in the Cochrane Reviewers’ Handbook, and circulated to the RAG.
    Action: Simon, The RevMan team, Phil

  10. Recalculation of analyses in Cochrane reviews after publication

    11. Mike explained this issue had been discussed at the last RAG meeting. Julian explained that the Statistical Methods Group would be discussing this issue at a future date and agreed to report back to the next RAG meeting.
    Action: Julian

    Monica and Andrew are continuing to discuss this issue, and Andrew explained that Wiley are developing their own software to provide the functionality of MetaView. Julian expressed concern on behalf of the Statistical Methods Group that the new software must be properly beta tested and piloted. He also suggested revisiting the issue of calculated data. Andrew confirmed that Wiley will discuss this issue with the Statistical Methods Group and the IMS team.
    Action: Monica, Andrew, Julian

  11. Notification of structural changes to the publisher

    12. Simon reminded RAG that there is a requirement to notify Wiley six months in advance of any structural changes in Cochrane reviews. Monica confirmed that the majority of structural changes in RevMan 5 would need to be approved by the Steering Group in their meeting on 29 February to 2 March 2004 and that Wiley would be informed about structural changes more than six months before releasing RevMan 5.
    Action: Monica

  12. Search strategies in a single-celled additional table

    13. Simon summarised the difficulties caused by using single cell additional tables for search strategies if they are longer than one page. Simon agreed to identify to Mike those reviews for which this could be a problem and to liaise with Rasmus and Jacob to propose ways in which these should be dealt with. Heather reminded the RAG that search strategies could be included in an appendix in RevMan 5. In the meantime, she agreed to inform TSCs that search strategies might appear better in The Cochrane Library if they were divided in the additional table into, for example, 10 rows per cell.
    Action: Simon, Rasmus, Jacob, Heather

  13. Deletion and withdrawal of reviews

    14. Mike reported that the current policy is that withdrawn protocols should be deleted from The Cochrane Library after the inclusion of a withdrawn notice for one issue. Simon suggested that rather than deleting these in future, there should be a perpetual statement indicating that the publication has been withdrawn/deleted, and perhaps some text explaining why the publication has been withdrawn. He confirmed that the last published version of withdrawn reviews and protocols will be traceable through the DOI. Simon’s suggestion about the perpetual inclusion of a notice about withdrawn protocols will be considered at a future meeting. In the meantime, the existing policy remains in place.

  14. New version of the Cochrane Reviewers Handbook

    15. Phil reported that the new version of the Cochrane Reviewers’ Handbook included a major revision to the statistics section and would be released with the next issue of The Cochrane Library on 26 January 2004 and uploaded to the Collaboration web site shortly thereafter. He asked if this new Handbook could be included in the next RevMan patch and stated the importance of this. Katharine agreed that it was easier for reviewers if the Cochrane Reviewers’ Handbook was packaged in RevMan rather than having to download it. It was agreed that Phil should send a message asking Mark Davies (the reviewers’ representative on the Steering Group) to send the Internet link for the Cochrane Reviewers’ Handbook to the reviewers e-mail discussion list. Monica agreed to discuss this issue with Phil to ensure that the patch containing the new Handbook is released within the next few weeks.
    Action: Phil, Monica

  15. Remit of the RevMan Advisory Group

    16. Mike noted that he had prepared a draft of the revised remit for the RAG and this was in the papers for this meeting. Everyone agreed to consider this so that it could be discussed and finalised at the next RAG meeting (if not before, by email).
    Action: Everyone

  16. RAG representation for methodology and diagnostic reviews

    17. Rasmus reported that RevMan 5 will be able to handle three types of review: health care interventions (ie as now), methodology and diagnostic tests. The reviewer will choose which kind of review they wish to prepare. Discussion took place about whether diagnostic and methodology reviews should have their own advisory group equivalent to RAG. It was agreed that it would be preferable to ask the Methodology Review Group, and the Screening and Diagnostic Studies Methods Group to suggest one representative each for the RAG. Mike agreed to invite them.
    Action: Mike

  17. Any other business

    1. Change of terminology from ‘reviewer’ to ‘author’

      2. There was general discussion of whether the term ‘reviewer’ should be changed to ‘author’, and the suggestion ‘review author’ was also mentioned. Mike agreed to put this item on the agenda for the next RAG meeting.
      Action: Mike

  18. Date of next meeting

    19. It was agreed that a meeting at the usual time (April) was not necessary and Mike agreed to canvass people’s availability for a meeting later in the year (in June or September 2004).
    Action: Mike

Document Actions