Minutes of the RevMan Advisory Group meeting held at the UK Cochrane Centre on 28 March 2001

Minutes approved on: 13 December 2001

Present: Clive Adams, Claire Allen (minutes), Gerd Antes, Chris Cates, Mike Clarke (Convenor), Katherine Deane, Monica Fischer, Sonja Henderson, Julian Higgins, Heather Maxwell, Tess Moore, Rasmus Moustgaard, Mark Starr, Lindsay Stead, Paul Whatling, Phil Wiffen.

Apologies: Phil Alderson, David Wilkinson.

Contents:
Welcome
Approval of the agenda
Approval of minutes and matters arising from the previous meeting (7 December 2000)
Combined wishlist (RevMan and MetaView)
Inverse variance meta-analysis in RevMan
Multi-user version of RevMan
Validation rules for Cochrane reviews before publication in The Cochrane Library
Presentation of most recent changes in Cochrane reviews
'Meta-analysis' of single trials - should it be avoided?
Needs assessment survey
Additional Graphs in Cochrane reviews
Membership of RevMan Advisory Group
Budget for the RevMan Advisory Group
Any other business
Date of next meeting
Apppendix

Welcome

Mike welcomed everyone to the meeting. Apologies for absence had been received from the people listed above. Mike advised that Helen Smith and Sue Wilson had stepped down from the Group.

Approval of the agenda

The agenda was approved, with the addition of 'Publication of minutes' to item 3.

Approval of minutes and matters arising from the previous meeting (7 December 2000)

The minutes were approved with no amendments and Rasmus agreed to upload them to the web site.
Action: Rasmus

Matters arising

1. Publication of minutes - Item 3.3.3

   Rasmus agreed to inform Jacob Riis of the discussions about publication of the minutes. Jacob will advise all entities of the web pages for the minutes of the Information Management System, RevMan Advisory and ModMan Advisory Groups.
   Action: Rasmus

2. Guidance on good outcomes - Item 3.3.11a

   Chris reported that the preparation of guidance on good outcomes was still pending until Jon Deeks had the results of some empirical work he was carrying out. Mike agreed to discuss this with the Handbook Group when appropriate.
   Action: Mike

3. Web-registering for RevMan 4.1 on CD-ROM - Item 3.12

   Monica agreed to remind all entities that there was a stock of CD-ROMs at the Nordic Cochrane Centre and that bulk ordering could be done via the Centre's web pages. A note will be added to this message to remind entities about the patch for RevMan 4.1 and availability of an Internet version for download.
   Action: Monica

4. Patch for RevMan 4.1 - Item 4.2

   Rasmus agreed to add the e-mail adminors@cochrane.de to the RevMan e-mail discussion list and to remove those RGCs who were listed individually. Monica agreed to ask Jini Hetherington to include this information in the introduction pack sent out to new RGCs.
   Action: Rasmus, Monica

5. RevMan wishlist

   Rasmus agreed to remind Jacob that he should feedback to people who had submitted wishes advising them that their wishes had been discussed by the RevMan Advisory Group and that minutes of the meeting are available on the web site.
   Action: Rasmus

6. MetaView bugs - Item 7

   Mark advised that no timescale was in place within Update Software to deal with the bug fixes.
   Action: Mark

7. Guidance on survival data - Item 8.1

   Mike agreed to write guidance for incorporating survival data, to be included in the RevMan User Guide.
   Action: Mike

8. Archiving of The Cochrane Library - Item 8.3

   Mark advised that discussions are underway with the Wellcome Library who are interested in making back copies of The Cochrane Library available through the Internet. These back copies would be searchable and people will have open access after an appropriate length of time has passed. He clarified that Update Software are willing to supply copies of versions of earlier reviews if requested and he agreed to inform all RGCs of this offer.
   Action: Mark

Combined wishlist (RevMan and MetaView)

Rasmus advised on the new format for the wishlist. He reported that each item has a unique ID number, which would remain with that wish regardless of how it was classified within the list. He is considering combining wishes that are similar in nature. Rasmus asked everyone to ensure that submitted wishes are clear and explicit about the action being requested. It was agreed that if wishes are submitted for any actions that are already possible, these would be added to the FAQ list. Rasmus will ask Jacob to check that this is noted in the help file.
Action: Rasmus, Everyone

The following subject areas, under which wishes could be grouped, were suggested:

• Statistics
• Printing
• Text
• Interface

The items on the wishlist (dated 27 March 2001) that had been classified as 'Discussed by RAG but require further investigation' and 'New items' were discussed (see appendix).
Action: Rasmus, Mike, Monica, Julian

It was agreed that Rasmus should add the inclusion of flowcharts, the ability to print the entire review as a single file, and the display of a warning if a review is highlighted but not selected when the export routine is run to the pending wishlist
Action: Rasmus

Inverse variance meta-analysis in RevMan

Julian reported that inverse variance meta-analysis had been recommended by the Statistical Methods Group and the Quality Advisory Group for inclusion in RevMan as soon as possible even though it was recognised that this would require a structural change. Rasmus agreed to investigate the possibility of doing this by using graphics files attached, but external to, RevMan since this would not require a mandatory release of a new version of RevMan. Rasmus expected that this could be implemented within one year, unless it would require a rewrite of MetaView. If this is possible and an interim version of RevMan is released, Rasmus agreed to consider including other items on the wishlist to this new version. He agreed to report on this to the next meeting of RAG.
Action: Rasmus

Multi-user version of RevMan

Rasmus reported that he had been developing a system that would allow more than one person to use a copy of RevMan at the same time. It is now ready for testing. Sonja and Lindsay agreed to test it. Rasmus advised that the new programme is for people using a network and have particular requirements for shared access (ie editorial bases of Collaborative Review Groups). Once Sonja and Lindsay have tested it, the new software will be made available to all RGCs for assessment. Rasmus agreed to ensure that the exported files would be identical to those from the single-user version. It was agreed that if the tests are successful future development of RevMan should build on the multi-user version, rather than either the single user version or both versions.
Action: Rasmus, Sonja, Lindsay

Validation rules for Cochrane reviews before publication in The Cochrane Library

Sonja reported the discussions that took place at the ModMan Advisory Group (MAG) meeting on 27 March 2001 in regard to the implementation of the validation rules related to the mandatory sections of Cochrane reviews. She advised that MAG would be recommending that these validation rules were included in ModMan 4.1 (this had not conflicted with anything previously agreed at RAG). It was expected that very few reviews would currently fail these rules and Lois Sims had agreed to notify the individual RGCs as soon as possible of those reviews that failed the validation checks.
Action: Mark

It was agreed that the validation checks in RevMan should include an option not to validate reviews on export (for example, for draft versions and non-primary versions). Mark noted that, from May 2001, Update Software will check all reviews and protocols to make sure that the titles and authors fields have been completed. Update Software will communicate any problems to the relevant Review Group Coordinator (RGC) immediately, so that these can be corrected prior to publication. Update Software will discuss with RGCs a realistic timescale for implementing the full list of validation rules. In preparation for this, they will provide a report to RGCs of reviews or protocols that would currently fail the validation rules. Mark noted that once the date for implementation has been agreed, this would not lead to the withdrawal of reviews from The Cochrane Library. Instead, if problems cannot be corrected in time, new reviews would not be published until they had been made valid and the previous version of an updated review would be published in place of the update that did not have text in the mandatory fields. Rasmus will add validation checks to the wish list for RevMan.
Action: Rasmus

Presentation of most recent changes in Cochrane reviews

Mike advised that RAG were awaiting suggestions from the Quality Advisory Group on guidance for what should be included in the 'Whats new' text box in RevMan. Mark advised that the content of this box should be revisited by the editorial bases with every submission.

'Meta-analysis' of single trials - should it be avoided?

This issue arose from a question of whether CRGs should encourage reviewers to make graphs if it is not possible to do a meta-analysis. Alternatives to the presentation of single trials as a "meta-analysis" were suggested. These included discussion in the text alone, use of tables or inclusion in a meta-analysis graph as sub-groups. Mike agreed to discuss this with the Handbook Advisory Group.
Action: Mike

Needs assessment survey

Mike reported on the current status of the survey. The pilot survey had gone well and there had been some suggestions for design changes. However, it was unclear how to proceed and whom to involve in the full survey. After discussion, it was agreed to try to survey all reviewers and co-reviewers with a published review or protocol in The Cochrane Library, all participants on RevMan or Protocol Development workshops over the past two years and contact people who had registered titles that had not yet proceeded to protocols. The survey will be distributed centrally. It will be announced using an e-mail which includes a link to a web site containing the survey. For those people who request it, the survey could be sent as an attachment or as a paper copy. Mike agreed to promote the survey through the Collaboration's e-mail discussion lists, CCINFO and Cochrane News.
Action: Mike

Additional Graphs in Cochrane reviews

Gordon Dooley, Tess and Rasmus had tried to identify ways of including additional graphs in Cochrane reviews without creating any new software. Sonja reported that the ModMan Advisory Group had decided not to endorse the proposal for the interim solution (ie where no software changes were required) because of the propensity for introducing errors. Mark noted that it would not be possible to introduce an add-on system that was not complex or error-prone.

Phil and Tess agreed that PaPaS and Heart would act as pilot CRGs for testing the suggested system of including additional graphs as suitable named .gif files. Julian advised that a statistician should check any additional graphs that were to be included. It was agreed that Rasmus and Gordon would work with Phil and Tess on this project and Mark would notify Gordon of this.
Action: Phil, Tess, Rasmus, Mark

Membership of RevMan Advisory Group

Mike noted that new members of the Group would be welcome - in particular reviewers. Suggestions for extra members should be sent to Mike.
Action: Everyone

Budget for the RevMan Advisory Group

Monica advised that the Cochrane Collaboration Steering Group had agreed a budget of £1,168 for the RevMan Advisory Group for this year, to cover travel, accommodation and meeting costs for the present meeting and the next one (December 2001). Monica will distribute expenses forms and advised that original receipts should be attached to all claims, which will be processed by the Nordic Cochrane Centre. Mike agreed to keep track of the remaining budget once people had claimed their expenses.
Action: Mike, Monica

Any other business

There was no other business.

Date of next meeting

Mike will canvass for availability for the next meeting, which would probably be held around the second week of December.
Action: Mike

Apppendix

Summary of discussion of items on the wishlist (dated 27 March 2001) that had been classified as 'Discussed by RAG but require further investigation' and 'New items'.

Item	Notes	Priority
043	Mark advised that headings will wrap in the future and it was agreed column headings should be 30 characters long.
044	Mike agreed to ask Miranda Mugford (convenor of the Economics Methods Group) for advice, once Monica had told him who had submitted this wish.
045	Jacob is still awaiting guidance from the submitter.
046	This had been discussed when the software was produced initially.	Low (section B)
047	Rasmus agreed to add this to the FAQ list, as it is already possible in RevMan 4.1
048	Dates for 'What's new'.	Low
049	Bold, italics and underline could be added to the next interim release of RevMan but superscript and subscript would require a structural change. When bold, italics and underline are added, the guidelines for authors should be amended to include information on when to use different textual formatting.	High
050	Julian agreed to ask the Statistical Methods Group for advice and Mike agreed to ask the Quality Advisory Group for advice.	Rejected
051	A structural change would be needed to allow more than 99 references per study	Low
052	Save data option relating to calculated data.	Rejected
053	All graphs to be able to be exported as bitmaps.	High
054	'Vector based' graphics would help reduce the size of graphics files and would allow the size of graphs to be altered without loss of resolution.	Medium
055	Copying graphs to clipboard.	Medium
056	Option to print several graphs on one page.	High
057	This can already be done. Rasmus agreed to add it to the FAQ list and ensure it was listed in the help file.	Medium
058	Option to select which columns are displayed on the summary screen. This had been part of the specification for MetaView 4.	High
059	Saving window size between sessions.	Medium
060	Reduce the gap between the scale and the body of the graph when printing.	High
061	Previous and next outcome buttons to be enabled without having to select outcomes.	High
062	Scaled font size when printing so that a graph fills the width of the paper.	High
063	Width of saved bitmaps should depend on displayed columns.	High
064	This wish was for bitmaps to be saved in colour. This had been the intention in the programming but there is a bug in the programme which means that although the files are saved as 16 colour files they only use black and white. It was agreed that the bug should be fixed so bitmaps are saved in colour and to add an option to save bitmaps in black and white (to reduce file size). It was also suggested to add an option to print in pure black and white. This should be added to the wish list as a separate wish.	High
065	ISRCTN is a unique study identifier that will be assigned to new randomised trials by Current Controlled Trials. It is also likely to be applied retrospectively in the future. The addition of a new field to store this would require a structural change to RevMan but it was assigned a high priority for any future development to the software because of its importance as a study (rather than a reference) identifier.	High
066	Mike agreed to suggest to the Handbook Advisory Group that the guidance in the Handbook should be changed to reflect the fact that 'Contributions to the review' should contain only the contributions of the authors who are named in the by-line for the review.
067	It was agreed that RevMan should call attention to the backup file and suggest that it be opened it if RevManit detects that it crashed in the last session. Note: It is already possible to open the backup file manually if RevMan crashes. Rasmus agreed to send a message to adminors about this [it is already on the FAQ]. Action: Rasmus
068	Rasmus agreed to request advice from the Statistical and IPD Methods Groups.
069	Tabbing between fields in reference windows.	High
070	The ability to see simultaneously the entire contents of both sets of contact details, when RevMan identifies a discrepancy on importing.	High
071	Importing references directly into the relevant category of study. This was rejected because this would mean that RevMan would have to create new studies for each reference. Instead it was agreed that:	Rejected
071	(a) it should be possible to move references in blocks	High
071	(b) it should be possible to import whole studies (including references) into the 'studies awaiting assessment' section (this would require a new import format)	High
071	(c) it should be possible to import references under an existing study.	High
072	User defined order should be the default in MetaView, if such an order has been entered into RevMan.	Medium
073	Track changes while editing option.	Deferred

RAG minutes | MinMan home