Minutes of the Software Development Group special meeting held at The Hilton Hotel, Terminal 4, Heathrow, London, UK on 30 May 2000

Minutes approved on ?

Present: Chris Cates, Mike Clarke, Monica Fischer (Chair), Julian Higgins, Peter Langhorne, Mark Starr, Phil Wiffen.

Apologies: Sonja Henderson.

In attendance: Claire Allen (minutes), Jon Godwin, Rasmus Moustgaard.

Contents:
Welcome and introductions
Guiding principles for development of the Information Management System and tools for accessing and maintaining information elements contained within it
What should constitute the Information Management System?
Which tools should be supported for accessing information elements in the Information Management System?
Remit of the Software Development Group
Report to the Steering Group
Any other business
Date of next meeting
Appendix

Welcome and introductions

Everyone was welcomed to the meeting. Monica explained that the purpose of the meeting was to take a fresh view on what should constitute the future Cochrane Information Management System and agree on recommendations on how this might be achieved. An immediate outcome of the meeting would be a brief report outlining these recommendations to the Steering Group and Fenella Rouse (who is writing the Collaboration's business plan). Following this meeting Monica would draft this report and circulate it for approval before submitting it on behalf of the Group.
Action: Monica

Jon Godwin participated at the request of Fenella Rouse. Fenella has commissioned him to provide independent advice with regard to the role of the Cochrane Collaboration Information Management System.

Potential conflict of interest

Potential conflict of interest was unchanged.

• Chris Cates: convenor of the newly formed Criticism Management Group, which will also look at non-software aspects of the comments and criticisms system.
• Mike Clarke: as editor of the Reviewers' Handbook
• Monica Fischer: involvement in the development of RevMan; funding of software development at the NCC dependent on decisions made by the SDG
• Julian Higgins: co-editor of the statistical section of the Cochrane Reviewers' Handbook and a writer of S-Plus code for meta-analyses
• Peter Langhorne: none known
• Mark Starr: involvement in development of commercial software, e.g. Metaxis; as publisher of the Cochrane Library
• Phil Wiffen: none known

Guiding principles for development of the Information Management System and tools for accessing and maintaining information elements contained within it

It was agreed to recapitulate the current principles to help keep these in mind for the remaining discussions. It was also agreed that throughout the meeting, if inspired, members could suggest additional principles.

At the previous meeting, it had been agreed that the primary principles should be:

1a: To facilitate production of Cochrane reviews by optimising use of resources and supporting reviewers.
1b: To ensure the scientific quality of the resulting reviews.

2: The following generic question should be applied to all aspects of the Information Management System "Will this help to get Cochrane reviews done?".

Monica suggested that a third principle be adopted: that duplicate storage of information should be avoided where possible. It was agreed that this was an important aspect, but that it had more to do with principles for implementing changes to the Information Management System. Therefore, probably two sets of principles should be maintained, where the first set should relate to the primary objectives of the Information Management System (such as 1 and 2) and the second set should allude to secondary issues such as aspects important for implementation.

Chris suggested, and this was supported, that the 1a and 1b were merged into one to put more emphasis on the scientific quality:

'To facilitate production of high quality reviews by optimising use of resources and updating reviewers.'

Following the meeting, in discussing the draft report to the CCSG and Fenella Rouse, it was agreed to add a footnote to principle 1 giving the definition of a Cochrane review: 'A Cochrane review is a systematic, up-to-date summary of reliable evidence of the benefits and risks of healthcare. It must adhere to a structured format that is described in the Reviewers' Handbook and be in the Parent Database maintained by the Cochrane Collaboration (from whence it will be published).'

What should constitute the Information Management System?

At the moment, there is no separation between the review standard and the software package, i.e., in order to prepare a Cochrane review, the reviewers must use RevMan. In the future, several choices could be offered to the reviewers providing greater flexibility about which software the reviewers can use whilst still ensuring that accepted reviews fulfil the Collaboration's standard. Checks can be put in place for various elements of reviews but ensuring quality of these reviews, however, need to be the responsibility of editorial teams. Today, overall responsibility of the quality of a review lies with the co-ordinating editor.

In addition to having the review production as the key objective of the future Information Management System, it was also agreed that investigations should go into how the review tracking process at editorial bases possibly could be better supported by software, commercial or in-house, and how internal communication could to be improved.

The SDG consider the following elements (blocks of data) part of a Cochrane review:

Elements of a review which may be used independently or elsewhere:
Search Strategy
Review - Synopses, Abstract, Analysis, Title (use independently but not in another review)
Study references
Bibliographic (other) references
Study characteristics
Study results
Contact details
Study registers
Modules (of all entities)
Topic list
Sources of support
Glossary

Additionally, the Collaboration needs to manage:
Documents
Finances
Communication
Distribution lists
Contact details
Web sites
Other documents (minutes, reports, etc.)

The assumption had been made, at a previous meeting, that a central repository database of reviews could be web based but not exclusively web based.

It was agreed that there will always be a need for paper copies of reviews (ie to peer reviewers who do not have access to RevMan). This should be taken into consideration when designing the new system. Consideration could also be given to developing a reader/review viewer, which will allow people to view and print a review without the use of RevMan.

An automatic tracking process for records of changes (possibly in a different colour text) to reviews should be taken into account with the new software.

Mike briefly went through the current status for feedback on the paper entitled 'Inclusion of additional graphs in Cochrane reviews'. The majority of comments received agreed with the option that additional graphs should be possible and that software should be provided.

The remaining of this item and items 4, 5, and 6 are minuted in the memorandum prepared for the Steering Group and Fenella Rouse (see the appendix).

Which tools should be supported for accessing information elements in the Information Management System?

See appendix.

Remit of the Software Development Group

See appendix.

Report to the Steering Group

See appendix.

Any other business

1. Update on plan for releasing RevMan 4.1 and MetaView 4.1

   NCC has identified some beta testers for MetaView and is now running it through the final beta testing phase. They are planning to provide a report to Update Software by 6 June and no major problems are expected. If all goes as planned, reviewers should receive RevMan 4.1 on CD-ROM towards end July / beginning of August. NCC recommends that MetaView 4.1 should not be used in the Cochrane Library until issue 1, 2001. Once the current beta test outcome is known, an e-mail will be sent to review groups letting them know about the timescale, but without committing to any firm dates.
   Action: Monica

2. Graphs document

   Julian agreed to look at the document from Mike, on including graphs and give an explanation/definition of forest plots.
   Action: Julian

Date of next meeting

Monica agreed to circulate dates in late September 2000. Post hoc note: the meeting has been scheduled for 3 October in Copenhagen.

Appendix

Memorandum

To: The Cochrane Collaboration Steering Group and Fenella Rouse

From: Software Development Group

Date: 8 June 2000

Subject: Recommendation on the future of the Cochrane Information Management System and the role of the Software Development Group

Background: The Software Development Group (the "Group" below) had a special meeting on 30 May 2000 at Heathrow airport to discuss the future of the Information Management System, and to agree on the contents of a related report to the Steering Group and Fenella Rouse. This report was drafted, circulated and approved by all members of the Group following that meeting. The Group is willing to provide additional information or expand on the contents of this report upon request.

Principles for the Information Management System

The Group agreed that the following principles should guide the development of the Information Management System:

• The primary objective should be to facilitate production of high quality Cochrane reviews by optimising use of resources and supporting reviewers.
• The following generic question should be applied to all aspects of the Information Management System "Will this help to get Cochrane reviews done?".

Building blocks and data elements in the future Information Management System

The current Information Management System (IMS) covers data elements stored in Review Manager and Module Manager. The Group recommends that the IMS is expanded to cover additional data elements (see figure). This will allow a more efficient integration of the data elements relevant to the preparation and maintenance of Cochrane reviews, and facilitate the automatic sharing of data between data elements (rather than relying on multiple data entry and storage as at present). For example, it should be possible to easily transfer study references from a central trial register to a review.

The figure groups the data elements into building blocks and provides a crude view of how these would interact in the future IMS. It is not intended to be a precise indication of how the elements will be integrated. However, one vision of how this integration might work is given in the appendix. The most appropriate way forward will be determined following the needs assessment that will take place later this year.

Tools for accessing the Information Management System

The Group envisages that there will continue to be a need for a suite of tools to help reviewers and editorial bases to support the preparation and maintenance of reviews (today's suite of tools consists of RevMan and ModMan). The Collaboration should move away from requiring reviewers and editorial bases to use mandatory software (like RevMan). This transition should be achieved by standardising and publishing the structure of the data elements in the Information Management System, and by allowing users of the system to add data to the building blocks providing that these data comply with the standard structure. Cochrane reviewers should continue to be able to prepare and maintain their reviews using tools that are of no direct cost to them. The planned needs assessment will help to determine which tools, if any, the Collaboration should offer to users of the future Information Management System.

Other procedures that should be supported by the future Information Management System include document tracking (e.g. from review title to published review) and internal communication.

Since the proposed new Information Management System (and any supporting tools) will not be available in the near future, the Software Development Group intends to maintain and support the current tools (RevMan and ModMan) for several years to come.

Remit of the Group

In light of the proposed expansion of the Information Management System, the Group wishes to expand its the remit accordingly. The new remit is proposed, but may not be limited, to the following:

• to specify the Information Management System and its building blocks
• to oversee the development of some of the Information Management System's building blocks (such as Reviews, Entities and Contact Details) and coordinate the integration of all of the System's building blocks.
• to oversee the development of any in-house tools, which should be available to users of the Information Management System. Review Manager and Module Manager should continue to be the responsibility of the Software Development Group until there is no longer a need for continued support of these tools.
• to oversee the development of a contact management system (including any in-house tools which may be required for accessing it).

To support the development of a new contact management system, the Group wishes to establish a Contact Management Advisory Group.

To better reflect its remit, it is proposed that the Software Development Group changes its name to: The Cochrane Information Management System Group

The Group proposes that the membership is expanded to include representation from the Cochrane CENTRAL/CCTR Advisory Group, the Cochrane Informatics Group, the convenor of the proposed Contact Management Advisory Group, and any other group, which plays a central role in the definition of building blocks in the Information Management System or the maintenance of key data elements. It may also be necessary to co-opt people possessing technical expertise from time to time. Because of the Information Management System's vital role in the Cochrane Collaboration, the Group wishes to ensure that either the Chair or the Deputy Chair of the Steering Group will continue to be closely associated with Group. In expanding its membership, the Group would also like to ensure a better international representation. Furthermore, the Group agreed that if the development phase of the new Information Management System is to proceed in a timely fashion, it will be necessary to meet more frequently than twice a year. Three to four meetings a year will be more productive.

Future funding requirements

At the moment, it is difficult to estimate the necessary funding required for developing the new Information Management System. We need more information on the needs of users (the planned survey will help to provide this) and to explore the possibility of using existing systems in the new Information Management System. Similarly, there may be existing tools, developed outside the Collaboration, which could support the preparation and maintenance of Cochrane reviews in the new system. The Group will investigate these issues more fully when the results of the planned survey are available. We recommend that the Steering Group should keep the Information Management System high on its list of funding priorities and should discuss this in detail when the Group is able to provide fuller advice on the funding required to develop the future Information Management System.

Funding is however necessary for supporting regular meetings to allow the Group to plan and oversee the development of the new system, to oversee the development of any new tools which may prove necessary, and to oversee the maintenance of existing tools. The Group would like the Steering Group to consider supporting these important meetings, so that they need not rely on ad hoc funding (contributed by Group members) or continue to be restricted to a predominantly UK membership.

Appendix: One possible implementation of the future Information Management System

The most important building block is Reviews. Certain data elements in Reviews may be copied from other building blocks (e.g. study references). As there is a continuing need to work with reviews outside of a central system (such as on a standalone computer or with a paper copy), it will likely be necessary to store all data elements relevant to a review within the review itself. In addition to storing such data elements within the review, there should probably be links to the building blocks in which the data elements are stored initially. Ideally, when a data element is updated in one building block, its updating in other building blocks should be automatic. Although it was agreed that a Cochrane review should exist as a single document only, no recommendation has been agreed on whether all reviews should be stored in a central database or if, for example, they should continue to be stored at review group editorial bases (as part of modules). A survey will be carried out during the remainder of this year to help to determine needs and preferences of users of the Information Management System in regard to this and other aspects.

Reviews is the most important building block because of its content, but the key building block for managing access to the Information Management System could be Contact details. Ideally, contact details would be stored centrally either on a single server or distributed on several servers. There should be only one record for each person and not, as for many people now, one record per entity of which he/she is a member. Only the designated person of a registered entity would have rights to create new records. Read and write access to other building blocks would be controlled through Contact details and be administrated by the designated person(s) (e.g, a review group coordinator could grant a reviewer access to the Reviews building block). Individual members of the Collaboration would be in charge of keeping their own contact details up to date, or designate this task to his/her primary entity.

Ideally, the Study register would be stored centrally on a server (or distributed on several servers). There should exist only one record per study (the register should be study-based and not reference-based). The Study register may be maintained external to the Collaboration, but it should be easy and free of charge for members of the Collaboration to contribute to and extract data from it. Study characteristics (e.g. types of participant) should be stored within Cochrane reviews and should be linked to from the Study register.

The general characteristic for the remaining building blocks is that they should be easily accessible to reviewers, people providing support to reviewers and people maintaining data elements other than those contained in Reviews. As an example, the building block Other documents would contain the Cochrane Reviewer's Handbook. When this document is updated, it should automatically be available to all users of the Information Management System.

IMSG minutes | MinMan home