Records relating to published information
Back copies of published issues. A complete set of the bulletin's past issues should be available in the editorial office. The issues should be stored in chronological order and easily accessible to members of the editorial team. It is often necessary to check what was said in the bulletin about a particular subject. The copies will be useful when deciding what topics to write about in future issues: it is helpful to look back over the years to see which articles need to be updated. Past issues can also help solve style problems when there is no house style rule, e.g. for an unusual table or graph. The problem may have been dealt with in a previous issue.
Documents relating to published articles. Records need to be kept about the articles produced by the bulletin, so that the development of an article can be traced from the start if necessary, e.g. if an error is discovered, or the article is challenged in some way. Including all the relevant documents makes it easy to understand how a subject originally came up and what evidence was available at that time, or enable the retracing of steps in the event of being challenged, or the discovery of where an error was introduced.
The archive must be organized in a way that makes retrieving information easy. It is usual to arrange archives in chronological order, the oldest at the bottom, and the most recent at the top, so that when one looks at a record, the most recent item appears first. Care should be taken not to have different archiving orders as this may undermine the structure of the record. The same order should be used for all the articles.
You should decide upon an order for sub-sections within an archive (alphabetical, chronological, thematic, geographical), moving from the general to the particular. This will make it more searchable and easy to retrieve information.
For a bulletin that publishes many articles per issue, and has existed for many years, the volume of documents used in the preparation of articles is so large that they cannot all be kept indefinitely. For example, for la revue Prescrire, the complete documentation for articles is kept for three months after publication (the average time for companies and their opinion leaders to complain about an article). After this, only the references quoted in the article are kept (they remain available for readers, on request, at any time) and the rare documents which cannot be easily obtained again (e.g. unpublished material not available on any web site, clinical reports from companies).
Consider whether it would be worth keeping a microfiche or digital copy of paper documents - modern paper is acid and can destroy itself. Modern inks made from synthetic substances offer no guarantee of permanent stability. See Box 14.1 for some practical steps.
Box 14.1 Archiving article documents
• Do not use rubber bands (they dry out and break), pins or paper clips (they rust and may damage the document).
• Use folders or document bags, not binders or hanging folders.
• Use standardised boxes if possible.
• Write a label on each box. Give each box a reference number so that the files can be identified and easily found.
• Use folders and sub-folders to separate the final issue of a bulletin from the records of its development (drafts of articles, correspondence, reviewers' comments).
• Remove all duplicates and blanks.
• Decide what to keep, e.g. only cited references, or all references, including uncited background papers?
• Ensure that no important element is missing, and print out important e-mail messages.
• Check that each document is identifiable, that is to say that the three questions who? when? and where? can be answered.
• Date photographs and label press cuttings to show where they are from.
• Identify the author of manuscript notes, reports, etc.
• Check the order of the records.
• Check the title and contents.
• Complete the file title with the date on which it was closed.
• Make a list of your contents in each box, and keep the lists carefully up to date so that they are accessible for research.
A record system is only effective if it is directly relevant to the bulletin’s needs. Disposing of records which are not needed is as important as creating them properly in the first place. To begin with, destroy only papers that are of no use: duplicates, rough notes and the manuscripts of published documents (unless they differ markedly from the final version).
Case study: Farmaka i fokus, Sweden
Farmaka i focus published an article about donepezil, and received a letter from the pharmaceutical company Pfizer: they did not consider the article well balanced. Thanks to its record-keeping systems the editorial board of the bulletin could easily retrace how the article had developed before publication and quickly respond to the company with robust reasons explaining how the conclusion of the article had been reached. This would have taken much more work if the records had not been kept, since the published article is often based on a much longer and more extensive draft.
Contributed by Malena Jirlow, Farmaka i focus, Sweden [http://www.janusinfo.org].