Article Text

Download PDFPDF

Uniform structured formats for scientific communications—how far should we go?
  1. Rheumatology Unit, City Hospital, Nottingham NG5 1PB, UK

Statistics from

Clear, efficient communication is a central aim of any scientific report. Clarity, however, can readily be lost when, as authors, we employ our individual literary style, omit detail that we but not the reader take “as read”, present information in long sections without subheadings, and expand reports with comment that relates more to the general topic than the specifics of the study. Following peer review a common request to authors from editors is firstly, to include more detail in the methods and results sections, and secondly, to remove extraneous information and extrapolation from the discussion. Peer review and revision, however, do not always result in optimal presentation of information. There is often disparity between what a study should report and what is actually published. In the case of randomised controlled trials (RCT) this presents important problems for inclusion in systematic reviews1 and the balanced appraisal of knowledge that may determine clinical practice.

The continuing education of investigators is clearly important if we are to maintain high quality research and communication. One—often underestimated—aspect of this is the experience of peer review. Although primarily established to guide selection for publication, peer review serves an important educational role in directing authors to potential problems and caveats of their study and in suggesting ways of improving presentation and discussion. It is for this reason that theAnnals editorial office always sends appropriately submitted reports to peer review, even if the Editor’s initial appraisal is that the report is sufficiently flawed that it would not be accepted even after major revision. The authors thus receive more than just a rejection letter from Annals. They obtain a critique of their work and an explanation as to why it was considered unsuitable for publication. Such practice, however, is not always undertaken, especially by larger and busier general medical journals.

CONSORT flow diagram for randomised controlled trial.

An effective means to tailor reports to better fit their purpose is to impose a generic structure on their presentation. The ordered division of a report into headed sections (for example, Summary, Introduction, Results, Methods, Discussion, Acknowledgements, References) facilitates the reader’s search for information, but importantly encourages the author to supply information under each heading. In line with many other journals, Annals in 1993 introduced a structured abstract in place of an unstructured summary. This simple manoeuvre means that authors now have to specify the objectives, methods, results, and conclusions of their study, aspects of which are often unclear or omitted from a “summary”. With widespread use of computerised literature searches based on abstracts such a descriptive precis is clearly beneficial. From 1997 we have also adopted the CONSORT (Consolidation of Standards for Reporting Trials) format for reporting RCT,2 requiring authors to supply a checklist of 21 items (table 1) and a flow diagram showing the progress of subjects through the various stages of the study. The example shown (figure) relates to the most commonly reported two group parallel design RCT.3 This strategy to enhance clarity and facilitate the appraisal of RCT will be formally assessed over future years.2

Table 1

CONSORT checklist for randomised controlled trials

Other attempts at standardisation have been supported by a body of international rheumatology journal editors who met first as a group in San Francisco in 1995. An audit by the group in 1996 revealed marked disparity in the referencing format used by rheumatology journals.4 The majority of these editors have now agreed to use a single style—the current, fourth edition of the Vancouver system5 6 —from January 1997. The following example shows how an original report is referenced in this format:  1 B rown A, Jones B, Smith C, Zahl D. Salicylate intake and subsequent risk of developing knee osteoarthritis. Ann Rheum Dis 1999;58:1984–6.

This agreement by rheumatology journal editors provides a rare example of successful closure of an audit loop in publishing. The group is also developing a common set of rheumatological acronyms and abbreviations and hopes to introduce this by the end of 1997.

Some, of course, may find uniform requirements and structured formats difficult and even restrictive. Such concerns were raised when the need for more informative structured abstracts was first debated.7 In general, however, the benefits of uniform structuring are likely to outweigh the disadvantages expressed in terms of author “freedom” and the difficulty of forcing diverse studies into a common mould. Indeed we might consider going further. For example, it is extremely common for reviewers and editors to request modifications to the discussion section of papers. Subdivision of this section (table 2) may have advantages in reducing overall length and undue extrapolation (a common problem), achieving a better balance of declared strengths and weaknesses of studies (caveats are often underemphasised or omitted by authors), and placing findings in the context of other relevant published work. Other headings could be added, depending on what the function of this section is perceived to be. The views of authors and readers of the Annals on this matter would be welcomed.

Table 2

A suggested “structured discussion”


Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.