When I use a word . . . The ICMJE requirements and recommendations

The Vancouver Declaration

In January 1978, the late Edward Huth, then editor of the Annals of Internal Medicine,1 and Stephen Lock, editor of the British Medical Journal, as it was known at that time, now The BMJ, proposed organising a meeting of editors of influential medical journals to discuss a standard format for citing references. Huth and Case have given a detailed account of the history of the event.2 In all, 19 journals were represented at the meeting, and the committee of 11 included editors from the Annals of Internal Medicine, The BMJ, The Journal of the American Medical Association, The Lancet, The New England Journal of Medicine, and other leading biomedical journals.

The report that emerged, the Declaration of Vancouver,3 named after the city in which the meeting had been held, was published in the British Medical Journal in May 19784 and elsewhere later.56 It included a description of the editors’ suggested method of citation, which has come to be known as the Vancouver method. However, the report went further than that. The editors had not restricted themselves to discussing methods of citation, but had encompassed other matters related to how to prepare and submit a manuscript, including details about the title page, the abstract and keywords, the text, the acknowledgments, how to format tables and illustrations and their legends, and what abbreviations were acceptable. Even the type of paper to be used was prescribed: “20.3 × 26.7 cm or 21.6 × 27.9 cm (8 × 10½ in or 8½ × 11 in) or ISO A4 (212 × 297 mm) white bond paper with margins of at least 2.5 cm (1 in).”

The report was generally warmly received, and despite some reservations about the supposed advantages of the new citation method over other methods, such as the Harvard system,7 the new format was widely adopted.

The original publication of the report in the British Medical Journal was headed “For Debate,” but when a slightly revised version was published in February 1979, soon after another meeting of the committee, this time in Montreal, the heading became “Style Matters”; the whole system had been widely accepted.8 The only major change to the original version involved tables of abbreviations; these had previously covered (i) standard units of measurements and statistical terms, (ii) so-called “combining factors,” from tera- (10¹²) down to atto- (10⁻¹⁸), and (iii) other common abbreviations of a range of biochemical and medical terms. To these the revised edition added a list of approved abbreviations of the names of 108 journals, from Acta Med Scand to Surg Gynecol Obstet.

In 1979 the revised version was published in several journals.910111213 And by December 1979, according to a list published in The BMJ,14 88 journals had signed up to the requirements, from American Family Physician to the Yale Journal of Biology and Medicine, with an additional entry for “World Health Organisation [sic] journals.”

Before long, the Vancouver Group changed its name to the International Committee of Medical Journal Editors (ICMJE), and over the years expanded its interests to include publication ethics and various other matters, such as authorship, conflicts of interest, editorial freedom, plagiarism, duplicate publication, fraud, and retractions.

Regular revisions of the document resulted in major changes in 1997, 2003, and 2010. Then, in 2013 the committee renamed the document, changing the word “requirements” to “recommendations”; the new document bore the overall title, as the latest version still does, “ICMJE Recommendations for the Conduct, Reporting, Editing and Publication of Scholarly Work in Medical Journals,” or simply “The ICMJE Recommendations.”15

The ICMJE recommendations

The 2013 version of the recommendations fixed the ICMJE’s criteria for authorship of a paper by the addition of a fourth item to the previous three:

1. Substantial contributions to: the conception or design of the work; or the acquisition, analysis, or interpretation of data for the work;

2. Drafting the work or revising it critically for important intellectual content;

3. Final approval of the version to be published;

4. Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Any would-be co-author had to fulfil all four criteria to be eligible, and the final criterion specified accountability as an important component. Without accountability, a co-author could claim that any scientific misconduct that was subsequently discovered was not their responsibility, since it had involved some aspect of the study or the publication in which they had not taken part. It also implied that even if there was an aspect of the work in which they were not expert or which they perhaps could not fully understand, they had sufficient confidence in their co-authors to accept responsibility for what those co-authors had done or written.

The new recommendations also extended the scope for declaring potential conflicts of interest, encouraged authors to pre-publish protocols describing proposed study designs and to make available data obtained during an interventional trial, and encouraged the publication of the results of a trial, whether the outcome was positive, neutral (i.e. lacking statistical significance in either direction), or negative. They also recommended that editorial decisions should be based on “the relevance of a manuscript and its originality, quality, and contribution to evidence about important questions, and not on commercial interests, personal relationships or agendas ….”

Since then the recommendations have been regularly reviewed and updated.

Comments

The ICMJE’s recommendations have proved highly influential in biomedical publishing and have played an important part in encouraging publication of protocols in advance of studies; in the establishment of trial registries and compliance with requirements for trial registration; in publicising potential conflicts of interest; in defining aspects of publications that raise concerns and may be considered for retraction; and in defining the nature of a peer reviewed journal. All these and other matters were discussed by Ed Huth and Kathleen Case in 2004 after the publication of the then most recent version of the requirements.2

The introduction to the August 2013 New ICMJE Recommendations stressed that the establishment of a standardised approach in the preparation of manuscripts would be of help to authors, but it is not clear to what extent that hope was fulfilled. It is in fact much easier for authors not to have to follow strict guidelines about the preparation of a manuscript, except perhaps that when a paper is rejected by one journal it is automatically ready for submission to another. Even then, if uniform standards were not applicable authors would not have to revise a manuscript each time they submitted it to a different journal. Indeed, there is a case for allowing authors to submit a paper in any format they think suitable and to ask them to format it in the journal’s style only after it has been accepted. The advent of preprint servers has shown that publication in an author-driven format is acceptable to the scientific community.16

On the other hand, uniform recommendations are much more likely to be of help to editors of journals, copy editors, and reviewers. Editors, for example, will prefer that papers submitted to their journals should be presented in uniform fashion, giving the journal an identity that is recognisable by reviewers and helps to enhance its reputation. Copy editors, in turn, will not have to go to great lengths to ensure that all manuscripts are properly prepared for publication; the advent of computerised submission systems has greatly contributed to their policing activities, since the computer does not allow an author to submit a manuscript unless all the requirements that the journal lays down have been fulfilled; after submission, authors may well be frustrated by repeated requests from copy editors to attend to the most minor aspects of presentation that are not covered by the computer. Regular reviewers, to whom editors look for reliable judgments, may also prefer to be presented with material in a recognisable format, allowing them to formulate their judgments more readily.

The uniform requirement/recommendations may also have encouraged journals to focus more on process and style rather than content. The formatting of articles of specific types is defined not, or not wholly, by subject matter, but by length, number of references, and limitations on the number of tables or figures, even though the topic may demand a different approach. The journal may not have another standard format under which the topic may be appropriately covered, even though the subject matter is relevant to its interests.

Some studies and a final thought

A specific example also shows how the Vancouver method of citation may have affected co-authorship. In an analysis of articles from random volumes of eight biomedical journals, Epstein found that most of the journals showed a trend towards increasing authorship numbers during the study period of 1982–92.17 General medical journals that published articles with a median of six to seven authors per publication published far fewer studies with seven authors than six. Since the Vancouver system restricted citation of more than six authors at a time, he suggested that senior authors, whose name typically comes last in the list, might prefer six-author papers to seven-author ones.

An internet-based survey of 1226 first authors of Cochrane reviews published between October 2016 and December 2018, of whom 666 (54%) replied, showed that the prevalence of gift authorship was 41%; this was only partly due to ignorance of the ICMJE’s authorship criteria.18

In a study of 41 Cochrane reviews with group authorships (median 32, range 6–91), 19 of which were published from 2015 to 2019, only 39% met ICMJE’s first authorship criterion, 41% the second, and 12% the third; in only two studies did all the authors meet all three criteria.19

It would be of interest to see a systematic review of the evidence that reflects the extent to which the requirements/recommendations laid down in the ICMJE documents have been effective in changing publication practices in biomedical journals.