Endorsement of the CONSORT statement by high impact medical journals: survey of instructions for authors

Introduction

The CONSORT (Consolidated Standards of Reporting Trials) statement of 1996, updated in 2001,1 2 gives recommendations for reporting randomised controlled trials and has been endorsed by the World Association of Medical Editors, the International Committee of Medical Journal Editors (ICMJE), and the Council of Science Editors. Studies indicate that it has helped to improve the quality of reporting of trials.3^–5 I sought to determine the extent to which leading medical journals had incorporated the CONSORT recommendations into their instructions for authors.

Methods and results

Using citation impact factors for 2001, I identified the top five journals from each of 33 medical specialties and the top 15 journals for general and internal medicine. I excluded selected journals that did not publish clinical research (based on explicit statement, inspection of journal contents, or PubMed search) and replaced them by the next on the list. The final sample of 167 journals was obtained after examining 232 journals. Thirteen journals represented two specialties.

Between January and May 2003 I examined the instructions for authors on each journal's website and extracted all text mentioning CONSORT or other publications relevant to randomised trials. I also sought any mention of the ICMJE's Uniform Requirements for Manuscripts Submitted to Biomedical Journals.

CONSORT was mentioned in the instructions of 36 (22%) journals (see bmj.com), more often in general and internal medicine journals (8/15; 53%) than in specialty journals (28/152; 18%). However, 9/36 journals referred only to the obsolete 1996 statement, whereas the other 27 journals (16% of the sample) referred to the latest version, gave the web address (http://www.consort-statement.org/), or both (table). No journal cited alternative reporting recommendations for randomised controlled trials.

Of the 167 journals, 72 (43%) referred to the ICMJE guidelines. Another incorporated much of the ICMJE text without attribution. Eleven of these journals cited the ICMJE guidelines only for particular issues, mostly reference style or authorship.

Only 24/72 journals gave the address; 4 others referred to versions on the websites of the CMAJ or the Lancet. Most of the remaining 44 journals cited an obsolete journal publication: one from 1999, 30 from 1997, five from 1991, two from 1988, and two from 1982. Four journals gave no reference. Journals that referred to CONSORT were much more likely to refer to the ICMJE guidelines (26/36; 72%) than those journals that did not (46/131; 35%).

Comment

In 2003, 36/167 (22%) of high impact medical journals referred to CONSORT in their advice to authors. The uptake of CONSORT by leading journals is encouraging, but 11/36 referred only to a superseded version of CONSORT. Also, many used ambiguous language regarding what was expected from authors. Similar problems were seen for the ICMJE guidelines.

This study reviewed electronic resources, which are volatile. One journal (Annals of Emergency Medicine) updated its guidelines during the study period.

The CONSORT statement was developed to help improve the quality of reports of randomised controlled trials. Its effectiveness will be influenced by the nature of its implementation; ambiguous statements are likely to be less effective than stronger directions. Journals supporting CONSORT should state unambiguously what they expect from authors.

In 2003, many journals gave out of date citations for both CONSORT and the ICMJE guidelines. This carelessness sets a poor example for authors. Journals should be more vigilant regarding the information in their instructions to authors, should be explicit in their expectations of adherence to specific recommendations, and should cite the web address to ensure that the latest versions are obtained along with any extensions.

Illustrative examples from journals' instructions to authors are on bmj.com