Widening the tightrope: new abstract guidelines for BJD authors

Abstract

Scientific abstracts are a tightrope. Authors must give enough information to inform reviewers, editors and readers within the constraints of a very short format. Give too little information, and readers will not understand what the investigators did. Give too much information, and risk exceeding the journal’s abstract word limit. Spend too many words elaborating the methods, and leave too little room to explain the findings. Striking the right balance is more difficult than it may seem, but it is incredibly important. Many readers, including clinicians and patients, will not read beyond the abstract either owing to time constraints or paywalls limiting access to fulltext manuscripts. Titles and abstracts are crucial to screening processes for systematic reviews. Abstract reporting is particularly important for randomized clinical trials (RCTs) because these may directly influence clinical decision making, but abstracts are also a key component of all original articles. The Consolidated Standards of Reporting Trials (CONSORT) extension for journal and conference abstracts was published in 2008 to promote improved reporting of RCT abstracts. The extension includes a 17-item checklist of items that should be included in the abstracts of RCT manuscripts (Table 1). The main CONSORT statement and checklist refers to the extension for abstracts, but does not itself give specific guidance for abstract structure or content. While many journals have endorsed the main CONSORT statement as a requirement for trial submissions, most do not mention the extension for abstracts. A review of the websites of the top 10 dermatology journals by impact factor found that none endorsed the abstract extension, even though nine had endorsed the main CONSORT statement. In last month’s issue of the BJD, McPhie et al. evaluated 198 abstracts of RCTs published over 5 years in the top 10 dermatology journals by impact factor. The mean proportion of essential items reported in the abstracts was only 42%, with particularly low reporting for items related to methods of randomization and funding source. The dermatology literature is not unique in this deficiency; similar results have been found for high-impact general medical journals. In order to identify trial and journal characteristics associated with better reporting, McPhie et al. calculated multivariable-adjusted odds ratios (ORs) for various journal and trial characteristics. Higher journal impact factor and having a registered trial protocol were each associated with better reporting, which are two factors that may indicate generally higher trial quality. However, the most significant association was with abstract word count – having an abstract word limit greater than 250 words was associated with 14 times the odds of better abstract reporting (OR 14 36, 95% confidence interval 6 76–30 52). In response to these findings, the BJD is making the following two changes to its guidelines for authors: (i) requiring all reports of RCTs to include a CONSORT extension for abstracts checklist and (ii) expanding the allowable word count for all original articles (not just clinical trials) to 350 words. While the BJD previously allowed abstract word counts of 250 words, which is considered the minimum for adequate Table 1 Items included in the Consolidated Standards of Reporting Trials extension for reporting randomized controlled trials in journal and conference abstracts