(for authors) Guidelines for writing abstracts

The article title (see What are JMIR's guidelines for article titles?) and abstract are the most read and most important parts of a manuscript.

A well-written and detailed abstract assists in the peer review process and also, if the manuscript is accepted and published, typically is a major deciding factor for readers/researchers to decide whether to read or cite an article. We strongly recommend that authors submitting their work for peer-reviewed publication spend a careful and appropriate amount of time, attention, and effort to craft a well-written abstract. This applies to any abstract, whether structured or unstructured. 

Which article types require an abstract?

• All article types, except the ones listed in the table below as "no abstract or none required", must include an abstract. (What are the article types for JMIR journals?)

* Editorials should also have an unstructured abstract, even if only 2-3 sentences, to convey the key content and points of the Editorial. A similar approach should be used for Viewpoints.

** Structured abstracts, unless otherwise noted, include: Background - Objectives - Methods - Results - Conclusions. For example, Implementation Reports have a minor variation on the structured abstract: Background - Objectives - Methods - Implementation (Results) - Conclusions. 

• For more information on the abstract requirements for specific article types, please see: What are the article types for JMIR journals?

What do I include in a structured abstract?

• Content-wise, an abstract should be robust and detailed enough to stand on its own. Details in each section of the abstract should reflect the major components of the study design and findings from it.
  
  • Some article types have recommended abstract checklists and authors should adhere to these when writing the abstract. For example, a PRISMA for abstracts checklist is available for systematic literature reviews. 
  • Abbreviations and terms introduced in the abstract must be reintroduced in the main text.
• The Results section should include appropriate quantitative data to the abstract, such as sample size, and the means (or proportions) of outcome measure which you compare between groups (i.e. do not only throw out P-values without citing the means or percentages of the groups you are comparing. Also, make sure to include df where needed, and test statistic eg t-score, chi-square, F-score). Please, see: How should P values be reported?
  
  • For JMIR Research Protocols manuscripts, a very short Results section is still required. This outlines the current status of the project (eg, if the study is recruiting participants or underway, and the anticipated completion date) (What should I put into the results section of a protocol or proposal?).
• Also, do not only mention the "positive" findings. Also explicitly state if the primary outcome was "negative" (eg no statistically significant differences found between groups).
• Abstracts must not include:
  • Hyperlinks (one consideration is for abstract Trial Registration information, only provide the URL and remove hyperlink);
  • References or citation numbers to references. Authors would be asked to remove these either during editorial processing or production of the manuscript, often requiring renumbering of all references.
• Randomized controlled trials (RCTs) require a trial registration number in the abstract, as recommended by the International Committee of Medical Journal Editors (ICMJE) (Does my trial (RCT) have to be registered?). See the section below on trial registration for more information (Trial Registration Numbers in abstracts).

Example structure (also read any recently published manuscripts with structured abstracts)

What do I include in an unstructured abstract?

Usually, at least 2-3 sentences (max. 450 words) are included to describe the key issue and supporting points or themes of the manuscript. 

See any recently published manuscripts with unstructured abstracts.

Trial Registration Numbers in abstracts

If a paper is a Randomized Controlled Trial (and referred to as such in the title), then it must have a Trial Registration number in the abstract and also in the manuscript body. A protocol for an RCT may also include a trial registration number if it has been registered. 

How to format the trial registrations in the abstract

International Standard Randomized Controlled Trial Number (ISRCTN) 14881571; http://www.controlled-trials.com/ISRCTN14881571/14881571

or

ClinicalTrials.gov NCT00102401; http://clinicaltrials.gov/ct2/show/NCT00102401

Note: Please remove hyperlinks from URL included in the trial registration.

Related articles

• What are JMIR's guidelines for article titles?
• What are the article types for JMIR journals?
• Why were abbreviations expanded in my abstract during copyediting/proofreading?
• JMIR House Style and Editorial Guidelines
• How do I edit the abstract/title in the metadata form on revision or before publication?
• How do I cite my trial registration?
• Does my trial (RCT) have to be registered?
• What reporting guidelines should I follow for my article?
• What should I put into the results section of a protocol or proposal?