Writing an Abstract for Publication
by Justin B. Moore, PhD, MS, FACSM
The Scholarship of Public Health addresses topics relevant to scientific publishing, dissemination of evidence and best practices, and the education of current and future professionals. This column presents some considerations and best practices when producing an abstract for your manuscript or presentation.
A number of years ago when the Journal of Public Health Management & Practice adopted a structured article format for full-length articles, I published an editorial introducing this change in the journal’s format. Naively, I thought that the editorial would be sufficient guidance. In the intervening years I’ve worked on the journal, I’ve fielded numerous questions about the format of several journal submission items (eg, cover letters, abstracts, references) and given feedback to authors on these and other components. In this post, I’ll address the abstract, but please stay tuned for a series of other articles addressing all components of a successful submission.
Seven years ago, I didn’t fully appreciate the difficulty inexperienced authors encounter putting together an abstract. It’s difficult for me to remember now the days when I didn’t have writing experience (ie, I’m getting old). Having written several hundred abstracts for manuscripts and conference presentations, I can tell you that experience is helpful. Much like all writing, practice makes one…if not perfect…less likely to be rejected. (In public health, it’s all about the probabilities.) Part of the reason is that abstract formats of journals and conferences vary immensely both between and within the genre. For example, I usually submit an abstract to two different conferences in September each year. One is a regional conference (the Southeast Chapter of the American College of Sports Medicine) that uses a unique format that requires headings (eg, purpose) and that the abstract fit into a box 5.75 inches wide by 5 inches high with a limit of no more than 25 lines of text. (See a recent call for abstracts here.) In my last submission, this included about 231 words exclusive of names, affiliations, headings, and funding disclosures. This is similar in word count to the format utilized by the American Public Health Association that enforces a 250-word limit. Juxtapose that against another conference to which I regularly submit, the Active Living Research annual conference, whose format allows for 450 words including the headings. Journals also vary widely. Some journals (especially those with brief formats such as JPHMP) allow as few as 150 words or as many as 300 words, which may or may not be inclusive of required headings. So the most important takeaway is that you need to identify the conference or journal prior to writing the abstract and study the instructions for authors before submitting.
Another area of variability is the desire or requirement of the journal for the authors to include specific information or detail in the abstract. For example, the American Journal of Preventive Medicine (and others) requires date of data collection and analysis in the abstract. The type of information required will vary by journal or conference, and much of the variability pertains to the presentation of data. For example, American Medical Association (AMA) and American Psychological Association (APA) formats (the two most commonly used by public health journals) are specific about presentation of “P” values (in AMA format; “p” values in APA format). For example, AMA format, which is followed by JPHMP, requires that statistical significance (ie, P values) should be expressed to two digits to the right of the decimal point unless the P value is less than .01, in which case the P value should be expressed to 3 digits to the right of the decimal point. (See this Editage post for a more thorough discussion.) Further complicating the issue is that some journals discourage presenting point estimates (eg, means) and P values in the abstract while others (eg, JPHMP) prefer their inclusion. So a good practice is to review the most recent issues of a journal (or listing of conference abstracts) before writing the abstract for a manuscript/presentation to ensure consistency with recently published/accepted abstracts.
As with any presentation of results or data, you want to stay within the confines of your evaluation or research design (eg, don’t use causal language for cross-sectional studies). But more importantly, don’t “bury the lead” of your study, and don’t rely on an individual reader’s statistical acumen or content knowledge in interpreting your results. The best abstracts and papers I read both give a measure of effect and the practical significance of that effect in plain language. For example, here’s a bad version of a hypothetical result:
The logistic regression model indicated that television viewing was significantly associated (P < .001) with sedentary time after controlling for physical activity, age, and gender.
Here’s a better version:
The logistic regression model indicated that television viewing (OR = 1.73; 95% CI, 1.43-2.09) was significantly associated with sedentary time in that children who watched >3 hours of TV per day were 73% more likely to spend >5 hours per day sedentary after controlling for physical activity, age, and gender.
This may seem relatively obvious, but I cannot tell you how many times I have read an abstract or an article with really interesting or important findings that were hidden by reporting only a P value or not giving context for the findings (if not obvious or previously placed in context in the background section). The latter can be especially important for clinical values where a small, absolute change can be important for health (eg, a one-unit change in hemoglobin A1c would be meaningful in a diabetes prevention study), or in policy/economic analyses where the population impact can be highlighted (eg, “for every $1 increase in the excise tax rate, the adult smoking rate decreased by 0.81%”). Therefore, a best practice for getting the attention of editors and reviewers is to hit them over the head with the public health importance of the findings. (Note: don’t actually hit us on the head.)
Finally, a piece of obvious advice: when writing a manuscript, write the abstract last. All too often, authors write the abstract first because that’s where it appears in most manuscript drafts. I would avoid doing this, as you’ll often re-do analyses, reframe the results, and shift emphasis in the discussion as the writing process evolves. As such, you may wind up with an abstract that bears only a cursory resemblance to your manuscript if you write the abstract first. With a presentation abstract this can be less of an issue, but you certainly want your analyses to be final before writing the abstract results and conclusions since results and conclusion sections of abstracts written while analyses are being conducted (ie, with a rapidly approaching submission deadline) can be necessarily vague, which can hurt you in the review process.
So in summary, a) identify the outlet for submission prior to beginning the writing process, b) review recently published abstracts to learn the style of the journal/conference, c) highlight the practical importance of your findings, and d) write the abstract last.
Justin B. Moore, PhD, MS is the Associate Editor of the Journal of Public Health Management and Practice and an Associate Professor in the Department of Family & Community Medicine of the Wake Forest School of Medicine at the Wake Forest Baptist Medical Center in Winston-Salem, NC, USA. Follow him at Twitter and Instagram. [Full Bio]