Structured abstracts: which journals require headings, and what they want.
Waraq · July 18, 2026 · RULE OF THE WEEK
Structured abstract headings are mandated mainly by medical and clinical journals. The Lancet requires a structured abstract, and its information for authors spells out five paragraphs: Background, Methods, Findings, Interpretation, and Funding. Frontiers in Neurology also requires structure, but as a single paragraph following the IMRAD sequence rather than under printed labels. Everywhere else the guide for authors decides. Some journals print headings, some forbid them, and almost none let you choose.
Is a structured abstract a different kind of writing?
No. It is the same science under mandatory labels. The journal is regulating the container, not asking for new content.
Authors lose days to this misunderstanding. Told that the target journal wants a structured abstract, they open a blank page and start over, as if the format demanded different sentences. It does not. An unstructured abstract that summarizes a study already moves through background, method, result, and conclusion, because that is the only order in which a study makes sense to a reader. The structured format takes that implicit sequence and turns each stage into a labelled, checkable section.
There is one honest exception. If your unstructured abstract cannot be cut into those sections, the format has told you something: a stage of the argument is missing. An abstract with no sentence that a Conclusion heading could sit above was incomplete before the headings arrived. Readers stumbled on the gap; they just could not name it.
Which journals require structured abstract headings?
The clearest mandates come from clinical and biomedical publishers, and the two rules below show the two shapes the requirement takes.
The Lancet's rule fits in one line: "The Lancet requires a structured abstract." The journal's information for authors then fixes the sections for research articles as five paragraphs, Background, Methods, Findings, Interpretation, and Funding, and caps the summary at 250 words in the current guidance, so read the current PDF for your article type before you count. Note the vocabulary: The Lancet says Findings where most journals say Results, and Interpretation where most say Conclusion. Submitting the generic labels to a journal that names its own is the kind of slip a technical check catches on day one.
Frontiers in Neurology takes the second shape. Its author guidelines say: "The abstract should be no longer than a single paragraph and should be structured, for example, according to the IMRAD format." IMRAD is the standard research sequence of Introduction, Methods, Results, and Discussion. So Frontiers wants the skeleton without the printed labels, one paragraph whose sentences march through the sequence in order. You can draft it with headings for your own discipline and then delete them before submission; the structure survives the deletion.
Outside these mandates, the spread is wide. Nature and PLOS ONE regulate length rather than structure, a difference covered in our comparison of abstract word limits across journals. Structure and length are separate rules, set independently, and your target journal may enforce one, both, or neither.
Re-cutting one unstructured abstract into required sections
The re-cut is a labelling exercise. Here is an unstructured abstract, then the same abstract under headings, with the science untouched.
The unstructured version:
Poor sleep is common in adults with migraine, but its relation to attack frequency is unclear. In a prospective cohort of 214 adults with episodic migraine, we recorded sleep with wrist actigraphy over 12 weeks and collected daily headache diaries. Nights with under six hours of sleep were followed by higher odds of a next-day attack (odds ratio 1.4, 95% confidence interval 1.1 to 1.8), and the association held after adjustment for stress and caffeine intake. Short sleep appears to be a modifiable trigger that migraine care could target.
The structured version:
Background: Poor sleep is common in adults with migraine, but its relation to attack frequency is unclear.
Methods: In a prospective cohort of 214 adults with episodic migraine, we recorded sleep with wrist actigraphy over 12 weeks and collected daily headache diaries.
Results: Nights with under six hours of sleep were followed by higher odds of a next-day attack (odds ratio 1.4, 95% confidence interval 1.1 to 1.8). The association held after adjustment for stress and caffeine intake.
Conclusion: Short sleep appears to be a modifiable trigger that migraine care could target.
Count the changes. One compound sentence split in two so the Results section could stand alone, and four labels added. No number moved, no claim strengthened, no sentence rewritten. That is the whole job, and it took minutes, not a day. For The Lancet you would rename two labels, Findings and Interpretation, and add a Funding line. For Frontiers you would delete the labels and keep the order.
The re-cut gets hard only when the original was disordered, with a conclusion in sentence two or methods scattered across the paragraph. Then the headings force a reorder, which is still cheaper than a rewrite and usually improves the unstructured version too.
How do you budget words per section when the total is fixed?
Give Results the largest share, Methods the next, and keep Background and Conclusion short. The headings do not come with extra words; they partition the words you had.
On a 250-word total, a workable split is 40 words for Background, 70 for Methods, 100 for Results, and 40 for Conclusion. The abstract above spends its budget in roughly those proportions. The split is not a rule any journal states; it reflects what each section is for. Background earns its place with one sentence of gap, not a literature review. Results carries the numbers, and numbers are what readers scan a structured abstract to find. A Conclusion longer than two sentences is usually smuggling in discussion.
Budgets tighten when the limit is low, and the first cuts should be phrasing rather than findings. "In order to" becomes "to," "due to the fact that" becomes "because." Our list of wordy academic phrases and their short versions recovers twenty words from most abstracts without touching a result. And since guides rarely say whether labels are counted, assume they are and leave ten words of slack.
How does a rule-based check catch a missing Conclusion heading?
Mechanically. A structured-abstract requirement decomposes into checkable facts: each required label present, spelled the journal's way, in the journal's order, over the total limit or under it. A checker holding the rule "The Lancet requires a structured abstract" plus the journal's section list can flag an abstract that ends at Interpretation with no Funding line, or one that says Results where the journal says Findings, before any human reads the science.
That is also why journals catch this so reliably themselves. The editorial office's technical screen runs the same kind of list, and a missing heading is among the easiest defects on it, the format-compliance cousin of the substantive problems that get papers desk rejected. When Waraq checks a manuscript against a target journal's rule set, the missing heading comes back as a margin note quoting the journal's own rule, which is the version of this check you can run before the journal does.
The general habit costs nothing. Before submitting, read your abstract against the journal's stated section list, label by label. The science is yours; the headings are theirs.