Shorten an abstract to its word limit
An abstract has to fit a fixed count of words or characters without losing what it reports. Fitting it rests on a budget for its content, the removal of repetition and wordy phrasing, shorter words, and linguistic devices that save words.
The word limit
Biomedical journals usually set an upper limit of 100 to 250 words for an abstract. Typically 200 to 250 words are allowed, and sometimes fewer for a conference abstract. Structured abstracts are typically 250 to 300 words. The limit must not be exceeded: online submission systems do not allow it, and editors otherwise reject a manuscript whose abstract is too long.
An abstract has to be complete, concise and interesting, and strict limits on its length and format make that harder. More journals were requiring structured abstracts, which can add to the required content and whose headings count toward the limit. The CONSORT extension for abstracts insists on content that further raises the count, and marketing and legal departments add demands of their own. With these demands, fitting in everything everybody wants becomes a delicate and complicated balancing act. An abstract cannot be dashed off at the last minute. Keeping within the limit is sometimes difficult. It seems that not every author has this problem: a physics preprint by four authors that asked whether apparent superluminal neutrino speeds can be explained as a quantum weak measurement had a whole abstract that read "Probably not."
Budgeting the content
Generally, most of an abstract should be methods and results. Long background, objective or conclusion sections waste space. A good rule of thumb is about 10 to 20% background, 30% methods, 40% results and 10 to 20% conclusions. In some cases, such as an unstructured abstract, the background can go altogether, because a reader who wants it can read the introduction.
The abstract is drafted backwards from the conclusion: first the key message or messages, then the results that support them, then the methods that produced those results, without detail. Anything that is not essential is eliminated. For example, antibody concentrations were measured on day 28 by enzyme-linked immunosorbent assay is enough without the assay's particulars, and adverse events were recorded up to day 56 needs nothing on relatedness, intensity or laboratory analysis. A sequence of fixation, dehydration and scanning electron microscopy reduces to samples were examined by scanning electron microscopy.
Negative results that can go without changing the conclusions are cut, though some may be essential. Rates of adverse events were similar in the subjects receiving the vaccine and subjects receiving placebo is unnecessary; for an abstract, the important information may be only treatment-related serious adverse events. References are generally not permitted in the abstract of a manuscript and are avoided.
Authors or a marketing department sometimes press the writer to squeeze non-essential messages into the abstract. The writer explains to them that the reader can always find those messages in the main text, and reminds them that the abstract absolutely cannot exceed its limit. This may be different for a conference abstract, where marketing considerations may take precedence and other statements in the abstract must be sacrificed. The CONSORT extension for abstracts is a guideline for what the abstract should include, not a law: the writer does their best to fit in everything on its list, and in the end judges what needs to be included.
Repetition and wordy phrasing
Each word of an abstract should be chosen as if whether the paper is read depended on it, and verbal clutter should be cut ruthlessly.
A detail given in the title, the methods or the objective is not stated again. A title that says the study was randomised, parallel-group and placebo-controlled makes that unnecessary in the methods. Once the methods name the assay, the results say antibody concentrations were… rather than the enzyme-linked immunoassay showed that antibody concentrations were…. The objective need not repeat an age range the methods already give either.
- Cut out, wherever possible, words and expressions that add nothing: previous studies of healthy aging have shown a strong relation between WML and cognitive decline becomes WML strongly correlates with cognitive decline in aging healthy adults.
- Delete meaningless preliminaries such as in this paper, we discuss… and openings announcing that something is well known or that many studies have shown it, so that the sentence just says what is. Write we examined… or this study examined… for the objective of this study was to examine….
- Drop known as in patients with known asthma were included, and give a result without an introduction such as we were able to show that or we showed that: mean systolic blood pressure decreased by 22 ± 11 mm Hg.
- Sentences that begin it is, it was, there are or there were can almost always be rephrased and are best avoided, in abstracts and in any other text, because they almost always end up complex and wordy: there were 10 patients who had severe adverse events in the treated group becomes ten patients in the treated group had severe adverse events.
- Drop show, reveal, demonstrate and found from statements of results, so that showed an increase and were found to increase both become increased, and the analysis revealed an improvement in clinical response becomes the clinical response improved.
- Phrase what was not done, and negative results, without not: patients with diabetes mellitus were excluded rather than not included, and resuscitation failed in 20 patients rather than did not succeed.
- Give a demographic result with only two alternatives, such as men and women or smokers and nonsmokers, for one of them only: 56.3% were women; 27.9% were smokers.
Some unnecessary words are common in abstracts, as in all texts, and each has a replacement.
| Replace | With |
|---|---|
| A bigger/higher/larger or lower/smaller + amount/degree/number | More/less |
| A decreased/increased number | Fewer/more |
| A higher/lower rate | Faster/slower |
| A total of seven | Seven |
| Above the normal range | Elevated |
| All of | All |
| Are/was/were able | Can/could |
| As a result of | Because of |
| At all times | Always, or delete |
| At regular intervals | Regularly |
| At the same time that | While |
| Based on the fact that | Because, based on |
| Both of these | Both |
| Bring about | Cause |
| By means of | Using |
| Classified into groups | Classified |
| Compare and contrast | Compare |
| Connecting words: therefore, thus, however, moreover, consequently, nevertheless, hence | Delete when possible |
| Considered to be | Delete |
| Despite the fact that | Although |
| During the course of | During, while |
| First(ly), second(ly), finally, etc. | Delete |
| For the purpose of | To |
| Found to be | Is |
| Give cause to | Justify |
| Great/large majority | Most/many |
| Had an effect on | Influenced |
| Has been found to be/is found to be | Was/is |
| Has the ability to/is able to | Can |
| In a considerable proportion of | In many |
| In conclusion, in summary | Delete when possible |
| In excess of | Higher than |
| In no case(s) | Never |
| In order to | To |
| In spite of | Despite |
| In the absence of | Without |
| In the range of | About |
| Is an important factor in | Delete |
| Is currently | Is |
| Is seen as | Is |
| Known to | Delete |
| Known to participate in | Participates in |
| Male/female children | Boys/girls |
| Male/female subjects | Men/women |
| Mix together | Mix |
| On the basis of | Based on, because |
| Outside the normal range | Abnormal |
| Past experience/history | Experience/history |
| Plays an important role | Contributes to, or delete |
| Prior to | Before |
| Seem to be | Seem |
| Subsequent to | After |
| The fact that | Because, that, or delete |
| The majority of | Most/many |
| The minority of | Few |
| Time interval | Interval |
| Time period, period of time | Period |
| Time schedule | Schedule |
| Under/over the age of | Younger/older than |
| Using a combination of | By combining |
| Very, extremely | Delete |
| Virtually | Delete |
| Whether or not | Whether |
| Widely/commonly held/believed/thought/considered that | Delete |
| With the exception | Except (for) |
Multiple hedges, also called multiple qualifications, are often used to avoid saying that one is sure of something, and they should be reduced as far as possible. They are often added under pressure from a legal department, or because scientists are often taught to leave the door open to a change in perspective or to evidence that contradicts a conclusion. In one example, these preliminary results suggest the possibility that the drug might be effective at reducing the incidence of the disease in some populations, although further studies are needed to confirm this becomes these results suggest that the drug will reduce the incidence of the disease.
Shorter words
Journals usually set the limit in words, but the instructions for a conference abstract may set it in characters, and then the length of words matters as well as their number. The abstract is written first for conciseness and revised afterwards to substitute shorter words. Wording that at first seems too simple to its writer is no cause for worry, because people are too used to hearing polysyllabic words where one or two syllables are good enough. A writer who feels they have gone too far, oversimplified or lost the meaning checks with a colleague. Word's thesaurus helps in finding the shorter word.
- use instead of utilize
- see instead of encounter
- do instead of perform, execute or conduct
- before instead of prior to
- stop instead of terminate
- halved instead of divided into two equal parts
Linguistic devices that save words
Besides removing unnecessary content and using plain language, linguistic devices reduce the word count, and they apply to conference abstracts as well as to abstracts for publication.
- Parallel construction gets rid of compared to, and it reduces the word count and makes the text easier to read: "12.2 months in the treatment group, compared to 3.1 months in the group that was treated with placebo" becomes "12.2 months in the treatment group and 3.1 in the placebo group".
- The design and setting entries of a structured abstract can be phrases rather than full sentences, as in "Design: double-blind, randomized, parallel-group study". An objective can start with "To" instead of "the objective of the study was to".
- Logical operators replace text: "in adults 60 years of age and older" becomes "in adults ≥60 years of age", and "two or more doses" becomes "≥2 doses". Closing the space between an operator and its number saves further words.
- Numerals replace spelled-out numbers, even at the start of a sentence. That a number must be spelled out there, or that a sentence must never begin with one, is a convention rather than a rule, and many medical writers ignore it. The journal's editor decides whether to enforce it. "A total of 247 subjects were enrolled" can be "247 subjects were enrolled".
- An unnecessary conjunctional "and" can be split into two sentences or, better, collapsed: "No treatment-related serious adverse events were reported, and no deaths occurred" becomes "No treatment-related serious adverse events or deaths were reported".
- An unnecessary relative clause, often starting with "who" or "that", can raise the word count and complicate the sentence, so it goes: "The patients who were treated with aspirin" becomes "The patients treated with aspirin", and "The values that were determined on" becomes "The values determined on".
- A nominalisation is a verb turned into a noun, often ending in -ment, -tion, -sion or -ing, and it always lengthens the sentence and often makes it clumsier. English is a verb-based language, so the meaning and activity of a sentence should usually go into verbs: "Measurement of the protein concentration was made using" becomes "The protein concentration was measured by", and "Administration of the vaccine was made by intramuscular injection" becomes "The vaccine was administered intramuscularly", or better "Subjects were vaccinated intramuscularly".
- A prepositional phrase can usually become an adjective, so "infection of the skin" becomes "skin infection", "cancer of the breast" becomes "breast cancer" and "abuse of nicotine" becomes "nicotine abuse". This works for many phrases with "of" but not all: "life quality" for "quality of life" does not sound right, so a writer in doubt checks with a colleague.
- A sentence that opens with a prepositional phrase, a habit especially of writers whose first language is not English, is often long and choppy and often leads to a there were or it is construction, so it starts with its subject instead: "By using whole body scanning in the emergency room, we saved more than 58% of diagnostic time in polytrauma patients" becomes "Whole body scanning in the emergency room saved more than 58% of diagnostic time in polytrauma patients", and "In 42 RA patients in 74 treatment courses, Drug A was given between 2004 and 2008" becomes "42 RA patients received 74 courses of Drug A between 2004 and 2008".
- Grouping common information compresses a list, but not with "respectively", which always makes the reader backtrack to find which number goes with which item, and so tires and confuses them. Both a list that repeats each item in full and one that uses "respectively" can be replaced in an abstract by "Decreases in the incidence of Herpes zoster by antigen dose were: 5 μg, 17.2%; 12.5 μg, 27.3%; and 25 μg, 55.2%".
- Omitting "the" saves words and is safe in some cases: in an abstract, "the patients" refers to the patients in the study, so "Patients treated with aspirin…" is understood as well. The definite article is extremely important in English, though, and removing it can change the meaning entirely; in the main text, "the patients" may refer to all patients or only to a group, depending on the context. Deleting "the" and "a" can also easily make an abstract sound like a series of notes, so other ways of shortening it are worth considering.
- Plurals also eliminate articles: "The vaccine was administered by microinjection" becomes "Vaccines were administered by microinjection".