Writing an abstract

In the modern age of scientific publishing, the abstract to a paper is perhaps the most important section of a manuscript, apart from maybe the paper. The way in which we access scientific publication is changing rapidly. Not only do we no longer show strong allegiance to any particular journal, we often spring to judgements about what is interesting, what is of quality based on very little information read in just a few minutes. Typically research is accessed by searching in PubMed, Web of Science, etc and the decision to go to the full work is made by a skim through the title and abstract. This is a very strong reason for authors to make sure that abstracts are well-written, informative and interesting. What used to be a scientific summary is now effectively an advertisement for the work.

What should go into an abstract?

Some journals (JHND) included prefer a structured abstract with clear definition of introduction, methods, results and discussion under sub-headings. This makes things a bit easier for the author in terms of structuring the abstract and even if the structured approach is not the requirement of the journal, it is a good idea to produce a first draft with those headings and then remove them before submission.

However written, the abstract should consist of four key elements:

Problem statement or motivation: Why did you do this research? What is the nature of the problem that you were investigating? Why would a reader want to go on and read the rest of the paper?

Methodology/approach to the research: How was the data collected? Who or what did you sample?

Results/findings: What did the research show? Explicitly state the principal findings of your work. Include some numbers and statistical information if you can.

Conclusion/implications: How do the findings of this research take the area forward? What are the implications for the field?

All of these elements are important and a successful abstract addresses each of them concisely. The balance of the sections is important and your readers are most likely going to have the greatest interest in your statements about the methodology and the results so make these as detailed as you possibly can. The problem statement can be confined to just a couple of lines, but should include key words and phrases that make the aim of the work clear very early in the piece. The conclusion will always be regarded as a somewhat subjective element that will be of little interest to most readers (who being intelligent folk will want to draw their own conclusions).

Let’s have a go at this using a fictional study:

Background: Maternal obesity is increasing in prevalence raising concerns about the impact upon obstetric outcomes and the possibility of long-term effects on infant health. This study evaluated an intervention to manage weight gain in obese pregnant women [Nature of problem stated in the first sentence. The second sentence concisely states the aim of the study].

 

Method: 430 obese pregnant women (booking BMI 38.3±3.2 kg/m2) were recruited from antenatal clinics and randomly assigned to either normal antenatal care or an intervention comprising eight 1 hr sessions with a dietitian, focusing on weight control, healthy eating and physical activity. Weight gain and pregnancy complications were monitored through to 6 weeks post-partum. [Not overly descriptive e.g. we don’t know where the work was done, what dropout rates, or recruitment rates were, or the full nature of the intervention. This is sufficient however, for the reader to have some idea of the nature of the study. Read the full paper for greater depth and detail.]

 

Results: Women in the intervention group gained significantly less weight across pregnancy than in the non-intervention group (Intervention 4.2±2.1 kg, non-intervention 9.8±2.4 kg, P<0.01). Weight loss to 6 weeks post-partum was also significantly greater in the intervention group (P<0.01). After adjustment for confounding factors, the intervention group were significantly less likely to develop gestational diabetes (OR 0.45, 95% CI 0.32-0.58), pre-eclampsia (OR 0.35, 95% CI 0.22-0.48), hypertension (OR 0.68, 95% CI 0.38-0.98) or suffer post-partum haemorrhage (OR 0.85, 95% CI 0.72-0.98). There was no effect of intervention upon rates of caesarian section or infant birth weight. [Note that a lot of statistical information is presented and as a reader it is possible to see all of the main findings of the study at a glance. There is still a need to read the paper and see more- what were the confounders, for example? Were other variables measured?]

 

Conclusion: Intensive dietetic intervention in obese pregnancy can prevent excess weight gain and significantly reduce obstetric complications. This suggests that wider use of such interventions may be a cost-effective approach to delivering healthy pregnancy outcomes. [Short, simple and relatively non-subjective. The first sentence is a fair conclusion based on the data presented, and the second would appear to be a reasonable statement widening the importance of the findings.]

How can you go about writing an abstract?

I won’t pretend- it is difficult and I still struggle to condense the essence of what is often a complex 6000 word paper into just 250 words. It is often done in a bit of a hurry at the end of the writing process, and yet given the importance in terms of advertising the paper on search engines, the abstract is worth taking some time over.

Most journals state 250 words as a limit for an abstract (although some will go for 200 or even less). Stick to this limit right from the start so that you don’t have to make abrupt cuts in length later on. Some online submission centres will not allow you to proceed if the abstract is too long. Others may truncate the abstract at 250 words and therefore lose some of your intended meaning. Given the need for brevity, it is worth remembering what you don’t put into an abstract:

  • Do not refer to other work- unless it is for a conference (which is a very different abstract style), the abstract for a paper has no references.
  • Do not define terms (e.g. don’t waste space by explaining what gestational diabetes is), but do define abbreviations
  • Don’t overuse Greek letters or mathematical symbols, as these sometimes fail to display properly in search engines

There are many different ways to compile an abstract and all authors will go about it in a different manner. These are two approaches that I make use of and find helpful for identifying and organizing the key information.

The cut and paste method

I used to adopt this approach a lot, but favour it less these days because it involves a lot more editing at the end of the initial draft. With the cut and paste method, the author works through the full paper and cuts and pastes key phrases or sentences from each section of the paper into the draft abstract. This can be a particularly effective way of pulling out key information from the paper introduction, the methods and the conclusion, but still leaves the results section to be compiled largely from scratch. Once all of the ideas are in the abstract, the first draft will need to be reworked so that the ideas fit together in a logical order and so that the emphasis of each section actually delivers the same story as the full paper. Some word processors will effectively do this for you if you set particular rules, but only rarely do they come up with something useable.

The central ideas method

This is incredibly helpful if the final product has to be a structured abstract. Start off with the four main headings of Background, Method, Results, Discussion. Now just jot down some bullet points under each heading and then use these to flesh out short, but full sentences. E.g.

Background

Obesity increasing.

Obesity leads to greater maternal complications and poor fetal health.

Aim to intervene to reduce complications

Method

430 women from clinic

Randomised to dietetic intervention

Weight and complications to 6 wks post partum

Results

Intervention- less weight gain

Intervention- better post-partum loss.

Intervention less pre-eclampsia, diabetes, hypertension, PPH

No effect on birth weight, CS or other complications

Discussion

Intervention prevented pregnancy complications.

Benefit- cost-effective implication for wider use.

Invariably the product of both methods will be too long (especially if using the cut and paste method) and most likely phrased in a clunky manner. The draft will need careful editing to remove any repetition, extraneous material and to introduce a good flow to the text.

Advertisements

One thought on “Writing an abstract

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s