How abstract can an abstract be?

No, this isn’t the start of a tongue-twister, but an investigation into how many ways there are to write a research summary or ’abstract’. I’ve recently covered the difficulties of converting a research paper into a lay summary, but what about the challenges faces by creating an abstract, using just one tweet per paper or using only the top 1000 words? I’m interested to see what information is lost (or gained) in each method.

Thinking about this, I remembered a few things I’d been meaning to tick off an old ‘To Do List‘ after attending Spot On London 2013 in November of last year. Since then, I’d wanted to write an abstract for ScienceGist and have a go at the UpGoer 5 text editor. Over the next three posts, I’m going to try a few ways to report the same research by asking the question: how abstract can an abstract be?



ScienceGist Logo

ScienceGist is an online repository of lay summaries of scientific papers. All papers in scientific journals start with a short summary, called an ‘abstract’, but they are rarely lay-friendly. I wanted to try and stick to some of the journal abstract rules to avoid just rewording or even lengthening my competition entry.

An abstract should:

– Be short, less than 400 words, although some journals require them to be as little as 150!

– Cover the goals and achievements of a paper

– Be almost always formal in style

– Not have references (I wanted my ScienceGist to lack any hyperlinks or references to information I dug out when researching my first article).

To get started, I went straight back to the original paper and had the same problem that people always have when writing abstracts – What do you include? There are online guidelines and tips that seem to focus around a series of short, punchy answers to key questions:

1. What is the topic?

2. What is the problem and why is it not solved?

3. What is your idea to solve it?

4. Experimentally, how did you do this?

5. What do your results mean?

6. So what? What’s the impact?

Here goes with applying this…

“(Q.1) Arthritis is a disease of the joints that causes its sufferers a considerable amount of pain and discomfort. (Q.2) There is, however, a treatment for rheumatoid arthritis that works very well for some patients, although it is costly and doesn’t work for everyone. This treatment is the use of anti TNF-α drugs, named after the chemical in the body that the drugs stop from working. For patients who will not respond to the drug, valuable time is wasted as their arthritis progresses, untreated, at an unnecessary financial cost. Unfortunately, at the moment there is no way to know in advance if the treatment will work. (Q.3) This study looks at investigating the contents of a patient’s urine to see if it is possible to predict their response to anti TNF-α treatment.

(Q.4) The scientists in this paper took samples of the urine of 20 rheumatoid arthritis sufferers before they were given infliximab or etanercept, anti TNF-α drugs. After administering the drugs for around a year, patients were characterised as those that responded to the treatment and those that didn’t. The scientists then went back and analysed the pre-treatment urine samples to see if there were high levels of any common chemicals in the responders’ urine. (Q.5) It was found that four chemicals in the urine (histamine, glutamine, xanthurenic acid and ethanolamine) were higher in the responders’ urine before treatment. The scientists also analysed changes in the urine during the therapy, indicating important alterations in metabolism. (Q.6) This work could be developed to prescribe drugs to patients simply based on the contents of their urine, avoiding unnecessary costs and disease progression. At the same time, the exact same research was carried out with psoriatic arthritis sufferers but almost all of the patients responded well to the treatment and this didn’t give them a big enough sample size of non-responders for a fair comparison.”

Word Count: 307

So…What do you think? Is this enough information for you to feel you’ve got the outline of the paper? Would you prefer some statistics in the abstract, less background, more background?

What about…? You may have noticed that I left out any mention of side effects. This is because the research paper itself doesn’t mention them. I found this quite striking. In my experience, the side effects are often of the most daunting prospects about new or unnecessary medication and yet it was not deemed important enough to include in the research paper. In a prior lay-summary, I didn’t include anything about the psoriatic arthritis (after advice from my proof-readers) but, as a crucial part of the experimental work, it gains a place in the abstract version.

If I needed to get this summary down to 150 words I think I’d work from the last paragraph (currently 178), but inevitably this removes some of the background information and assumes prior knowledge.

Summaries for the general public are a valuable way to convey complex research but, from these early experiments, I’m getting more of an idea of how and why dry, but accurate, research can lose its scientific integrity by the time it’s seen in a magazine!

Leave a Reply

Your email address will not be published. Required fields are marked *