Everyone knows you’re not supposed to just read the abstract, right? But how many social media posts; how many statements that include the words “evidence-based…”; how many strongly held convictions; are made on the basis of having read an abstract without reading the full paper?
I’ve done it. Have you?
Spin
The job of an abstract is to provide a brief overview and summary of a piece of research (check out another blog I wrote HERE, on how to interpret a research paper). From this perspective there is no reason to doubt that what appears in the abstract faithfully reflects the methods, results and interpretation of the study. Unfortunately, it is not the case; abstracts commonly misrepresent the findings of their study; sometimes called ‘spin’, ‘inconsistency’ or ‘mislead’. The problem spans numerous healthcare fields, including in physiotherapy/rehabilitation research (1, 2).
The reasons for spin include bias and compromised integrity on the part of researchers, perverse incentives in academic research systems, and failure of peer-reviewers and editors at journals. The implication is that the content of abstracts cannot be trusted as the basis for making clinical decisions.
A small ray of light is that the abstracts of Cochrane systematic reviews are less likely to contain spin.
What are the specific problems?
Researchers that have looked at the problem of spin/inconsistency, have identified a few specific issues that have important implications for readers of research articles.
1. Recommendations not supported by findings
Sometimes the clinical recommendations in the abstract do not align with results of the study. For example, researchers might find that lower limb strength is correlated with pain intensity in a sample of people with back pain and then recommend in the abstract that clinicians provide lower limb strengthening for their patients.
2. Selective reporting of beneficial outcomes
Studies often measure a number of outcomes, but the good quality studies specify one (preferably) or a very few primary outcomes before the study starts, and then base the conclusions on the findings of them. Abstracts sometimes pick and choose which results to report based on what showed positive results.
3. Benefits reported despite high risk of bias
As well as what the study found (e.g. intervention A was more effective than intervention B), readers need to make a judgement about how reliable the information is, i.e. how much confidence or trust they should have in the findings. This is where risk of bias matters. There are two problems for people who only read the abstract:
- researchers make recommendations in their abstracts that do not take risk of bias into account, and;
- there is not enough information in the abstract for readers to judge risk of bias.
Implications
The reality is that readers cannot trust that information appearing in an abstract reflects the findings of the study. And further, abstracts do not contain enough information to assess risk of bias. This means that using only the information in an abstract could lead to implementing something in the clinic that is ineffective or even harmful.
Access to full text articles
The fact that full text articles sit behind paywalls, which limits access to clinicians and the public with an interest, is a huge problem for evidence-based practice, but beyond the scope of this brief post. That said there are options for readers who locate an article of interest. My recommendations are a direct email request to one of the article authors, or contacting someone with a university/academic appointment. Become friends with a researcher, some of them are just like regular people!
At Physio Network, we understand that time is also a huge barrier to keeping up to date with new research. Reading the full text of all relevant papers is desirable, but it isn’t always feasible. That’s why we started our Research Reviews – where industry experts review new papers and summarise them for you into 5 minute reads. Learn more about these reviews HERE.
In sum
Abstracts are worth reading. They are worth reading to decide whether you should read the full text, so they are good for screening out what is not worth following further.
They are not a good basis for forming strongly-held beliefs or for making clinical decisions. They are also not good for starting twitter arguments with strangers!
📚 Stay on the cutting edge of physio research!
📆 Every month our team of experts break down clinically relevant research into five-minute summaries that you can immediately apply in the clinic.
🙏🏻 Try our Research Reviews for free now for 7 days!
References
Don’t forget to share this blog!
Related blogs
View allElevate Your Physio Knowledge Every Month!
Get free blogs, infographics, research reviews, podcasts & more.
By entering your email, you agree to receive emails from Physio Network who will send emails according to their privacy policy.
Leave a comment
If you have a question, suggestion or a link to some related research, share below!