Drawn ideas/getty imagesReducing fat consumption may not be the best way to reduce the risk of heart disease and death, according to the most discussed study on social media in 2017. The paper, which instead recommends cutting carbohydrates, is top of the 100 scientific articles of greatest online impact, as ranked by Altmetric, a UK company that monitors the influence of scientific output not by number of citations, but by mentions on websites, social networks, Wikipedia, news portals, and blogs. Many of the studies discussed last year relate to issues such as gender stereotypes, stigma around mental illness, and the effects of climate change. Others address topics such as planets capable of supporting life and the risk of robots taking jobs from humans. The full list is available at altmetric.com/top100/2017/.
The paper that tops the 2017 ranking was published in The Lancet in August, and was authored by researchers from several countries, including Brazil. The study evaluated the diet of approximately 130,000 people from 18 countries over the last 10 years. According to the Scopus database maintained by publisher Elsevier, it has only been cited by 10 other scientific articles so far, but it had a major impact on a number of audiences: it was mentioned in more than 7,400 Twitter posts, 422 Facebook pages, and 130 news sites. “The data provided by Altmetric gives us a more complete picture of the immediate reach of research in society,” says cardiologist Álvaro Avezum, director of the Research Division at the Dante Pazzanese Institute in São Paulo, and one of the authors of the article. The number of citations, he says, is a well-established parameter for assessing the impact of scientific output, but it does not tell us how the results are received by the public.
The Lancet, BMJ, and JAMA dominate the list with 22 articles
The Altmetric ranking has been published annually since 2013 (see Pesquisa FAPESP, issue no. 250). Last year, the 100 articles on the list were mentioned more than 200,000 times in total. Most comments were made on Twitter (165,000), followed by news sites (25,000) and Facebook (6,164). “Social media is the most important source for assessing public engagement through sharing or commenting on scientific articles,” says British bioinformatics expert Euan Adie, founder of Altmetric. “For impact, it is more interesting to look at newspapers, magazines and so on,” he adds.
An article published in the journal Research Policy in May was ranked second, having been discussed not only in 7,300 tweets, but also in renowned newspapers and websites such as Yahoo and Spanish media outlet El País. The study found that one-third of 3,659 PhD students at universities in the Flanders region of Belgium were at risk of developing a mental illness. “Although the problem concerns the university environment, the pressures faced by those who choose to pursue an academic career and the psychological disorders related to life at graduate school are evident in everyday life, which is why there is so much public interest in this subject,” says biologist Atila Iamarino, one of the founders of the ScienceBlogs Brasil network.
49% of the articles are available to read for free
In third place was a paper from the Journal of the American Medical Association (JAMA), authored by researchers from various US institutions. The study examined 2011–2014 data from Medicare, the US government health insurance system, and found that patients aged 65 and over treated by female physicians had significantly lower mortality rates than those cared for by male doctors in the same hospital. The study suggests that differences in practice patterns between male and female physicians may have important clinical implications for patient outcomes.
“As you can see, many of the most popular papers are related to health and fitness,” says Euan Adie. Just over half the articles on the list (53%) were published in medical journals or journals aimed at those who work in medicine. “It is a field where scientific advances and discoveries directly affect people’s lives, unlike, for example, quantum physics.” One of the most successful articles in 2017 proposed a controversial methodology for diagnosing depression based on Instagram photos. The study, published in EPJ Data Science by researchers from the universities of Harvard and Vermont, USA, collected 43,950 photos published on the app by 166 participants. The images were then analyzed by software, which helped to identify that participants suffering from depression tended to post photos that were bluer, darker, and grayer than those published by other users.
69 papers have at least one author from the United States
For epidemiologist Carlos Augusto Monteiro, a professor at the School of Public Health of the University of São Paulo (USP), Altmetric’s list serves as a warning to the scientific community about how research results are being shared with the public. “Many health and medical studies are published on social media as definitive, even though the results do not always reflect a consensus within the scientific community,” says Monteiro. This is the case, he says, for the research that topped the Altmetric ranking. “Despite dominating the headlines, the study published in Lancet reaches very decisive conclusions, such as suggesting changes to dietary guidelines, and has been criticized by epidemiologists around the world for adopting questionable methodologies.”
Pharmacist Anderson Martino-Andrade, from the Federal University of Paraná (UFPR), featured on the Altmetric list as one of the authors of an article that identified a decline in sperm concentration among men in the western world of more than 50%. The study was widely reported on news sites, being mentioned 396 times across several countries. “We analyzed 185 studies published on the subject between 1981 and 2013, which examined a total of 42,000 semen samples. Some possible explanations for the falling concentration of spermatozoa include increased rates of obesity, smoking, stress, and exposure to pesticides,” says Martino-Andrade. “This is a controversial subject among researchers, because some people dispute the theory, denying that human reproduction could be affected. The public should be informed about these conflicts, which are very common in science,” he says.
The journal Nature has the highest number of articles in the ranking (16)
Altmetric’s Euan Adie notes that articles do not always gain notoriety on the internet because of their scientific contribution. “Many papers attract attention on social media due to errors detected after publication, or because the subject matter itself is controversial,” he says. The latest edition of the ranking did not include any cases of error or fraud, but he can recall a number of unusual situations from previous years. “There was once an article that made the top 100 because the author had included a marriage proposal in the acknowledgements section.” In 2014, a paper was widely discussed on social media because of an oversight by the authors and the publishers of the Journal of Ethology. The article, which described how levels of melanin—a pigment that gives the skin its color—differ in different environments, was published with an unusual comment about one of the references cited in the text: “Should we cite the crappy Gabor paper here?” The paper was replaced by a new version, but it was too late to avoid the deluge of comments on the internet, almost all of which were mocking the mistake made by the authors.
In order to evaluate the national and international importance of circulating scientific articles on social media, researchers from Brazil, Canada, and the US investigated how papers on the Zika virus were shared by Facebook and Twitter users in the first half of 2016, when the outbreak was declared an international emergency by the World Health Organization (WHO).
The study, published in the journal PLOS ONE in January, showed that although English is the most-used language on both social networks, about 24% of the posts on Facebook were made in other languages, particularly Portuguese and Spanish.
The research also found that comments on articles with at least one Brazilian author were 13% more likely to be written in Portuguese on Facebook than on Twitter. “The results suggested that scientific communication on social media regarding Zika was dominated by English, even though Brazil was the epicenter of the epidemic,” says Germana Barata, a researcher at the Laboratory of Advanced Journalism Studies of the University of Campinas (LABJOR-UNICAMP) and lead author of the paper.
“On Twitter, researchers publish more in English and thus reach a more global audience. On Facebook, posts written in non-English languages only have a local impact,” explains Barata, who is working in the field of information science during a postdoctoral fellowship at the Simon Fraser University in Canada.
According to her, the analysis indicates that scientific information would be better disseminated to the affected populations if publishers considered sharing the results via Facebook in Portuguese or Spanish.