Rise of the zombie ants
Why hype is creeping into scientific papers.
11 January 2021
When Jean-François Doherty first dipped into research papers on parasitic host-manipulation five years ago, he felt as though he was reading science fiction.
The technical jargon was peppered with colourful words and phrases such as ‘zombie’, ‘hijack’ and ‘mind control’. Even ‘puppeteer’ was used to describe host manipulation, whereby a parasitic organism significantly alters the appearance or behaviour of its host.
The “ample use of anthropomorphisms and words borrowed from science fiction” bothered Doherty, a PhD student studying host manipulation by hairworms at the University of Otago in New Zealand. “I knew these words were objectively inaccurate.”
A 2019 study on the taxonomy of a pathogenic fungus (Ophiocordyceps unilateralis), for example, described its hosts as ‘zombie ants’.
Another 2019 study, focused on jewel wasps (Ampulex compressa), stated that they ‘hijack’ the brains of cockroaches by triggering a neurochemical ‘storm’.
In October 2020, Doherty published a commentary in Proceedings of the Royal Society B on the pitfalls of sensationalism in host-manipulation research.
His analysis of more than 165 host-manipulation research and review articles indexed by the Web of Science found that sensationalistic words began creeping into the literature in the late 1990s, roughly a decade before they became regular fare in news articles.
Hopeful terms now more common
The use of exaggerated and emotive language isn’t just an issue in parasite research.
Buzzwords such as ‘breakthrough’, ‘game-changer’, and ‘miracle’ are common in news articles about cancer research, even when there is no clinical data to support such claims.
In the scientific literature, words that invoke hope have also become more common. A 2015 study published in The BMJ found that the appearance of positive words such as ‘novel’, ‘innovative’, and ‘unprecedented’ in PubMed abstracts jumped from 2% in 1974 to 17.5% in 2014.
This spin can mislead readers. A 2017 paper published in BMC Medical Research Methodology reviewed 17 studies on discrepancies between what was claimed in abstracts of biomedical studies and what was reported in the full text.
The review found that nearly half of these studies uncovered inconsistencies between abstracts and their full text, with 19% citing major discrepancies. Two studies cited examples where non-significant results were framed in overly optimistic terms in the abstracts.
The pressure on researchers to produce results with social and economic impact could be driving hype in biomedical research, says Timothy Caulfield, who studies research ethics and public representations of science at the University of Alberta, Canada.
“You’re less likely to see it in a field such as particle physics, where people understand that the results are going to be more incremental, and not about a therapy that’s going to benefit them and their kids,” says Caulfield.
Hype in COVID-19 research
Although the global research community was quick to join forces in response to the pandemic, the rush to discover and share life-saving solutions has driven hype and misinformation, says Caulfield.
“There’s an incredible time pressure involved in getting research results, and on top of that, there’s fear,” says Caulfield. “If you take scientific plausibility, plus emerging evidence, plus social value, the results can sometimes get twisted or hyped.”
In March 2020, a small but controversial study on the therapeutic potential of the anti-malarial drug hydroxychloroquine and the antibiotic azithromycin reported “significant” and “promising” results. The authors of the study, published in the International Journal of Antimicrobial Agents, recommended that COVID-19 patients be treated with these drugs to “cure their infection”.
The paper quickly gained the media’s attention, though many critics in the scientific community described the study as poorly designed and its conclusions as deeply flawed. A wave of research investigating anti-malarial drugs as effective treatments for COVID-19 followed, despite a dearth of evidence.
Nikolaos Frangogiannis, a cardiologist at the Albert Einstein College of Medicine in New York, says the problem runs deeper than simple word choice. “Due to shortened peer review and the urgency to discover something new, there’s a real tendency for researchers to overinterpret their findings.”
The misinformation problem
Frangogiannis points to a July 2020 study in JAMA Cardiology, which analyzed the cardiac MRI scans of 100 patients who had recently recovered from COVID-19. The results state that 78% of the participants showed signs of heart abnormalities.
The study’s authors say the results suggest that COVID-19 patients experience “ongoing perimyocarditis” (inflammation of the heart) after COVID-19 infection, which may indicate a “considerable burden of inflammatory disease in large and growing parts of the population”.
Although many patients showed a slight increase in indicators associated with inflammation, the evidence wasn’t robust enough to support the study’s conclusions, Frangogiannis argued in an editorial published by the European Heart Journal in October 2020.
Media outlets were quick to report on the study's results, some producing dramatic headlines such as “Coronavirus may have ‘devastating impact’ on the heart” (BBC Science Focus) and “Heart of it: Coronavirus could cause SAME damage as a heart attack in 78% of patients” (The Sun).
“When you have messages based on overinterpreted findings that are conveyed to the general public, it leads to misinformation,” says Frangogiannis.
“It may spread fear in some cases, and in others it can lead to more enthusiasm for therapeutic strategies that are not particularly effective.”