NEWS

Ghost authorship haunts industry-funded clinical trials

Drug companies make big contributions to analysis in the trials they fund but can fail to report their contributions.

Search for this author in:

Testing at a startup seeking answers to epilepsy

Academic researchers don't always get access to full trial data sets.Credit: Pat Greenhouse/The Boston Globe/Getty

An analysis of industry-funded clinical trials has found that drug companies are often heavily involved in research — but are not always transparent about it.

Kristine Rasmussen, a medical researcher at the Nordic Cochrane Centre in Copenhagen, and colleagues searched 7 high-impact medical journals, picking out the 200 most recent phase III and phase IV trials of drugs, vaccines and medical devices1.

The team found that both academics and their industry partners contributed to the design and reporting of most of those trials.

However, fewer than half the trials had academics involved in data analysis, whereas 73% had funders involved (see ‘Industry funders’ influence’).

Rasmussen suggests that lack of time or statistical know-how could mean that many clinicians are happy to leave analysis to funders. “We need to be extra aware when the data are analysed by funder only,” she says — particularly if no analysis plan has been published before the study begins.

Source: K. Rasmussen et al. Br. Med. J. 363, k3654 (2018).

To delve further into these collaborations, the researchers sent a survey to the lead academic authors on the 200 papers, asking 20 questions about their experiences working with industry. Around 40% of the authors approached completed the survey.

Among the responses, the researchers identified some instances of industry involvement that academics perceived as curtailing their academic freedom. Only 79% of the academic authors who replied reported having access to the entire trial data set, for example (see ‘Industry funders’ influence’). And 11% also reported disagreements with the funders, although these were usually minor and were resolved after discussion.

Ghost authorship

The responses also highlighted inaccuracies in declaring the contributions of industry employees.

About 21% of the academic authors indicated that a funder, or one of their contracted employees, had been involved in the design, analysis, or reporting of the research in a way that had not been declared in the paper. This “ghost authorship” could potentially be more widespread than this, write Rasmussen and her colleagues, as academic researchers who had a relatively small role in a study may not have been aware of the extent of industry involvement.

Rasmussen says she was surprised by how common these undeclared contributions and associated issues were. “It’s incredibly inaccurately reported,” she says. “The roles of the funder were often downplayed or even omitted in the publications, funder employees rarely had first or last authorship despite having played a role in every single part of the trial.”

The study adds “interesting new information” on the roles played by different parties in industry-funded trials, says David Nunan, senior researcher at the Centre for Evidence-Based Medicine at the University of Oxford. He is particularly surprised to see that only 40% of trials involved academic authors in data analysis.

But Nunan says that the poor response rate to the survey — only 80 of 200 lead academic authors filled it in — means that the figures the researchers give on the experiences of academic authors may not reflect the true situation.

Rasmussen acknowledges that it is possible the low response rate could have skewed the results in some way. She is particularly intrigued by 10 authors who opened the survey but never responded, suggesting that they could have had something they didn’t want to disclose. “Those would be interesting to talk to,” she says.

doi: 10.1038/d41586-018-06986-x
Nature Briefing

Sign up for the daily Nature Briefing email newsletter

Stay up to date with what matters in science and why, handpicked from Nature and other publications worldwide.

Sign Up

References

  1. 1.

    Rasmussen, K., Bero, L., Redberg, R., Gøtzsche, P. C. & Lundh, A. Br. Med. J. 363, k3654 (2018).

Download references