“We are no longer cobblers who assemble entire shoes,” says Henning Tiemeier, a Professor of Epidemiology at Erasmus MC. “You can’t assemble an iPhone on your own, either. That is done by assembly line workers working at the many production units of an Apple factory.” Certain academic studies must be regarded the same way, in Tiemeier’s opinion.
Tiemeier’s work has received a lot of attention lately, not because of its content but rather because of its amazing quantity. In 2016 Tiemeier published a whopping 76 articles, according to a high-profile analysis by Stanford professor John Ioannidis, published in Nature. This makes Tiemeier one of the most prolific scientific authors in the world. The epidemiologist had not even read three of the 76 articles published under his name, he indicated in a response to the Nature article.
But Tiemeier is not the only highly prolific author in Rotterdam. Erasmus MC has a startling nine researchers who publish more than seventy-two titles per year – an average of at least one every five days. This means Erasmus MC has more extremely prolific authors than any other institution in the world. By way of comparison, the average academic working in the Netherlands, all fields combined, publishes 1.8 articles per year, according to figures released by the Rathenau Institute in 2016.
When authors publish close to one hundred articles per year, it seems unrealistic that they are making a ‘substantial contribution’ to the studies in question, as required by VSNU’s recently amended integrity code. Critics, including Frank Miedema, a Professor of Immunology at the Utrecht University Medical Centre, pointed out as much in a response published by De Volkskrant. Ioannidis asked in his Nature article what being an author actually means these days.
Tiemeier was willing to explain his side of the story, he told us via Skype from Harvard University in the USA, where he became a professor in early 2018. He explained that sometimes using a production line system is necessary in his field (epidemiology).
You didn’t even read some of ‘your’ articles!
“I did indeed state in my response to Ioannidis that I hadn’t read some of my 2016 publications – a very small number, three. However, I did do a lot of other things for those studies, such as data analysis, and I can definitely vouch for the quality of those analyses. But I did not wish to deny that I never read the final draft of the article that ended up being published. I made that admission on purpose, to engender a debate on authorship. I think this is a debate we mustn’t evade, nor should we pretend that nothing is wrong.”
What do you think is wrong?
“If two hundred people are listed as contributing authors to an article, it is clear that not all of those people made an equally significant contribution. In such cases, we are not authors in the sense of accepting accountability for all the analyses performed – but it does result in a high level of productivity. This is something that needs to be discussed. Right now there is no good way to deal with such large numbers of authors. That is the problem.”
But surely you don’t have to be listed as an author for every article with which you have been involved?
“It’s not as simple as that. Sometimes we need one hundred thousand human research subjects in order to be able to make a reliable statement. No research group in the world can get that many subjects on its own. For that reason, we collaborate with other research groups, in consortia of up to twenty groups all over the world. Each of these groups has some five or six people working on data collection and data analysis, and all these people must be credited for their work, which means it’s easy to arrive at over a hundred authors.”
So Ioannidis says: list these people in your acknowledgements.
“You can’t do that. Many journals won’t even allow you to do so. You can only be held accountable for your part of the work if you’re officially listed as an author – for instance, if something is found to be incorrect. Even if you didn’t make much of a contribution, it’s still essential to the results. Moreover, researchers sometimes devote weeks or months of their lives to such studies. In such situations, I don’t think it’s justified to relegate them to a quick mention in the acknowledgements, just because the project to which they contributed happened to be huge.”
Critics say: if you’re just providing data, you don’t meet the requirements for being listed as an author. Authors must have made a substantial contribution. Currently that is not the case. Is anyone still monitoring the quality of the overall product?
“That is indeed the trouble with these major studies conducted by consortia of research groups. No one has inspected all the raw data. I’d love to make all the data available to the public – open access. If we do that, we can go back to being asked for our expertise on the subject, rather than having to oversee our PhD students while they perform routine analyses for a month for yet another study. If we do that, the number of authors and publications will automatically be reduced. You can already see this happening at the UK Biobank, which has recently begun making data publicly available. We are not allowed to do that in the Netherlands, due to EU privacy legislation, which we interpret differently than the Brits do.”
So in your opinion, how should things be done?
“We must work our way towards open access. NWO (the Dutch granting agency – ed.) must pressure the parties concerned. They also wish to bring about open access, but they are not actually doing it. For instance, I myself applied for a grant a while ago to establish an open-access database. NWO rejected my application because it did not concern a project related to research infrastructure or a particular field of research.
“What I’d particularly like would be for us to be able to hire more paid analysts to perform the analyses, rather than using PhD students to do so. They don’t care about whether they are listed as authors or not.”
Many of your studies only demonstrate very minor effects. One could say not all of these studies need to be published.
“But minor effects are what our profession is all about! Suppose we are investigating the correlation between chemicals and autism. There are hundreds of these sorts of toxins, such as organophosphates or phthalates. It is absolutely inconceivable that all of those have a significant effect on complex cognitive phenomena such as autism. The same thing is true for genes. Theoretically, they can’t have a significant effect.
“In other words, epidemiologists simply have no choice but to focus on minor effects, but there are risks inherent in that. Either because we are not using the right models, or because we run the risk of accidental findings.
“But in general, the quality of scientific publications is improving. There is no doubt about that! [We are seeing an improvement] in the people, and in our research methods, which are getting better all the time. It’s hard to keep up with even the absolute top-tier publications these days because so many of them are getting published. Although I do admit there is indeed a truckload of really poor academic literature in my own field of study, and the number of those poor publications is increasing, too.”
Some of the extremely prolific authors mentioned in the article are quite upset. They think they are being called out for being bad researchers.
“I think being so defensive about it is a bit churlish, and moreover, it’s unnecessary. It’s the wrong attitude, and I’ve told my colleagues so. Look, Ioannidis is not saying, ‘Tiemeier’s work is of poor quality or unethical.’ He doesn’t say so anywhere. His part in this debate is to provoke people and get us to discuss things. We must take his concerns seriously, or things will never change. The system does need to be overhauled. I agree with him on that. And I want to help find a solution.”
