Picture a scientist employed by a pharmaceutical company publishing research results that were important bricks in a wall of evidence that could massively benefit their employer down the track.
Imagine a geologist employed by a fracking corporation publishing data showing that fracking posed few, if any, risks to adjacent water tables, again helping make the case for a policy allowing uninhibited fracking.
Now imagine, in both cases, that the researchers concerned, while not concealing their employers’ details, declared that they had no conflicting or competing interests. You could hear the incredulous laughter resonating throughout the research community and beyond.
Two such statements recently caught my eye at the end of research papers authored by people employed by or contracted to tobacco companies.
In a 2014 paper, scientists from Philip Morris reported what happened to rats forced to inhale a prototype “modified risk” tobacco product. Compared to a standard cigarette, the rats had “much smaller” gene expression changes when exposed to e-cigarettes. The Philip Morris authors declared simply that “there are no conflicts of interest.”
And in a 2015 paper on e-cigarettes that concluded “e-cigarette aerosol showed little cytotoxicity”, the authors declared:
The authors report no conflicts of interest and are employees of British American Tobacco or contracted by British American Tobacco. … All work conducted was funded by British American Tobacco.
What are we to make of these statements?
Having any sort of financial relationship with a funder with deep commercial interests in the outcomes of the research is the most basic example of a conflict of interest that needs declaring. And of all the types of financial relationships that authors should declare, employment by a company which may directly benefit is peerless.
The World Association of Medical Editors (WAME) argues that:
Commercial sources of funding, by companies that sell drugs and medical devices, are generally seen as the most concerning, perhaps because of many well-publicized examples of bias related to ties to industry.
WAME does not explicitly mentioned tobacco companies, but the history of that industry’s conduct in this regard is long and well documented.
Professor Lisa Bero, co-chair of the global Cochrane collaboration and now at Sydney University’s Charles Perkins Centre where she specialises in studying bias in research, has reviewed the history of how the tobacco industry uses research to pursue its ambitions and manipulate research on the risks of smoking. She has also found broadly similar strategies in at least five industries: tobacco, pharmaceutical, lead, vinyl chloride, and silicosis-generating industries.
These include:
- manipulation of the research question to obtain predetermined results
- funding and publishing research that supports industry interests
- suppression of unfavourable research
- distorting the public discourse about research
- changing or setting scientific standards to serve corporate interests
- disseminating favourable research directly to decision-makers and the public, bypassing the normal channels of scientific discourse.
In 1977, for example, Philip Morris internal correspondence shows a study investigating nicotine withdrawal was given the go-ahead but that officials discussed burying the work if “the results with nicotine are similar to those gotten with morphine and caffein (sic)”.
The industry understood the extreme sensitivity of such internal correspondence and documents for litigation and subsequently embarked on the now infamous large-scale Orwellian-named document “retention” process of document destruction.
The track record of the tobacco industry in dubious research and research communication practices makes it globally unique in being the only industry which many universities ban from research engagement. Google “university tobacco research funding policies” and you can spend hours reading the list of prestigious universities which have done this.
Conflict of interest declarations are not admissions of wrong-doing. There is, of course, plenty of sound science conducted inside industry and I make no comment here about the quality of the science in the two papers I have named.
Competing interest declarations are, rather, alerts to editors, reviewers and readers to consider the provenance and context of the work that has been done and to give it particularly close critical attention in light of those contexts.
Declarations also allow researchers to conduct meta-analyses of industry funded and non-funded research on the same areas to see if the funding source is associated with the study outcomes. Repeatedly and unsurprisingly, this has been found to be the case with tobacco, pharmaceuticals and medical devices, food and others.
For this reason, the Cochrane community’s policy on who is allowed to sponsor, author, review and edit reviews of evidence in medical science explicitly excludes all those with any commercial interests.
Perhaps the tobacco industry scientists on these two papers hold the interesting view that by stating their company affiliations, this entitled them to state that they had no conflicts: by naming their employers, this somehow gave them a free pass to say they were unconflicted. The “no conflicts of interest” statement is traditionally used by authors who in fact have no competing interests.
Some research journals such as PLoS Medicine, PLoS One, PLoS Biology, the Journal of Health Psychology, all journals published by the American Thoracic Society and Tobacco Control now refuse to consider manuscripts with tobacco industry funding because of the long history of misconduct.
When I edited Tobacco Control for a decade, I declined to introduce such a policy, perhaps naively believing that rigorous peer review would be a fail-safe.
Journals that accept manuscripts from tobacco industry employees and sponsored scientists need to consider their position regarding their competing interests declaration policies in cases such as these. Being employed by a company with direct commercial interests in the research in question should never allow authors to mock the process by declaring they have no conflicting interests.