Monday, February 01, 2016

Violation of Conflict of Interest Principles and Seemingly Dishonest Disclosures Confirmed: Authors Should Have Revealed Past History of Funding from Big Pharma

Last Thursday, I revealed that in five recent articles on the effectiveness and use of smoking cessation drugs that were published in JAMA (one article) and Addiction (four articles), several of the authors failed to disclose the fact that they have received funding from pharmaceutical companies. For the JAMA article, the authors in question simply failed to disclose that they had received funding from the pharmaceutical companies. But for the four Addiction articles, the authors actually made a statement in which they actively denied having received pharmaceutical funding. They stated:

"The authors have received no direct or indirect funding from, nor do they have a connection with the tobacco, alcohol, pharmaceutical or gaming industries or anybody substantially funded by one of these organizations." 

As I documented in the previous commentary, there does not appear to be any doubt that the relevant authors have indeed received funding in the past from pharmaceutical companies. This history of Big Pharma funding is readily acknowledged by these authors in prior publications. Presumably, the "justification" for their failure to disclose these same conflicts of interest now is that several years have gone by, so the relevant financial interests were in place more than 3 years ago.

Both of the relevant journals refer authors to conflict of interest disclosure forms or policies. For JAMA, the authors submit the ICMJE conflict of interest disclosure form. For Addiction, authors are provided with instructions regarding conflict of interest disclosure and also told that the journal adheres to the Farmington Consensus, which includes a conflict of interest disclosure instruction.

It has been assumed by many that the ICMJE conflict of interest form only requires authors to disclose funding that occurred within the past 3 years. But this is patently false. The form actually requires authors to disclose all "relationships or activities that readers could perceive to have influenced, or that give the appearance of potentially influencing, what you wrote in the submitted work." This requirement does not specify that the relationships or activities must be current or within the past 3 years. There is no time frame given, so the clear intent here is to make sure that authors disclose any prior history of funding from corporations that could give the appearance of a potential conflict. I don't think that there is much doubt that the long and significant history of pharmaceutical funding and relationships for the relevant authors of the recent smoking cessation articles could give the appearance of a potential conflict. Thus, it seems clear that they should have been disclosed.

The confusion on this point may be present because in part 3 of the ICMJE form, it asks authors to disclose "all sources of revenue paid (or promised to be paid) directly to you or your institution on your behalf over the 36 months prior to submission of the work." If the form ended there, then there would be no issue with authors failing to disclose potential conflicts that occurred more than 3 years ago. However, the form does not end there.

In section 5 of the form, it requires that authors disclose "other relationships or activities that readers could perceive to have influenced, or that give the appearance of potentially influencing, what you wrote in the submitted work." And the word "other" is defined as meaning "anything not covered under the previous three boxes."

Thus, section 5 is asking authors to disclose any relationships or activities that existed more than 3 years ago if they could be perceived as potential conflicts of interest. Certainly, the extensive relationships that the relevant authors had with Big Pharma prior to the past 3 years could give the appearance of a potential conflict. It seems quite clear that they need to be disclosed.

Regarding the Addiction articles, the journal instructions state that "authors should declare sources of funding, direct or indirect, and any connection of any of the researchers with the tobacco, alcohol, pharmaceutical or gaming industries or any body substantially funded by one of these organisations." Importantly, the instructions do not tell authors not to disclose any funding or connections that occurred more than 3 years ago. In fact, there is no time frame given. Thus, any funding or connection with pharmaceutical companies should be disclosed, regardless of whether it falls outside of some arbitrary 3 year window.

Moreover, the Farmington Consensus clearly states: "Authors should declare to the editor if their relationship with any type of funding source might be fairly construed as exposing them to potential conflicts of interest." I think it is clear that a history of funding from multiple pharmaceutical companies, as is the case for three of the relevant authors, certainly does expose them to a potential conflict of interest. And importantly, the guidelines do not state that: "Authors should declare to the editor if their relationship with any type of funding source within the past 3 years might be fairly construed as exposing them to potential conflicts of interest."

The Rest of the Story

For all of these reasons, I think it is clear that the prior financial relationships between these authors and pharmaceutical companies should have been disclosed. The three year window mentioned in one section of the ICMJE form does not get authors off the hook from having to disclose conflicts that may have occurred prior to the past 3 years. And for very good reason. The perception of a potential conflict of interest is not going to suddenly disappear 36 months after an investigator has not received industry funding.

There is no credible argument to be made that journal readers or the public are not going to perceive a long history of funding from multiple pharmaceutical companies as not presenting a potential conflict as long as more than 1,095 days have gone by. There is no reason why the perception of a potential conflict will suddenly subside on the 1,096th day.

This is precisely why the ICJME form includes section 5 -- they want to make sure that there are no other potential conflicts, even beyond what was disclosed in section 3 for the past 36 months. And it is also why the Farmington Consensus does not specify a particular date at which a conflict will magically no longer be perceived as a conflict.

The principle here is most important: a conflict of interest does not represent wrongdoing, and it is actually in the authors' best interest to err on the side of being most inclusive. After all, conflict of interest procedures are largely in place to protect the author and the institution, not just the journal and the readers. Conflict of interest disclosures are not some sort of punitive procedure, forcing authors to pay penance for past wrongdoing. There is no wrongdoing. The sole purpose is to make sure that any financial relationships (no matter when they occurred) that could be perceived as potentially representing a conflict are disclosed. Importantly, a disclosure does not mean that an investigator was actually influenced by the conflict.

The rest of the story is that in both cases (the JAMA article and the Addiction articles), the past history of funding from, or relationships with Big Pharma should have been disclosed. The 36-month window in section 3 of the ICJME form does not relieve the obligation to disclose all potential conflicts that could reasonably be perceived as potentially influencing the conduct or reporting of the research.

Finally, there is an additional problem with the disclosure in the Addiction articles. Not only is there a failure to disclose the prior history of funding of some of the authors from Big Pharma, but the disclosure itself appears to be a false statement on its face. The disclosure states that the authors have received no funding, directly or indirectly, from pharmaceutical companies. But several of the authors have received funding from pharmaceutical companies. So this appears not only to be a failed disclosure, but a false declaration as well, which is seemingly worse than simply failing to disclose a particular conflict because it has the appearance of being dishonest.

No comments: