How does peer review deal with Conflict of Interest?

Four hypotheticals.

  1. A paper is presented to the review board of a major medical journal. The authors are all researchers working for Merck, the paper is a safety assessment of a new drug by Merck. The conflict of interest is stated up front.

  2. A paper is presented to the review board of a major medical journal. The authors are all researchers working for Merck, the paper is a safety assessment of a new drug by Merck. The conflict of interest is not stated up front but rather discovered after the fact.

  3. A paper is presented to the review board of a major medical journal. The authors are all leaders of a non-profit science organization in favor of universal vaccination, the paper is a safety assessment of a new vaccine against an infectious disease. The conflict of interest is stated up front.

  4. A paper is presented to the review board of a major medical journal. The authors are all leaders of a non-profit science organization in favor of universal vaccination, the paper is a safety assessment of a new vaccine against an infectious disease. The conflict of interest is not stated up front but rather discovered after the fact.

How would a review board at a major scientific journal handle each of these cases? Would their proceedings differ in any significant manner?

(Short version: how would they react differently when faced with papers with disclosed/undisclosed financial/ideological/other conflicts of interest?)

The conflict of interest HAS to be stated upfront. I’ve seen it several ways.
Sometimes at the end of a paper there is a section labeled “acknowledgements”.
Relevant grants and stipends will be listed. Or under “disclosures” , each author lists any connections.

So if the conflict is unstated, the paper will be rejected? To what degree is this applied? Solely to financial conflicts? What about if the people writing the paper formerly worked for a corporation that benefits from the paper?

Good question. I can find out tomorrow.

This is from the viewpoint of engineering journals, which I am familiar with, not medical journals.
First, while a paper is assigned to an editor, it is sent out to a number of reviewers who are familiar with the area. I’d think it extremely likely that at least one would be aware of the association with the authors with a non-profit group. I’m not sure I’d even consider this a conflict of interest - every writer has subject they feel are important, and being involved with these is expected.
Getting funding is another thing, and is also harder to determine.

However, assuming that this is a reputable journal, and not one set up to further a particular position, it might not matter all that much. The paper needs to be reviewed on its merits. Even with no ethical issues, the author of a paper might well leave out stuff that does not support his thesis and skew the writing in his direction. I’ve seen papers where important metrics were left out of the results - I wasn’t the only reviewer who noticed.
The real danger is with the editors. An editor aligned with authors in some way could select a set of reviewers who will tend to accept the paper, and might accept one with marginal results. And an editor and reviewers clueless about the subject matter might miss gaping holes in the work.

We get authors all the time reporting on work which has become or will become products. A lot of it is really good, and the bias in the work is no greater than the bias in the work of an academic trying to get more funding.

All conflicts of interest must be stated upfront. Most journals that I’ve seen require a statement to this regard somewhere in the paper, so if there’s no conflict, you state that there’s no conflict. If it is later discovered that there WAS a conflict of interest, then it would be treated similarly to any other falsification in a paper. Depending on how serious the conflict is, the paper could be amended or even retracted. The author’s reputation would probably be harmed, perhaps seriously, depending on the details.

It’s mainly a cultural thing within science, but it’s a very strong cultural thing. Science has to be honest and aboveboard for it to work, and anything that harms our credibility hurts everyone.

And with stated conflicts, does this have any relevant impact on the paper’s review?

It certainly would. I can’t speak from direct knowledge - there are other, more experienced scientists here who would be able to answer better - but if, say, they were being paid by the vaccine manufacturer, I’d expect the authors to make some statement about how their research was conducted independently of any pressure, etc, etc. I would also think that the journal they publish in would be a notch or two lower than if there were no conflict. The existence of the conflict by itself just makes the paper less trustworthy. That doesn’t mean the research is worthless or shouldn’t be published; they just need to let the audience know what’s going on.

What about the review board? Is it disclosed where they work?

In my experience with the journals I’ve been published in, after the paper has been accepted but before it is published, you have to submit and sign a form (pdf) specifying what a conflict of interest means and indicating that all of your conflicts of interest have been disclosed.

If the conflicts were disclosed at this point, the editor would probably just publish an amended version of the paper with the conflicts indicated. But if they were severe its possible that they could be sent back for re-review, or simply rejected by the editor.

If the conflicts were discovered after publication, its possible that the journal could put out an errata in a later issue linked to the paper indicating that certain conflicts were not disclosed.

We generally assume conflict of interest = financial.

I’m unaware of any consideration of “ideological conflicts of interest” or even how that could work.

If I’m a Republican (but not receiving money from them), I should “disclose” that fact if I publish a scientific journal article that shows evidence that global warming has a non anthropogenic source?

No, I don’t think that should happen.

If they are paying me (or someone associated with them is paying me) such that I have a financial interest in the research, yes, I should disclose that fact, but not simply if I have an “ideologic” interest in the research.

What I mean by ideological conflict of interest is, for example, someone who has a severe stake in something being true/false yet no direct financial claims involved. Say, the leader of an animal rights non-profit when examining a study about whether or not a certain type of animal can feel pain - they clearly have a vested interest in the study going a certain way, even if this is not a financial interest. Or is that simply a non-issue?

How are conflicts of interest, or the lack thereof, verified? Obviously a few people are going to lie on the form. Is there a standard background check process that pops out a list of an arbitrary person’s conflicts of interest (to a reasonably high level of reliability), or is it pretty much ad-hoc?

Absolutely, if the reviewer has a conflict it should be stated and then the editor will decide how to proceed. Most likely the invitation to review will be rescinded. At the journals I’m experience with, authors have to answer conflict of interest questions as part of the submission process. Having a conflict is not necessarily a deal breaker at all.

I would think that potential conflicts of interest by the reviewers are more significant. After all, the author of the paper is going to be competing for the same jobs and funding as they are, so it’s in their interest to fail the paper. Also the paper may promote or denigrate a reviewer’s own theory, in which case the reviewer has a bias. Especially if that particular scientific community is small or well-connected.

Well a person with an animal rights background wouldn’t be invited to review an animal study paper in the first place. Reviewers have to be vetted to ensure they can provide fair, factual reviews. Usually this means they are subject experts.

I’ve seen the opposite effect. A reviewer in the same tiny field as an author tends to give better ratings because more papers will tend to inflate the importance of the field, and thus the chance of getting funding for it. If your specialty is the left toe of the wombat, you want enough people working in the field so you can start a left toe of the wombat journal.

Woe betide anyone writing an article about the left toe of the wombat and not quoting any of the reviewer’s papers. This is obviously a sign of insufficient research.

None of these are conflicts of interest. The institution where you work is printed right under your name in the article title. Every reader gets to decide for himself what that implies about your trustworthiness, but I’ve never personally given a damn about who pays someone’s salary when evaluating a paper. It’s either sound or not. If I think I need to do something bizarre and unscientific as consider the author’s personal motives to decide whether the results are to be trusted or not, then obviously the paper is too light on the experimental detail and needs to be substantially revised. If you have to be an amateur psychologist to figure out whether the paper’s conclusions are justified by its data, then obviously the paper is in need of enormous revision. I would never recommend such a paper for publication. On the other hand, if a paper concluding that Jews ran the world was submitted, by A. Hitler working for the government of Germany, and the data were sufficient to support the conclusion, then I’d recommend publication. Doesn’t matter who the author is. When it matters who the author is, you have left science and wandered over into pseudo-science, a weird modern religious cult. (Some journals have experimented with double-blind reviews, in which the reviewers don’t know who the author is.)

What is normally called a “conflict of interest” is when there is some fact about your personal situation that could affect your authorship, or review, and it is not readily discoverable from the article (or your review) itself. For example, you work for the FDA but Merck pays you a fat retainer to advise them on drug trials. (Which would not be possible under FDA employment rules, but let’s say.) In that case, you would be expected to add a note to the paper, or to the editor, to the effect that Merck was paying your bills and that might be relevant if you are writing a paper on the effectiveness of a Merck drug.

Would that matter? Not to me. See above. If your paper doesn’t stand on its own, I don’t recommend publication. If it does, I do. Doesn’t matter who you are, nor who helps pay your mortgage. I would quit the field if it ever started seeming like that was important.

A conflict is generally not between the author and the research (as described in the OP), but between the reviewer and the author.