Conflicting Interests

Jan 19 2012 Published by under Uncategorized

When reviewing papers and especially proposals, you are supposed to declare conflict of interest if you are unable to be objective. For instance, reviewing the work of past collaborators generally falls under this category -- hence you include PhD and postdoc advisors, any people you may have advised, any coauthors on papers or proposals or co-editors in the last however many years on the lists of people whose materials you should not review.

But then there is a murky area of colleagues and friends who are not collaborators, i.e. there is no paper trail of collaboration (manuscripts, books, proposals coauthored).
Still, I should not review papers or proposals authored by, for instance, the woman who was the maid of honor at my wedding or any close personal friend for that matter. But what about not-so-close friends? Should everyone with whom you ever happened to have coffee or lunch at a conference be excluded? I would say this is taking the conflict of interest too far: it eliminates too a large portion of the reviewer pool -- you actually want competent people in the field to review; conferences serve as places where competent people meet and discuss science, often while consuming foods and beverages. They may end up liking one another.

Just because you like someone and know them a little doesn't mean you cannot be objective in reviewing their work. I think there is more danger in not being objective if you personally dislike someone than if you somewhat know and like them. But, honestly, how often do people recuse themselves from reviewing work of a competitor or someone they really hate rather than trash perfectly good papers or proposals in an attempt to sabotage the party they dislike or are in competition with?

The issue of properly gauging where the line of the conflict of interest lies is critical in panel review of proposals. We all want to have someone on the panel who likes our proposal and champions it, but how often does it happen that a person passionately champions a proposal authored by someone they have never heard of versus the proposal of someone they know and respect? In contrast, how often does the champion know the PI a little too well (e.g. drinking buddies over multiple decades)? At which point does being a good champion border on conflict of interest?

For instance, would you review a proposal or claim conflict of interest if the PI is:

1) Someone whom you knew in grad school way back when you were both students but you now rarely keep in touch?

2) Someone with whom you occasionally exchange emails and mostly talk shop, but you think they are an OK person and a great scientist?

3) Someone you invited to give a talk at your institution (or they invited you to talk at theirs)?

4) Someone you see regularly at conferences and may socialize in that context, but have minimal interaction between conferences?

(feel free to insert your own case...)

Dear readers, how do you decide when to claim conflict of interest and recuse yourself from reviewing because you know the author(s) a bit too much?
What level of professional and personal acquaintance is  still OK to comfortably agree to do the review and not feel like you have perhaps been playing favorites?
Has your answer been changing as your career progressed?


8 responses so far

  • Alex says:

    If somebody sees my presentation and chats with me and says "Yes, this is what the field needs" you'd damn well better believe I add that person to my list of suggested reviewers.

    And junior faculty should absolutely be inviting people in useful positions (e.g. journal editors, members of grant review panels, conference program committee members) to give seminars. In fact, seminar committee is one of the few committees that junior faculty should make a point of being active on.

  • I just declined to review for one of my PhD students!

    But I do review for people I know in my sub-field ALL THE TIME. If I didn't review them and if they didn't review me, then there wouldn't be anybody left knowledgeable on the material to review us. Sometimes I suggest R&R, sometimes I suggest rejection. I'm always polite whether I know the authors or not and I try to offer constructive suggestions even if it is a reject. I do a lot of reviews for one of the top journals in the field, and the editor would cut me off and think poorly of me if I started R&Ring when the other reviewers are saying to reject. I care about my reputation in the editor's eyes.

  • I wouldn't decline to review under any of the four circumstances you listed, nor would I decline to review for people who are close friends whom I have had in my house for parties or dinner or whatever. There are very few circumstances which create a conflict of interest sufficient to decline review, and include (1) you are related to the person, (2) you are involved in an active scientific collaboration, and (3) you are at the same institution.

    I have reviewed many papers by my former post-doc mentor, and she has reviewed many of mine. This is because it has been years since we have collaborated.

  • Anon2 says:

    I agree that none of the four situations you list constitute a conflict of interest. I have been pondering this issue lately because I am on a grant review panel that is reviewing several grants from people I know and I also have a grant that will be reviewed by a panel consisting of several friends from graduate school and my PhD advisor. Of course, my PhD advisor can't review my grant, but I would be very upset if the friends from grad school claimed a conflict and refused to review my grant. I WANT them to review my work because they have similar backgrounds and interests (though we are not direct competitors) and they would be the best judges of my work. On my own panel, there were a few grants where I was initially unsure about potential conflicts, so I spoke with senior colleagues to get a sense of what really constitutes a "conflict". The consensus was that I have a conflict with people from my institution and people I have mentored or collaborated with - people whose funding success could directly or indirectly benefit me. But as others have said, those who I know scientifically (or personally) and those whose research is similar to mine are the people whose grants I SHOULD be reviewing. Although this is not a quantitative measure, I ask myself "would I want this person to be reviewing similar grants from my lab"? If the answer is yes, then I should do the same for them. If the answer is no, there may be a conflict.

  • Jim Thomerson says:

    The first paper I was give to review was written by friends. They were describing a new subspecies based on an arithmetic mistake. I kept them from being embarrassed.

    Much of my work has been in a fairly restricted area where, at most, a dozen people are involved, including my ex students and coauthors. It is pretty incestuous, but I think we have all been objective and helpful to each other.

  • Devesh says:

    Those cases are very tight calls. Personally speaking they fall under conflict of interest.

    In scientific community a lot of good researchers get to know each other the ways you have listed. And, slowly tiers of researchers are formed. Now, if everyone starts counting that as conflict of interest then it is very possible that tier-1 people, so to say, will all get tier-2 reviewers and vice versa. Guess what? tier-1 will publish more and get more grants, tier-2 will suffer.

    And that's why I wouldn't decline reviewing their papers etc. in the cases you mentioned.
    However, what I must ensure is that I am fully "unbiased" while reviewing. Even if he is my good friend, and has submitted a weak grant -- reject it. You may not be friends afterwards once he gets to know, but at least we are servicing science the best we can. And that's the only thing we are responsible for.

  • Another Prof says:

    I wouldn't declare a COI for (2) and (4); if I were to, most people in my community will be ruled out!

    For (3), I will declare a conflict depending on how recent the talk was and if money was involved. For example, if I visit someone at a research lab the month before, and they pay me as a consultant for the visit, I'd say it is a conflict. If the visit was 5 years ago, and we haven't really kept in touch much since the visit, not so much.

    For (1) again, it depends on how long ago I used to know them. When I was a postdoc, I'd declare it as a conflict; now not any more.

  • Snow-Bound says:

    My field is double-blind, which seems to eliminate many of these potential conflicts. Of course, it's often possible to have a good sense of the author, which raises another question: if you're reasonably certain who the author is, but the review is double-blind, when does one declare a conflict?