Tracey Koehlmoos: Internal conflict of interest bias?

Tracey Koelhmoos If you work as a scientist, you also volunteer as a peer reviewer in the name of collegiality and service. I review about five or six manuscripts a month from a surprising array of journals–which always leaves me to ask, “How did they get my name and email address?”

Generally these are journals in which I have not published or worse still, been rejected from, and are manuscripts by authors I do not know.

Last night I finished reviewing an article about an intervention, but it caused me the most amazing internal conflict. Actually, I suspect that there will be a flurry of articles related to the particular non-state sector intervention coming out in the near future–or at least that is my judgment based on the stream of related articles I have been reviewing for the last four or five months.

On the one hand the manuscript makes a clear reference to one of my recent systematic reviews. On the other it uses the word privatisation to describe this particular non-state sector intervention and pretty much concludes that through this intervention the poor will suffer and not have access to health services.  (My review is just one source cited in this manuscript, and it literally says that there is some weak evidence of client satisfaction and no evidence of impact and a number of other important factors—all of which are within our grasp to evaluate and should be).

I was so torn. A citation for the long version of my new review would be fantastic. I am still smarting from the recent (and from the reviewers’ comments “unwarranted”) rejection of the manuscript version from the Bulletin of the World Health Organization, so that I am easily eight months away from being able to share the manuscript version of the review as an article, indexed in PubMed or available in a database or anywhere. It would raise the profile of the long version and potentially of the free review library in which it is published. Also, my underwhelming Hirsch-Index might improve—and I might say that the H-Index is meaningless but if I wish to be promoted again in my lifetime, I do think about it once in a while and generally feel powerless to improve it. Mostly, I was just amazed that some author, some stranger, some lone researcher, somewhere else on the planet had grasped so quickly something I had written.

The paper is however ideologically completely opposed to what I know to be current thinking about the non-state sector for health in developing countries. The language used about privatisation was very 1990s and clearly pre-dates Anne Mills’ seminal work from 2002 “What can be done about the private sector for health in low income countries?” [1]  Also, the author had skipped an entire body of literature around the need to engage the private sector in order to achieve the Millennium Development Goals–including a prominent debate piece on the issue [2].

Rather than turn the article around and say “I cannot review this because the author is using dogma,” I reviewed it. The background was a bit dated, the methods were strong and the discussion was very good. In terms of the research design and findings, it would be a good addition to the literature on this intervention. As a reviewer I know there is grossly limited evidence despite a lot of enthusiasm and investment for this intervention; the author had drawn a negative conclusion despite a reasonable analysis.

In my review of the article I wrote to the author and the editor. I explained we were on different sides of the private sector debate, and gave references to an abundance of more recent material that highlighted the change in thinking about the non state sector for health in developing countries. After all, the word “privatization” conjures images of the failing health systems of the former Soviet Block nations – grey and grim – and not at all related to the systems in poor countries where private practitioners are slogging away in rural areas and where there are huge gaps in the ability of the public sector health services to fill positions in less desirable locations.  (Clearly, the author did not read my last BMJ blog about the tide turning for the private sector in developing countries–but I thought it would border on literary narcissism to suggest it.)

Did I do the right thing? I was very clear in a note to the editor about being on the polar opposite side of the debate on the role of the private sector for health in developing countries. Thus warned, if the editor detects that my review forms a conflict of interest, she will no doubt disregard my input as a peer reviewer. Of course, the author might also complain because she perhaps wanted to make the point that she so emphatically concludes.

I should add that this was not an anonymous peer reviewer. As I do in all of my peer reviewer comments, I add my name at the beginning. Open peer review is important to me as there is little I dislike more than receiving what seems like a personal attack via peer review (i.e., “Clearly the lead author does not know what health systems research is and has no experience in the field.”) The peer reviewer felt confident in writing such comments under the cloak of anonymity.

I was completely transparent to both the editor and to the author. What would this be called: Conflict of Interest Bias? Peer Review Bias? Internal Conflict of Interest Bias? It leaves me wondering how often this happens and no one says a word.

[1] Mills A, Brugha R, Hanson K, McPake B (2002) What can be done about the private health sector in low-income countries? Bull World Health Organ 80: 325-330.

[2] Hanson K, Gilson L, Goodman C, Mills A, Smith R, et al. (2008) Is private health care the answer to the health problems of the world’s poor? PLoS Med 5(11): e233. doi:10.1371/journal.pmed.0050233

Tracey Koehlmoos is programme head for health and family planning systems at ICDDR,B and adjunct professor at the James P. Grant School of Public Health, BRAC University, Dhaka, Bangladesh.

  • Jcoynester

    I think we should be restrained in invoking “peer review bias”. It lets authors off the hook. So, the evidence of your bias is that you have published a systematic review on the topic, That should be a strength in your qualifications as a reviewer and should not exclude you even if–and, arguably, especially if you disagree with the author.

  • Thanks for this excellent thought, Tracey. I suspect there aren't many peer reviewers, commentators or practitioners who don't experience “conflict of interest bias” as a matter of routine. After all, if one is interested in a topic, it often comes with an opinion or conclusion – or “bias”.  Perhaps we should move from discussing these crudely as conflicts and find an alternative term, like “balance of interests”?  Or even “dialogue of interests”?  That might help develop a less fraught dialectic in other settings, as well.

  • Tracey

    Thank you Robert and Jcoynester for responding to the blog. I do feel that there must be a set of rules for peer review–that perhaps I did not receive during training or now when I am asked to complete the task. I suspect one of those rules is not to talk about it, which I have violated. I appreciate your responses because it makes me feel that my own experience was not so singular.

  • Richard Smith

    Generally there is no training for peer review. It's the blind leading the blind.