I am interested to know if journals have policies regarding multiple "rounds" of letters on the same topic. For example, the journal publishes a letter to the editor, and a reply from the authors of the article commented on. Then either the authors of the first letter or somebody else submits a letter commenting on the reply. Do you publish this second letter? With another reply? We have informally decided not to do this—one "round" is enough. What do others do? Decide on a case-by-case basis? Standing policy one way or the other?

David C. Cone
Senior Associate Editor, Academic Emergency Medicine
_______________________________
The short  answer is that it is decided on a case–by–case basis. 

If the letters are peer reviewed they should be viewed in the same light as any qualitative submission.

Sam Sussman
_______________________________
We have not a lot of pages available for correspondence, and we distinctly do not encourage follow-up letters to letters. We might make an exception for something very critical, but mostly we turn such letters down. We do not peer review letters, but make editorial decisions regarding how much they contribute to knowledge.

Michael D. Lockshin
Editor-in-Chief, Arthritis & Rheumatism
_______________________________
Good point—we do not peer review letters, either. They come to one of our senior associate editors (me), who reviews them with the associate editor who served as the decision editor for the paper being commented on.

Do other journals do true peer review on letters to the editor?

Dave
_______________________________
I haven't peer reviewed letters, but occasionally we do receive manuscripts disguised as letters, presumably so that the authors can get around peer review. We just direct them to submit these manuscripts that they called letters as the manuscripts they are.

Kimberly
_______________________________
We peer review letters, and have been doing so for many years. Reviewers often provide insights and facts that are unknown to me. Such information has led to both rejection of submitted letters and improvements in letters before publication.

David E. Bruns
Editor, Clinical Chemistry
_______________________________
Thanks for raising this important but neglected issue. What I find worst about letters is that, quite commonly, when correspondents have explained carefully that there are major flaws in a study, the study's authors get away with a cloudy reply. This smoke screen often succeeds confusing the readers, many of whom are not experts in the area in question, and who therefore do not know whether to believe the authors or their critics.

I have, as a critic, even been exposed to mere lies on part of the authors when they have tried to get off the hook, but only to have my second letter (where I document, with citations from the author's own article, that they are lying about their research), rejected. This is frustrating and does not serve the scientific community well.

This is an area that needs a lot more attention than it has so far. Editors should not let authors get away with beating about the bush, but should tell them that their reply is not satisfactory and that they need to reply honestly to the most pertinent criticism.

We have started to do research on letters and I would be most grateful to learn about articles that describe reviews of letters (eg, purpose, content, relevance).

Peter C Gøtzsche
_______________________________
I tip my hat to the editors of Headache. In this journal, I published a letter:

Berger, VW (2006). "Do Not Use Blocked Randomization", Headache 46, 2, 343.

The authors replied to this letter without really addressing the points it raised, yet doing just enough to appear to (as you said sometimes happens), and I felt the need to set the record straight. Sure enough, my follow-up was published:

Berger, VW (2006). "Misguided Precedent Is not a Reason To Use Permuted Blocks", Headache 46, 7, 1210-1212.

If only all journals were so enlightened! Along the lines of dishonest authors, I once reviewed a randomized trial for a journal, and found that the trial was in fact not randomized. The manuscript clearly stated in the methods section that assignment was alternated, not randomized. I pointed this out in my review, and the manuscript was rejected (for entirely different and unrelated reasons, I can assure you). Anyway, some time later I found the same manuscript published in a different journal, and it retained the word "randomized" (actually "randomised" as it was published in a British journal) in the title and abstract, but it had taken out the part from the text that would have allowed a careful reader to discern that this was not a truly randomized trial.  I contacted both journals involved, and was told by both that we must honor confidentiality, and so cannot report these authors to any ethics boards. About all that came of this incident was a sanitized description as one of the examples in a paper regarding general research misconduct:

Berger, VW, Ioannidis, JPA (2004).  "The Decameron of Poor Research", BMJ 329, 1436-1440.

It still bothers me that not only are these authors rewarded for their duplicity by having their dishonest manuscript published in a prestigious journal, but also that nothing has changed in the system. If the same events unfold today, then the same results will follow. Moreover, is there anybody who oversees author honesty to make sure that the published version of a manuscript matches the version submitted to a different journal that rejected it? Should such a body exist? Would it help to ensure the integrity of at least that portion of the research that gets published? Do authors have the absolute right to confidentiality, or is it only a prima facie right that can be taken away when it is abused (for example, by dishonesty)?

Vance
_______________________________
Squirrelly responses to critical letters may be a clue that the author's reply (and the original article) were written by medical writers directly or indirectly employed by industry. I know that busy editors don't want to be  detectives, but I suggest that named authors of non-responsive "responses" be telephoned rather than  e-mailed. An author who seems to be unfamiliar with the reasoning behind his or her response should be asked whether a medical writer "assisted" with the response,  at which point the editor should have a good idea whether or not  the original article was ghostwritten.

(Letters to the editor attacking good studies may also be ghostwritten. )

Adriane Fugh-Berman
_______________________________
I agree that the current letters system is an inefficient way of updating the scientific record. Many journals already have "eLetters" or "rapid responses" that allow readers to make their comments and criticisms public without delay. These also enable the reader and author to engage in a to-and-fro conversation regarding a particular topic and generally don't have cut-off dates for publication. (This also solves the point Dave raised regarding multiple "rounds" of correspondence, which could be a problem for a paper journal). At PLoS we are developing tools that allow readers to directly annotate published papers. This means that a reader can attach comments to specific sections of a published paper and those comments are themselves open to subsequent discussion. Annotations could be used to highlight minor or technical aspects, to add to or correct parts of the paper, to link to external content (such as entries for relevant material in public databases), or to start more extensive discussions. This annotations system has been launched as part of the PLoS ONE Web site (www.plosone.org), and you can see examples by selecting a paper under the "most annotated" link on the homepage. The annotations are accessible via the "See all commentary" link on the right hand side of a published paper. Anyone can comment and all contributions will be published, with moderation by journal staff to remove inappropriate content. (See our commenting guidelines at http://www.plosone.org/static/commentGuidelines.action).

Emma Veitch
Publications Manager, PLoS Clinical Trials
_______________________________
This is an interesting discussion. In addition to letters to the editor, some journals are creating weblogs (blogs), or electronic discussions following a paper's publication. They can carry on as long as the discussants wish, and they are often mediated. BMJ does this, as do other journals. Our new publisher wants us to lead their efforts in this area, which could be very interesting but also very time-consuming to manage.

Bill Tierney
_______________________________
We have a small letters section that mainly consists of papers that we reject, but then offer a letter (400 words, 10 refs, 1 table or figure if necessary) when a paper contains one "bite sized" piece of information that we judge advances the field. We very much encourage rapid responses to papers. We publish all that are not libelous, trivial, or otherwise worthless, and these remain permanently attached to the e-version of the paper so that readers can consider them alongside the paper. They are citable too. See  http://tc.bmj.com/cgi/eletters?lookup=by_date&days=90

Simon Chapman
Tobacco Control
_______________________________
We encourage Letters to the Editor commenting on recent articles published in our journal. They are not peer reviewed as such, but are assessed in-house to make sure they are factual and non-inflammatory, etc.

We also publish Research Letters. These are usually brief communications that do not justify a full article. They are, however, peer reviewed similarly to original articles, and thus authors of accepted Research Letters can include them on their CV as peer reviewed publications.

Rob Siebers
Editor, New Zealand Journal of Medical Laboratory Science
_______________________________
We also encourage letters to the editor to:

• comment (pros and cons) on recently published papers
• present results in very brief format (shorter than a Brief Report), as a reaction to a published paper or spontaneously as “research letter”
• provide a forum for spontaneous short opinions, hyphothesis launching, hot topics in the discipline…

One year and a half ago, we published an “editorial note” encouraging the submission of letters to the editor. We revised the “history” of this type of publication in the journal (we’re now celebrating our 20 birthday), showed examples of the topics covered by the letters published, explained why we love them, and gave some editorial advice on how to write them.

The PDF file of this note is freely available (in Spanish) at:

http://external.doyma.es/prepdf/water.asp?
pident_articulo=13080132&pident_usuario=17690&pident_revista=
138&fichero=138v19n05a13080132pdf001.pdf&ty=0&accion=
L&origen=doyma&web=www.doyma.es&lan=es

(or upon request directly from me!)

Esteve Fernández
_______________________________
Esteve,

All points of view have been very interesting as well as your editorial note. It would be important to know what other strategies you use to encourage letters to the editor in order to change some "cultural gap." In our journals (university medical journals or from medical associations) in Latin America, I feel  letters are used infrequently and improperly.

To educate our readers and to be a more interactive journal, we have designed a simple blog for Salud Uninorte (http://www.saluduninorte.blogspot.com/ ). We view this as an extension of the journal and feel that it is an adequate place to inform and to discuss editorial topics or articles of interest published by the journal. We  frequently receive e-mail but not formal letters (at least not in the way we would prefer).

Carlo V Caballero-Uribe
Editor, Salud Uninorte