Correspondence article support

What correspondence articles are

Journals publish short pieces in response to the papers they print. A letter to the editor comments on a published article, raising a question, pointing out a problem or adding evidence. The authors of the original paper are normally invited to reply. A commentary or viewpoint offers a longer interpretation of a topic or a paper, and a research letter reports a small study or a short analysis in a compressed format. In evidence synthesis, correspondence is often used to discuss methodological issues in a published review, such as the choice of model, the handling of heterogeneity or an error in extraction, and to report a re-analysis.

These pieces matter because the published record is a conversation. A paper that has an undetected error or a questionable inference remains in the literature and gets cited, and correspondence is the established, peer-moderated way of putting a concern on the record so that readers who find the paper also find the response. A good letter is brief, specific and fair, and it makes the literature more accurate.

What the service includes

When a letter is warranted

Not every disagreement deserves a letter. A letter is worth writing when there is a substantive issue that affects the conclusions or the use of the paper, when you can show it with evidence, and when readers would benefit from knowing. Examples include a data extraction error that changes a pooled estimate, a statistical method that is inappropriate for the design, an important omission of the search or the studies that bears on the conclusion, a result that is inconsistent with the text or a table, or a new study that changes the picture. It is less worthwhile for differences of opinion that rest on taste, matters already acknowledged in the paper's limitations, or points that could be made in a paper of your own.

Before writing, it is usually good practice to check whether a correction or a response has already been published, and to consider contacting the authors directly, which sometimes resolves the matter faster and more courteously. If the authors acknowledge an error, a correction may be better than a letter.

Structure and length

Journals set strict limits on correspondence, often a few hundred to about a thousand words and a small number of references, although the figures differ widely and must be checked for the specific journal. The brevity is useful discipline. An effective letter is built around one main point and perhaps one or two supporting ones, and it follows a clear order.

• Reference to the article, with its full citation and the specific claim, table or figure under discussion.
• The concern, stated precisely and in neutral terms.
• The evidence, which may be a calculation, a corrected analysis, a reference or a new dataset, given with enough detail to be checked.
• The implication, stating how the point affects the paper's conclusion or its use, without overstating it.
• A constructive close, such as a suggestion for how the issue might be resolved or a question to the authors.

Tables and figures are usually limited, so the central evidence should fit in the text or in one compact table. Supplementary material may be allowed for a re-analysis.

Evidence and re-analysis

The strength of a letter depends on whether its point can be verified. A claim that a meta-analysis was done incorrectly is far more convincing when it comes with a corrected analysis from the same data than when it is asserted. Where a published paper makes its data and code available, the analysis can be repeated and the discrepancy shown. When it does not, the numbers reported in the paper, such as the effect sizes and standard errors in the forest plot, are often enough to reproduce the pooled estimate and to test an alternative. This work is done carefully and documented, because a letter that contains its own error is worse than none. See meta-analysis and statistical analysis.

References should be accurate and relevant. The standard methodological sources, such as the Cochrane Handbook and the reporting guidelines, are used to show what is expected, and the claims are matched to what the sources actually say.

Common kinds of correspondence in evidence synthesis

Re-analysis letter

Reports that the results of a published meta-analysis change when an error is corrected or a more appropriate method is used, with the corrected figures and the method described briefly.

Methodological comment

Questions the choice of a model, estimator, effect measure or handling of dependent data in a review, and explains what the alternative would show.

Missing evidence

Points to relevant studies that a review did not include, for example a recent trial or a registry record, and says how they might alter the conclusion.

Reporting and transparency

Notes that a review does not meet the reporting guideline in a way that limits the use of its findings, such as absent search strategies or a registration discrepancy.

Research letter

Presents a brief original analysis, such as an updated meta-analysis including newer studies, in a compressed format.

Commentary

Interprets a body of evidence or a controversy, usually at the invitation of an editor or in a dedicated section of a journal.

Preparing before you write

A few steps make a letter stronger and save time. Read the article and any supplement fully, and note the exact statement you are questioning and where it appears. Check whether the journal has already published a correction or other correspondence on it. Read the journal's instructions on letters, since the length, the format, the time window and whether a competing-interest statement is needed vary. Collect the evidence you will cite and verify each number against its source. Decide in a single sentence what the letter is claiming, and what should change as a result, and if that cannot be written, the letter is not yet ready. Finally, ask a colleague who was not involved to read it for clarity and fairness. A piece prepared this way is short because it has been cut to its point, not because it is thin.

Tone and fairness

Correspondence is read by the authors of the original paper, by the editor and by the community, and its tone reflects on the writer. A letter should criticize the work and not the people, describe what the paper reports and avoid attributing motives, acknowledge what the paper does well where it is relevant, and be precise about what is disputed. It should disclose any competing interest, such as a relationship with a rival, a related study or a financial interest, because failure to disclose can damage the letter and the writer. Journals can and do reject letters that are personal, speculative or defamatory, and the Committee on Publication Ethics provides guidance on the handling of concerns about published work. Where a concern is serious, such as suspected data fabrication, it should go to the editor through the proper channel and not into a letter.

Replying to a letter about your own paper

When a letter is published about your paper, the journal usually invites a reply. A good reply addresses the points made, concedes what is correct, explains what is not, and does so concisely. Where the criticism has merit, acknowledging it, and correcting the record if needed, enhances credibility. Where it does not, the reply gives the evidence and avoids a defensive tone. If a letter reveals an error in the original paper, the authors should consider whether a correction is needed. Replies are subject to the same limits as letters and to the journal's deadlines, which are often short, so it is wise to start early.

Deliverables

Optional extension: when a letter needs to be longer

Limitations and boundaries

Journals decide which letters to publish, and many are declined, often for space or because the point is not judged to be important enough. The window for commenting on an article can be short, and some journals do not accept correspondence on older papers. No service can guarantee publication. We write only what can be supported, we do not write letters that are personal attacks or that conceal a competing interest, and we do not help to submit anonymous or false-name correspondence. The authors of the letter are responsible for it. This service provides research and publication support and does not provide clinical advice.

How long does it take?

A short letter is quick to draft once the point and the evidence are settled, and the time goes mostly into verifying the evidence and checking what the journal allows. Where a re-analysis is needed it takes longer. Journals often set short deadlines for responses to a paper after its publication, so tell us the date and the journal at the start.

Frequently asked questions