Letters to the Editor: The Fastest Route to Publication

A PhD student emailed me last year, frustrated. Three years into her program, zero publications, and a supervisor telling her she needed "something on the CV" before the job market. She had read a trial in her field that she thought had a serious analytic problem — the authors had treated a time-varying exposure as fixed at baseline — and she had written four pages of notes about it.

Those notes became a 480-word letter. It was submitted eleven days after the trial appeared online, accepted three weeks later, and published with an author reply. Total writing time: maybe six hours. That is the shortest path from idea to indexed publication that exists in academic publishing, and most early-career researchers never use it.

Here is how to use it well — and how to avoid the version of this that makes you look careless rather than sharp.

What a letter to the editor actually is

In most biomedical and scientific journals, a letter to the editor (sometimes "correspondence," "letter," or "comment") is a short, peer-reviewed or editor-reviewed contribution responding to something the journal recently published. It is indexed in PubMed, it carries a DOI, it is citable, and it appears on your ORCID record like any other output.

There are two broad types, and journals treat them differently:

• Correspondence on a published article. You identify a methodological issue, offer an alternative interpretation, add context the authors missed, or report data that qualifies their conclusion. The original authors are usually invited to reply, and both pieces publish together.

• Research letters. A short original report — often a brief analysis, a small case series, a secondary analysis — presented in 600 to 1,200 words with one table or figure. JAMA, The Lancet, NEJM and many specialty journals run these. They are peer reviewed properly and are genuinely competitive.

This post is mostly about the first type, because that is the one with the shortest timeline and the lowest barrier to entry. But be clear with yourself about which you are writing, because the submission category you choose changes who reviews it.

What a letter is not

It is not a place to advertise your own work. Editors can smell a letter whose real purpose is to get three self-citations into print. It is also not a venue for grievance — if you believe there is research misconduct, that is a different process entirely, and the COPE flowcharts for raising concerns about published work are the right starting point, not the correspondence page.

When a letter is worth writing

The letters that get accepted almost always do one of five things. Check your idea against this list honestly:

1. A specific, demonstrable methodological problem. Immortal time bias. Collider stratification. A per-protocol analysis presented as intention-to-treat. A composite outcome driven entirely by its softest component. These are the bread and butter of published correspondence.

2. A reporting gap against an established standard. If a trial report omits allocation concealment or a systematic review doesn't state whether a protocol was registered, you can anchor the critique in EQUATOR Network guidelines — PRISMA for reviews, CONSORT for trials, STROBE for observational studies. An editor reading "the report does not state whether outcome assessors were blinded (CONSORT item 11b)" takes it seriously immediately.

3. Data that qualifies the finding. You have a cohort, a registry extract, or a national dataset showing a different pattern. Even a single contrasting figure is publishable as correspondence.

4. Context the authors genuinely lacked. Regional practice differences, a regulatory change, a prior trial in a different population. This works best when you are closer to the setting than the authors are.

5. An interpretation that overreaches the data. A non-inferiority trial discussed as if it showed superiority. A cross-sectional association described with causal verbs. These are common and worth flagging.

What does not work: "interesting paper, we agree, here is our similar experience." Editors reject these at a very high rate. The letter must change what a reader would conclude.

The timing window is shorter than you think

This is where most people lose. Journals impose a deadline on correspondence, and it is usually measured from the date of publication — often the online-first date, not the issue date.

Typical windows:

• NEJM: within 3 weeks of publication

• JAMA network journals: within 4 weeks

• BMJ rapid responses: open-ended, but formal letters are selected from early responses

• Most specialty journals: 4 to 12 weeks, sometimes unstated

Check the journal's "Instructions for Authors" page before you write a word. If no window is stated, assume eight weeks and move fast anyway — relevance decays, and an editor is far more likely to run correspondence while the original paper is still being read.

Practical implication: set up alerts. A saved PubMed search in your area, run weekly, is the cheapest letter-generation system there is. For example, a standing search like semaglutide cardiovascular outcomes trial will surface new trial reports in days rather than months.

Word limits and structural constraints

Letters are small. Treat the limit as a hard design constraint, not a target.

The author cap matters more than people expect. If you have six collaborators and the journal allows three, you need an authorship conversation before submission, not after acceptance. The ICMJE authorship criteria apply to letters exactly as they apply to original research — substantial contribution, drafting or critical revision, final approval, and accountability. Honorary authorship on a 400-word letter is still honorary authorship. If your institution uses CRediT taxonomy, a short contributor statement is a clean way to document who did what.

A template you can fill in today

Nearly every successful correspondence letter follows the same four-move structure. Here it is, with sample sentences you can adapt.

Move 1 — Anchor (1–2 sentences)

Name the paper, the journal, and the finding you are responding to. Be neutral and precise.

Chen and colleagues report that early initiation of anticoagulation reduced 90-day recurrence in patients with atrial fibrillation and recent stroke. We would like to raise a concern about how the exposure window was defined.

Move 2 — The issue (3–6 sentences, the core of the letter)

One issue. Specific. With the evidence that it matters.

Before (vague): "The statistical methods seem inappropriate and may have biased the results in favor of the intervention."

After (specific): "Patients were classified as early initiators only if they survived long enough to receive treatment within 48 hours, yet follow-up for all groups began at admission. This creates immortal time bias favoring the early group. In a comparable registry analysis, re-anchoring follow-up to treatment initiation attenuated a similar hazard ratio from 0.68 to 0.91."

The second version gives the editor something to evaluate. The first gives them a reason to desk-reject.

Move 3 — Why it changes the reading (1–3 sentences)

If the effect is substantially attenuated by this adjustment, the clinical recommendation for a 48-hour threshold may not be supported. Readers applying this threshold in practice should know the magnitude of the uncertainty.

Move 4 — A constructive ask (1–2 sentences)

Invite a response rather than demanding a retraction. You are asking colleagues a question in public.

We would welcome the authors' view on whether a landmark or time-dependent analysis was conducted, and whether the results were materially different.

Then: references (3–5, all real, all checked), a one-line conflict of interest statement, and ORCID iDs. Register one at ORCID if you haven't.

The mistakes that get letters rejected

Raising four issues instead of one. A letter with four criticisms reads as a hostile review. Pick the strongest and let the rest go. If three of your four points are weak, the weak ones discredit the strong one.

Personal or dismissive tone. "The authors appear not to understand survival analysis" will be rejected even when it is true. Write "the analysis does not appear to account for…" instead. Editors protect their authors from being embarrassed in their own pages.

Missing the submission category. Submitting correspondence through the "original research" portal routes it to the wrong pipeline and adds weeks. Check the manuscript type dropdown.

Undisclosed competing interests. If you ran a competing trial, say so in the letter. Editors find out, and the discovery is far worse than the disclosure. ICMJE's disclosure form standard applies here too.

Citing yourself three times in five references. One self-citation, if genuinely relevant, is fine. Three is a tell.

Writing to a journal that doesn't publish letters. Some journals accept online comments only, which are typically not indexed. If indexing matters to you, confirm first. For unfamiliar journals, run them through Think. Check. Submit. and check listing status in the DOAJ — predatory outlets will happily "publish" your letter for a fee and give you nothing citable.

What happens after you submit

Expect one of three outcomes within two to six weeks: rejection (usually without review and without detailed feedback, which is normal for correspondence), acceptance, or acceptance conditional on trimming. Peer review, where it happens, is often a single editor or one external reader.

If accepted, the journal will typically send your letter to the original authors and invite a reply. Both publish together. Read the reply carefully when it appears — occasionally the authors will show you were wrong, and the gracious response is silence, not a second letter. A public exchange that runs to three rounds rarely makes either party look good.

If rejected, the letter is usually not resubmittable elsewhere, because it responds to a specific paper in a specific journal. Budget for that risk. It is a small risk against a small time investment.

Building the habit

The researchers who publish two or three letters a year are not doing anything exotic. They read new papers in their field closely, they keep notes while reading, and they can tell within a day whether a concern is letter-worthy. The bottleneck is almost never writing — 500 words is an afternoon. It is noticing in time and having the source material at hand.

That is a workflow problem, and it is worth solving deliberately. Keeping your reading, your annotations and your drafting in one place means a flagged concern on Tuesday can be a drafted letter on Thursday without re-reading the paper from scratch. A workspace like PACR is built around exactly that continuity — literature search, reading, summarizing and writing connected so nothing gets lost between steps. You can surface new publications in your area through its research and discovery tools, keep the critique notes attached to the paper that prompted them, and draft the letter against the source without switching tools.

Whatever system you use, the discipline is the same: read with a pen, check the reporting guideline, and when something genuinely doesn't hold up, write the 500 words while the paper is still fresh. If you want to try that workflow end to end, you can set up a PACR workspace and build the reading-to-draft pipeline once, then reuse it.

A good letter is a real contribution. It corrects the record, it tells editors you read carefully, and it puts your name next to a paper people are already reading. For the cost of one focused afternoon, there is nothing else in publishing that comes close.