Preprints: What They Are and Whether to Post One

A preprint is a complete manuscript posted publicly before peer review. It has a permanent link, is citable, is free to read, and can appear within days of you finishing it.

It has also not been reviewed by anyone with expertise in your field, and that fact travels much less well than the findings do.

Preprints are now a normal part of medical publishing and orthopaedics has adopted them more slowly than most fields. This post covers what they are, what they are not, and when posting one makes sense. It sits within our research integrity series.

Screened is not reviewed

The single most important thing to understand.

Preprint servers do not conduct peer review. Because of the sensitivity of medical content, medRxiv does run a brief screening process — a basic check for offensive or non-scientific material, and for content that does not meet ethical research standards.

That is a floor, not an assessment. Screening does not evaluate whether the methods are sound, the statistics appropriate, or the conclusions supported. A preprint can be screened, posted, indexed and cited while being completely wrong.

The distinction matters most at the point where it is most often lost: a preprint reported as "a new study shows" carries an implication of vetting it has not received.

Accurate as of July 2026. Preprint server policies and journal positions change — check the current policy of your target journal and server.

Why post one

Speed. Peer review in orthopaedics routinely takes months. A preprint is available in days. For time-sensitive findings — a safety signal, a device problem — that gap is meaningful.

Priority. A timestamped public record establishes that you had the finding first, which matters in competitive areas.

Feedback before submission. Comments from readers can improve the manuscript before a journal sees it.

Access. Free to read immediately, without an APC — see Open Access and APCs.

Visibility for early-career authors. Work exists and is citable while it sits in review.

Funder requirements. Some funders now expect or mandate preprint deposition.

Why not

No quality signal. A reader cannot distinguish your careful study from a poor one on the same server.

Misinterpretation risk. This is the serious one for clinical work. A preprint suggesting a technique is superior can reach patients and journalists with no filter, and the caveats peer review would have imposed are absent. Retracting or correcting a preprint that has been widely reported is far harder than never posting it.

Peer review may change your conclusions. If review substantially alters what you claim, the earlier version remains public, and people will have read and cited it.

Journal policy. Most journals now accept previously preprinted work, but not all — and posting first can foreclose your preferred venue.

Field norms. Preprints are well established in basic science and less so in clinical orthopaedics. Some senior colleagues remain sceptical.

Journal policy has moved a long way

The old convention — that prior public posting disqualified a manuscript — has largely dissolved. Most major journals now accept previously posted preprints, and several actively encourage them.

The integration has gone further than many surgeons realise: 144 journals and counting are integrated with medRxiv, either accepting direct submission from the server (medRxiv-to-journal) or facilitating simultaneous posting on submission (journal-to-medRxiv).

Check before you post, not after. Individual journal websites carry their policy, and SHERPA/RoMEO aggregates them. Where the policy is ambiguous, email the editorial office — a two-line query removes the risk entirely.

The main servers

For clinical orthopaedic research, medRxiv is the default.

When it makes sense for orthopaedic work

Good candidates:

• A safety signal — an implant problem, an unexpected complication pattern — where delay has a cost measured in patients
• Methodological or protocol papers where feedback improves the final version
• Registry analyses where the data source is well established and the finding is descriptive
• Work by early-career authors who need the output to exist while review grinds on
• Where a funder requires it

Poor candidates:

• Small underpowered studies with a surprising positive result — precisely the ones most likely to be wrong and most likely to be picked up
• Anything whose headline could change clinical practice on its own
• Work where you have not yet had a statistician look at it
• Anything you are not confident enough to defend publicly, in its current form, today

That last test is the useful one. A preprint is publication in every practical sense except peer review. If you would be uncomfortable seeing the finding quoted in a newspaper tomorrow, do not post it.

If you do post

• Post the complete manuscript, not a summary or an extract.
• Label it clearly as a preprint that has not been peer reviewed — servers do this, but say it in the abstract too.
• Include the limitations section. The temptation to trim it for a preprint is exactly backwards; there is no reviewer to insist on it.
• Declare conflicts and funding, as you would for a journal.
• Use the versioning. When review changes the manuscript, post the updated version rather than leaving the original as the only public record.
• Link to the published paper once it appears, so readers land on the reviewed version.
• Be careful with media. If a journalist contacts you about a preprint, say explicitly and early that it has not been peer reviewed.

Citing preprints

You can cite them, and sometimes you should — but do it honestly.

• Identify it as a preprint in the citation, with the server and DOI.
• Cite the specific version you read, since later versions may differ.
• Weight it appropriately in your discussion. A preprint is weaker evidence than a peer-reviewed paper of the same design, and treating them as equivalent misleads your reader.
• Check whether it has since been published. If it has, cite the published version.

In a systematic review, whether to include preprints is a decision to make and state in the protocol in advance — see CONSORT, PRISMA, STROBE. Including them without saying so is a methodological problem.

The honest position

Preprints solved a real problem. Peer review is slow, and a system where useful findings sit invisible for a year serves nobody. Speed has value, and free immediate access has value.

They also removed a filter, and the recent history of medical publishing shows what happens when that filter is absent and the finding is dramatic. Both things are true, and neither cancels the other.

The workable stance for a clinical specialty like ours: use preprints for work whose value is time-sensitive or whose findings are descriptive, be cautious with small studies carrying surprising positive results, and never post anything you would not be willing to defend, publicly and unreviewed, on the day it appears.

The summary

Screened is not reviewed. Most journals now accept preprinted submissions and many are directly integrated with medRxiv — but check your target journal before posting, not after.

Post the full manuscript with its limitations intact, version it as review changes things, link to the final paper, and tell any journalist who calls that it has not been peer reviewed.

And apply the newspaper test before you upload.