Posting and citing preprints
Posting a preprint makes a manuscript available before peer review and calls for clear disclosure, version management and careful citation.
Benefits and risks
Medical writers should explain preprint options so authors can make informed choices, and support posting, disclosure, updates and citation if authors choose to proceed.
Common arguments for and against preprints can be arranged as paired PREPRINTS acrostics:
| For | Against |
|---|---|
| Primacy | Public misinterpretation |
| Record | Revisions |
| Engagement | Errors |
| Promotion | Permanence |
| Reproducibility | Rejections |
| Impact | Incompatibilities with journals |
| No cost | Not peer reviewed |
| Transparency | Trustworthiness |
| Speed | Scooping |
Preprints can appear within hours or days, while peer-reviewed publications may take months or years; authors can cite their own preprints meanwhile and list them on grant, promotion and job applications. Citing a preprint can be preferable to a conference abstract because readers can inspect the full manuscript. Preprints are free to post and read and typically receive a DOI, making them traceable, citable parts of the scientific record.
Open sharing increases engagement, inclusivity and transparency, and feedback before journal submission may improve manuscript quality and help address reproducibility problems. Authors should plan to track and answer comments, including potentially difficult exchanges and deliberate trolling.
A bioRxiv comparison found peer-reviewed articles with preprints had on average 49% higher Altmetric Attention Scores and 36% more citations than articles without them; where both versions existed, the peer-reviewed one was preferentially cited. In a preliminary comparison of BMJ Open articles published from June 2018 to September 2022, 308 articles matched to medRxiv preprints were compared with 1,000 randomly selected unmatched articles. Articles with preprints had higher citation counts at one, two and three years (P<0.00001, P<0.00001 and P<0.0001 respectively) and significantly more Twitter shares at all three points; excluding COVID-19 articles gave similar findings except for year-three Twitter shares. The preliminary analysis suggested a possible benefit in citations and Twitter shares.
Concerns about credibility arise because preprints lack peer review and a journal’s reputation; the press or public may treat them as fully reviewed, allowing poor-quality, misleading or biased information to spread and harm patients. Fear of being scooped can discourage posting, although public timestamps support claims to primacy; some regard priority claims themselves as unnecessary or inappropriate.
Journal compatibility and disclosure
Before posting, authors should check the target journal’s policy and alternatives, using journal websites or an aggregator such as Sherpa Romeo. In 2024, most journals and publishers considered preprinted manuscripts, with Springer explicitly encouraging primary-research preprints and not treating posting as prior publication. Preprints can undermine double-blind review because they expose author identities.
Posting before journal submission allows constructive feedback to be incorporated; simultaneous journal submission and preprint posting misses that opportunity. In 2024, more than 100 journals offered medRxiv-to-journal or journal-to-medRxiv integration, and some publishers invited posting on their own servers at submission.
The preprint’s first page should disclose that it is preliminary and has not been peer reviewed, explain peer review’s role in assessing manuscript quality and suitability for publication and warn against using the findings to guide clinical practice or reporting them as established news. The 2021 AMWA-EMWA-ISMPP statement also recommended watermarking and a clear disclosure within the article to distinguish unreviewed work.
Server choice and posting
Server scope matters: medRxiv’s 2024 scope excluded case reports, narrative reviews, editorials and opinion pieces. The ICMJE recommendations described in 2024 advised choosing servers that:
- Clearly identify work as not peer reviewed.
- Require interest disclosures and funding sources.
- Provide a clear route to report concerns, preferably with public commenting.
- Retain withdrawn-preprint metadata and notices giving withdrawal timing and reasons.
- Allow authors to identify subsequent peer-reviewed publication.
Posting advice is to use one server and post before journal submission. The 2024 medRxiv process was:
- Create an account and sign in.
- Select the subject area.
- Enter the title and abstract, author approval, competing interests, declarations covering author assent, ethics, participant consent, trial registration, legal responsibilities and reporting guidelines, data availability and its link, funding, and the clinical-protocol link.
- Complete author information and distribution or reuse licence options.
- Upload the manuscript; following approval it receives a timestamp and DOI.
Authors can then invite feedback through social media, email and other channels and revise constructively before journal submission.
Journal submission and version management
At submission, authors should disclose the preprint and DOI according to journal instructions, usually in the cover letter, submission system or manuscript. The DOI should expose the full version history, including versions previously rejected by another journal.
New preprint versions may be posted after journal rejection. The ICMJE advice described in 2024 discouraged posting the final published article or interim versions produced during peer review.
Once published, the preprint should link by DOI to the peer-reviewed article to direct readers and subsequent citations towards it. In 2024, medRxiv generated this link automatically, while PsyArXiv required manual entry of the publication DOI. Some journals also expected a manuscript statement identifying the earlier preprint, as in Addiction’s statement giving its URL and DOI.
Under medRxiv’s policy described in 2024, preprints could not be removed; authors could withdraw work if they no longer supported its findings or conclusions or discovered fundamental errors, leaving the original accessible with a withdrawn watermark and an explanation.
Citation guidance
Preprints can be found through indexes including Google, Europe PubMed Central and OSF Preprints. The 2021 joint statement and ICMJE recommendations described in 2024 agreed on clearly identifying cited work as a preprint, but differed on reference-list inclusion.
| Guidance | Citation treatment |
|---|---|
| AMWA-EMWA-ISMPP joint statement | Cite only like a personal communication in the text, using the link, DOI or both; clearly disclose preprint status and omit a bibliographic reference-list entry. |
| ICMJE recommendations | Clearly identify the preprint, include “preprint” after the reference-list citation, consider identifying it in the text, and provide its link and DOI where assigned. |
A preprint-server cofounder challenged the joint statement’s exclusion from reference lists because indexing services such as Google Scholar depend on those entries; the associations published a response.
Writers should monitor cited preprints for subsequent publication and replace the citation with the peer-reviewed article whenever appropriate; proofs after acceptance are usually the last opportunity to update it.