bioRxiv and medRxiv are not simply upload boxes. They moderate before posting, and that moderation determines which health claims reach the public first.
Both platforms are run by Cold Spring Harbor Laboratory. medRxiv is operated with Yale University and BMJ and launched in 2019; bioRxiv has been running since 2013. Every submission passes through a screening desk before it appears. The screeners are not peer reviewers, but they decide whether a manuscript enters the public record at all.
What the screening teams check before posting
bioRxiv screeners look for plagiarism, non-scientific content, offensive or defamatory material, and dual-use research concerns. They do not judge sample size, statistical choices, or whether a conclusion is likely to be correct. Their job is scope and basic integrity, not validation.
medRxiv performs that same check and then adds a clinical safety layer. A paper about a drug, device, behavioral intervention, or public health measure gets read with one question in mind: could someone act on this in a way that causes harm? If the answer is yes, the manuscript must show enough supporting evidence to pass.
That difference explains why some papers are fine on bioRxiv but never appear on medRxiv. A lab-bench study with a strong mechanistic claim is one thing. The same claim attached to a patient-facing treatment recommendation is another.
No one waits months for a preprint. The servers post most accepted manuscripts within days because their value depends on speed. That speed is also why moderation remains narrow. You cannot fact-check a global research pipeline in 48 hours.
Why medRxiv rejects clinical claims that bioRxiv would let pass
medRxiv requires human subjects research to carry evidence of ethical approval and informed consent. Clinical trials need a registration number, typically in a public registry, before the manuscript is posted. Authors who skip trial registration often get a desk rejection, not a request for revision.
The platform also screens out manuscripts that read as individual medical advice. A case report can pass on bioRxiv but face a higher bar on medRxiv because a single dramatic outcome can be misread as a treatment signal. Screeners do not decide whether the treatment works; they decide whether the presentation is safe for a public audience that includes patients and clinicians.
That is a form of misinformation moderation, and it predates COVID-19. The pandemic simply made the stakes visible to everyone.
The policy boundary for health claims
medRxiv does not ban controversial topics. It bans presentation that a reasonable reader could mistake for clinical guidance. A paper can report that an intervention showed a signal in a preliminary analysis; it should not report that the intervention should be used now. The difference often comes down to one sentence in the abstract.
This is where authors get frustrated. They know their work is exploratory. The public does not. medRxiv screeners apply the public standard, not the author's intent.
When COVID-19 forced faster, tougher screening
In February 2020, a bioRxiv preprint claimed the SARS-CoV-2 spike protein contained HIV insertions. Outside researchers spotted basic sequence alignment errors within days, and the authors withdrew the paper. It had already travelled widely on social media and in some news coverage. The episode showed how quickly a flawed preprint can become a public narrative.
medRxiv's caution banner exists for exactly that reason. Every manuscript carries a warning that the work is preliminary, has not been certified by peer review, and should not be reported as established information. The language is specific because the problem is specific: journalists and patients sometimes treat a PDF on a server as a finished study.
Nature reported in 2020 that medRxiv strengthened its screening in response to a flood of COVID-19 submissions. Manuscripts making treatment claims with weak evidence could be refused before posting. Read the current criteria on the medRxiv FAQ. bioRxiv states plainly that screening is not peer review in its own FAQ.
What screening still will not catch
No preprint server can promise that posted findings are true. Screening does not replicate analyses, inspect every image, interrogate datasets, or test whether a statistical model was appropriate. A paper can pass screening, receive wide attention, and then be withdrawn after outside criticism. That is a feature of rapid preprint culture, not a failure unique to bioRxiv or medRxiv.
Retraction metadata has made the aftermath more visible. Crossref now pushes retraction and withdrawal notices to preprint servers, so a paper that is later retracted can be flagged where readers first saw it. We covered that shift in our earlier report on Crossref retraction metadata reaching preprint servers.
Preprint overlay journals offer another signal. They take strong preprints and arrange formal peer review outside the traditional journal queue, which helps readers distinguish between a draft that has simply been posted and one that has survived outside scrutiny. You can see how that model is developing in our coverage of preprint overlay journals.
How to keep your preprint from bouncing
- Register the trial before enrollment if the study is a clinical trial. medRxiv screeners check for a registration number; missing registration is a fast rejection.
- Write limitations into the abstract instead of burying them in the discussion. A strong causal sentence with no caveat will draw extra scrutiny on medRxiv.
- Disclose funding, conflicts, and ethical approvals in the manuscript. Missing disclosure reads as a screening problem, not a formatting issue.
- Check the scope before you submit. bioRxiv and medRxiv do not accept the same types of work, and the health sciences gate is tighter.
- Read your abstract as if you were a journalist. If you would not want a headline drawn from a sentence, rewrite the sentence.
What this means for your next submission
Preprint moderation is not a verdict on your science. It is a filter on how your science will be read before peer review. The researchers who move fastest through bioRxiv and medRxiv are the ones who treat the screening questions as part of the writing process, not as an afterthought.
This week, take your next preprint abstract, highlight every sentence that could be read as a health claim, and soften the overreach. Ten minutes of that work prevents most screening rejections and makes the paper easier to report accurately. That is the whole game.
Photo by Julian Rösner on Unsplash