Wikipedia talk:Identifying reliable sources (medicine)

The section "Respect secondary sources" in the second paragraph talks about sources in away we don't use in normal articles, let alone medicine:

• If conclusions are worth mentioning (such as large randomized clinical trials with surprising results), they should be described appropriately as from a single study...If no reviews on the subject are published in a reasonable amount of time, then the content and primary source should be removed.

In other parts of wikipedia we simply do not include WP:EXTRAORDINARY claims based on new primary sources. Such work is not encyclopedic: we summarize reliable sources and we have no means of verifying any controversial new claim given only a primary source. In my opinion the content "A large study..." should have been reverted until additional sources were available.

Reviews per se are also not essential. A primary sources with numerous citations becomes a reliable source. The citations inform us about the reliability of primary source. Johnjbarton (talk) 02:58, 6 February 2026 (UTC)

A surprising result is not necessarily an WP:EXTRAORDINARY claim; this provision supports the opposite case of what EXTRAORDINARY describes as "Surprising or apparently important claims not covered by multiple mainstream sources". If there is a "Surprising or apparently important claim" that is "covered by multiple mainstream sources", then MEDPRI gives advice on how to handle it.

Keep in mind that most Wikipedia editors aren't scientists in the relevant field(s) and can't properly evaluate primary research for themselves. It's usually best for Wikipedia editors to avoid primary sources and look to either medical textbooks or peer-reviewed review articles. WhatamIdoing (talk) 03:43, 6 February 2026 (UTC)

Just to add on, even if an someone is a WP:EXPERT editor, we still can't evaluate primary research as anonymous editors (regardless of the number of citations of paper has).

With that said, the quoted text is one I have wondered about, moreso the idea that one could wait to remove the primary source and instead see if it gets coverage. That doesn't line up with other WP:PAG. I'd probably swap out the last sentence of then the content and primary source should be removed with something like then the findings of the primary source likely aren't WP:DUE to include. KoA (talk) 03:51, 6 February 2026 (UTC)

I disagree. We can evaluate primary research if it is cited. Citations are a public form of review. The citing work attaches the citation to commentary which informs us about the primary research. It may be a phrase, a sentence, or sometimes multiple paragraphs. This makes the citing works secondary sources about that aspect of a topic covered in that phrase/sentence/paragraph.

But the important bit in my complaint is that we should not include the primary until we see citations or reviews. The existing text in this article is backwards. Johnjbarton (talk) 18:00, 6 February 2026 (UTC)

A WP:SECONDARY source provides some sort of new evaluation or analysis, not just a plain citation. "One study found that the Sun is really big" is not secondary source material. "One well-designed study found that the Sun is really big" is (trivial) secondary source material. WhatamIdoing (talk) 07:00, 7 February 2026 (UTC)

"It's usually best for Wikipedia editors to avoid primary sources" So why does the second paragraph suggest including the primary? Johnjbarton (talk) 18:02, 6 February 2026 (UTC)

The overall section you are quoting from tries to balance the desire among editors to report a research finding that has widespread press coverage, but for which the academic secondary non-news literature hasn't yet got around to citing.

Wrt your last paragraph, a primary research paper with multiple citations might well be cited widely because it turned out to be wrong, or unsatisfactory, but somehow relevant to the new research or the review or textbook. For example, imagine the publication of a large scale RCT that demonstrated to the world that Wonderpam pills were merely "placebos with side effects". That research paper and subsequent reviews and textbooks might lament that so many case studies and pilot studies and badly designed or inadequately sized RCTs were conducted and appeared to indicate its worth, before we learned this sad truth, and cite each and every worthless and misleading one of them. -- Colin°^Talk 13:11, 6 February 2026 (UTC)

It is my understanding that Wikipedia represents verifiable knowledge, not "desire among editors". We routinely, day by day, suppress the desire among editors to report all manner of non-encyclopedic content.

As far as I can tell this paragraph is exactly a cautionary tale. It illustrates why we should not be citing new primary sources with challenging results. We should wait for secondary sources (including contra or confirming primary sources citing the new one). The text in this document gives bad advice. Johnjbarton (talk) 17:53, 6 February 2026 (UTC)

I think you'll find that it's more complicated than that.

For one thing, it's possible to have a non-academic secondary source. The general policies (e.g., WP:PSTS saying "Wikipedia articles should be based on reliable, published secondary sources, and to a lesser extent, on tertiary sources and primary sources") don't differentiate between secondary sources in the popular press and secondary sources in scholarly publications. This means that we could have WP:MEDPOP sources that are secondary but not what we'd like to be using.

Second, the MEDPOP sources could make inclusion WP:DUE.

Third, because Wikipedia's mission is education rather than rule following, it is sometimes appropriate to mention a misunderstood or overly hyped primary source. That often means writing something like "One study found a correlation between X and Y, though correlation does not prove causation" or "One study found that household bleach destroys viral particles on non-porous surfaces such as glass; however, experts say that drinking bleach is dangerous and does not prevent or cure COVID-19." WhatamIdoing (talk) 07:10, 7 February 2026 (UTC)

As far as I can tell all the commentators on this post agree that this article should not promote citing newly published and unconfirmed sources with unusual findings. There is a lot of variation in points of view, but I hope we agree to make a change to the document. I boldly did so and welcome your improvements and further comments. Johnjbarton (talk) 17:53, 7 February 2026 (UTC)

See this RfC for the full context, if needed. The relevant part of the proposal is written as follows:

Should this article mention that Khelif has reportedly stated that she has the SRY gene?

My question is simply the one in the title: if we were to include that in Imane Khelif, would we need to support it with MEDRS-compliant sourcing?

I'm aware that you've discussed very similar things before and concluded that we didn't need to do that (see 1, 2, 3), so I apologise if I'm asking you to repeat yourselves. I just want to make sure, since there's been a bit of disagreement over the sourcing requirements, that I'm not getting the wrong end of the stick or missing some nuance that means that we DO need MEDRS-compliant sources this time.

The secondary sourcing that we have is at Le Monde/AFP (and countless other iterations/edits of the AFP story in other newspapers of record and state media), The Independent, El País (Spanish newspaper of record), Público/Lusa (Portuguese newsaper of record/reputable news agency), Berlingske (Danish newspaper of record - paywalled, but text and a translation is here), and possibly others - all high-quality news sources reporting independently of one another, but not MEDRS-compliant by my understanding.

Thanks in advance. Clicriffhard (talk) 17:29, 12 February 2026 (UTC)

If the claim is that someone said she said she had SRY, no medicine is involved, no more than if she said she likes green apples.

I personally object to reporting rumors in wikipedia. If a source said she "stated that she has the SRY gene" and the source is WP:reliable then we should not say "reported". We should not cast doubts on sources. (I have no idea if her statement is encyclopedic knowledge, that is an issue for WP:BLP) Johnjbarton (talk) 17:37, 12 February 2026 (UTC)

Thank you.

I actually agree about "reportedly", but that's how the RfC has been phrased. The actual secondary reporting says that "Khelif confirmed she has the SRY gene", "She confirmed that she has the sex-determining SRY gene", and so on.

And yes, fully understood that BLP considerations apply regardless. The question is purely whether this is in scope for MEDRS. Clicriffhard (talk) 18:06, 12 February 2026 (UTC)

To say that someone HAS this gene, yes. To say that someone has CLAMED that they have the gene, no. Blueboar (talk) 17:37, 12 February 2026 (UTC)

Thank you. Clicriffhard (talk) 18:19, 12 February 2026 (UTC)

• Looking at the discussion there, it seems like you are the only person repeatedly raising WP:MEDRS; the comments you're replying to seem to be more about WP:BLP / WP:EXCEPTIONAL standards. As you mentioned, MEDRS applies to anything that would imply something biomedical about a condition (ie. condition X has attribute Y or implies Z) which covers a lot of what's appropriate on a page like that and does limit what we can say - many of the individual suggestions and edits people have made or are proposing there would violate MEDRS in that regard. But something like "figure X said shocking thing Y about their own medical history" is more about BLP. I'm particularly bothered by this exchange, , which appears to be what led to your post here - you were the one who introduced MEDRS to that conversation. Why do that and then come here to rebut the argument you made yourself? Throughout that conversation, why do you keep pushing them to frame their position as an MEDRS argument if you think that that isn't a valid argument to make? --Aquillion (talk) 17:48, 12 February 2026 (UTC)

  Looking at the discussion there, it seems like you are the only person repeatedly raising WP:MEDRS

  That's just not the case. Simon made it explicit that he was talking about MEDRS here, and there are numerous mentions of MEDRS from others in that thread.

  Of course I'm aware that we need to align with BLP requirements regardless, but that isn't the question that I'm here to clarify. Clicriffhard (talk) 17:59, 12 February 2026 (UTC)

"Far closer" is not the same as saying that something falls under MEDRS, and that's after you pressured them repeatedly by bringing up MEDRS yourself, and ignored more detailed responses from numerous other people. It feels like this is a WP:MOTTEANDBAILEY situation where you're so focused on reducing an argument to an easy-to-answer question that you've reduced it past the point where it has any relevance to the actual discussion whatsoever. Was bringing this question to this talk page really necessary for a single statement of I think we should, at minimum, cleave far closer to WP:MEDRS standards. I don't think a single, subsequently contradicted, statement reported in a single news publication is enough to opine about a BLP's medical history? Especially when you are, by far, the person who has brought up MEDRS the most in that discussion? --Aquillion (talk) 18:04, 12 February 2026 (UTC)

I asked Simon to clarify which policy or guideline he was referring to for his sourcing considerations, and his only response that referenced any policy or guideline referred to WP:MEDRS. Others also referenced MEDRS explicitly, and then someone raised a need to evidence "medical tests", so I came here to ask.

If anyone's inclined to assume that your account is accurate, I'd be grateful if they could simply read the conversation from here onwards, because it's completely misleading. Why, I do not know. Clicriffhard (talk) 18:15, 12 February 2026 (UTC)

@Aquillion it’s you who is constructing a motte and bailey, by suggesting that the use of sources ‘close to’ MEDRS is not a question relevant to MEDRS. Stating that Clicriffhard has mentioned MEDRS himself is also a non-sequitur. Please focus on the question. Riposte97 (talk) 20:48, 12 February 2026 (UTC)

Note: This has already been brought up in MEDRS before, for literally the same article. Wikipedia_talk:Identifying_reliable_sources_(medicine)/Archive_32#Does_WP:MEDRS_apply_to_medical_information_about_individuals? User:Bluethricecreamman (Talk·Contribs) 18:23, 12 February 2026 (UTC)

Thanks, but please also note: that discussion is linked in my question, and was referenced by me in the conversation at Talk:Imane Khelif. People don't appear to have accepted the conclusions of that discussion (and/or their applicability to this month's new developments and reporting), so I'm here to clarify whether there's some nuanced distinction that applies here or not. Clicriffhard (talk) 18:37, 12 February 2026 (UTC)

See also "Notable cases" under Wikipedia:Biomedical information#What is not biomedical information?

More generally, if a person says "I have ____" (any medical condition), the type of medical condition shouldn't matter for the sourcing. Editors need to not screw up the identification of the medical condition (e.g., having a naturally occurring SRY gene does not mean having a Y chromosome; having a gene does not mean having a functional gene; having an SRY gene doesn't even mean being male; having any disorders of sex development does not mean being trans; etc.), and they need to treat the information in an unbiased, boring fashion. Having a DSD should be treated as being no more interesting (or salacious) than having any other congenital condition. WhatamIdoing (talk) 05:47, 13 February 2026 (UTC)

Thank you, that's very clear. And yes, agreed - it's only relevant or significant to the extent that it's affected or stands to affect her boxing career/public life.

One quick follow-up: where secondary sources have said, for example, "Khelif confirmed she has the SRY gene, located on the Y chromosome that indicates masculinity" (1), am I right in thinking that the part I've bolded there is effectively a biomedical statement about what the Y chromosome is/does, and that (even if it were accurate, which it appears not to be in all cases) we should leave it out given the nature of the sources? Clicriffhard (talk) 11:28, 13 February 2026 (UTC)

If the sources are WP:MEDPOP, it's safe to assume that they're trying their best but aren't necessarily getting the details right. Here are facts that are actually true (numbered for convenience):

1. Khelif confirmed she has the SRY gene.
2. The SRY gene is normally(!) located on the Y chromosome.
3. The SRY gene is present and functional in biological males (i.e., individuals that produce sperm).
4. The SRY gene is present on an X chromosome in some people with disorders of sex development.
5. The SRY gene is non-functional in some people with disorders of sex development.

The first three appear to be intended by the MEDPOP sources. The last two are (apparently) relevant facts that editors should bear in mind while they endeavor to make the article be not-wrong. For example, editors should not repeat the oversimplification of MEDPOP source by claiming that the SRY gene "is" located on the Y chromosome. The Y chromosome is its normal location, but it's not its only possible location. In the context of this article, we wouldn't want to imply that normal is relevant. WhatamIdoing (talk) 18:39, 13 February 2026 (UTC)

Thanks again, and sorry for the delayed acknowledgement. I think I understand; now I just have to figure out how best to relay that to the handful of editors who are yet to accept that fact 1 is "actually true". Maybe tomorrow... Clicriffhard (talk) 23:49, 20 February 2026 (UTC)

Keep in mind #5: Just because a gene is (self-reportedly) there doesn't mean that it matters.

I suspect that this article will be easier to write in July 2028 than it will be this year. WhatamIdoing (talk) 05:20, 21 February 2026 (UTC)

 Courtesy link: Phoenician and Punic Sardinia

I have always thought of this guideline as applying solely to determining what sources are compliant for use in articles on biomedical topics, but now I wonder if it also applies in the other direction, so to speak, as well. What I mean is, suppose we have a non-biomed topic like Phoenician and Punic Sardinia and an editor adds the following content citing a source published in a biomed journal, like Sarno et al. (2021) from Annals of Human Biology (diff):

Sarno et al. (2021) analysed ancient Punic individuals from the necropolis of Tharros in Sardinia (5th – 3rd century BC). The study demonstrates that Punic-related individuals from Tharros fall in an intermediate position between modern Southern European Iberians and North African populations, while present-day Sardinians form a clade with pre-Phoenician Sardinian populations dated between the Neolithic and the Nuragic periods. The results shows that female mobility from Iberia and North Africa to Sardinia was associated with Punic communities.

I start to rebel already at the expression the study demonstrates as appropriating way too much power to one study (maybe concluded would be okay?), and the rest of it seems to be attributing too much to one study's conclusions. But it occurs to me that maybe we shouldn't be using this primary source at all in the article, and that the sentence quoted is a good example of why we have WP:PRIMARY on the one hand, and MEDRS to boot.

I am hesitating what to do about it. My first reaction is to take it out per MEDRS, but then I wasn't sure it applied in this case. Alternatively, it already qualifies as intext attribution so I suppose one could keep the content along with the source, but tone the wording way back, making it clear that it's less a done deal and more just this one study's conclusion. But I'm really not sure and would appreciate feedback. Mathglot (talk) 07:03, 3 March 2026 (UTC)

I think there's a small group of (non-medical) editors who are experienced with the problems of combining genetics with ethnicity (and therefore sometimes with politics). I can't remember where I last saw a dispute on this, but it was recent (probably this calendar year), and I think they'd be best able to give you sound advice. WhatamIdoing (talk) 05:55, 4 March 2026 (UTC)

Could someone take a look at the summary of medical evidence for Pilates (edit | talk | history | protect | delete | links | watch | logs | views)? Talk:Pilates#Proposal_to_update_"Effectiveness"_section_with_2024_review identifies a 2024 review that might be used, that has not received a response. --Hipal (talk) 19:24, 18 May 2026 (UTC)

Can someone look over this with the edit summary, Important to distinguish betw. scientific evidence & other forms of evidence; empirical & anecdotal evidence, observations of doctors, do indicate effectiveness. //I agree that better med cites can be obtained; but if old ones remain here they should be properly summarized - editors shouldn't sub their own opinions for medical evidence. Pls. read the actual articles cited. //Important to mention other therapies, so as not to suggest study was limited to Pilates. If disagree, go to talk page --Hipal (talk) 18:11, 19 May 2026 (UTC)

It looks okay to me.

Also, see the note at the very top of the page. Wikipedia talk:WikiProject Medicine is a better place to ask these questions. WhatamIdoing (talk) 21:03, 19 May 2026 (UTC)