Wikipedia:Reference desk/Archives/Science/2019 June 23

Science desk
< June 22	<< May | June | Jul >>	June 24 >

Welcome to the Wikipedia Science Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

June 23

Autoimmune encephalitis

Is autoimmune encephalitis a disease? On one hand, it can result from a number of autoimmune diseases, so I'm guessing "no" because I envision a disease causing symptoms/syndromes/etc., not other diseases. But on the other hand, Category:Encephalitis is a subcategory of Category:Inflammatory diseases of the central nervous system, so presumably encephalitis is a disease. I can't see how both of these ideas could be right, so I must be misunderstanding something. PS, note that Autoimmune encephalitis is a member of Category:Autoimmune diseases, but that's because I just put it there before asking this question. Nyttend (talk) 04:27, 23 June 2019 (UTC)

As with encephalitis generally, it's a symptom, but with a slightly specified cause. You could refer to it as a medical condition without confusion, I think. "Disease" has an incredibly broad definition that could include this. Someguy1221 (talk) 05:59, 23 June 2019 (UTC)

Does a woman who had 3 or more abortions have elevated cancer risk?

After reading the Wikipedia article on abortion, I am wondering if there is increased risk of uterine, cervical, or ovarian cancer after several abortions. Also, is it well established that these cancers are caused by hpv viruses? In the 90s, my biology PhD friend told me that many abortions leads to high likelihood of cervical and other cancers. Maybe he was wrong, or the science has changed?(He was in favor of abortion rights, by the way).One of the hypothetical causes I read recently about was abrupt hormonal changes after an abortion.Rich (talk) 16:50, 23 June 2019 (UTC)

HPV vaccine works at reducing cancer risk, but this doesn't mean that all these cancers are from HPV.

Also beware spurious correlation. For instance, having children late or not at all is known to increase breast cancer risk, and obviously women having several abortion are more likely to belong to this group. They may also have a lifestyle that expose them to more STD, including Papillomaviridae. Even if abortion by themselves are not the cause, this may cause correlation between abortion and cancer. Gem fr (talk) 17:37, 23 June 2019 (UTC)

Agreed. They may be having unprotected sex with strangers, and hence need multiple abortions. Thus, the multiple abortions aren't the source of the cancer, but rather the HPV and other diseases that will likely result from unprotected sex with strangers. SinisterLefty (talk) 18:02, 23 June 2019 (UTC)

Note also that anti-abortion activists have a known habit of spreading lies about the health risks of abortion. Someguy1221 (talk) 19:37, 23 June 2019 (UTC)

Spreading lies is usual tool of all activists. Pro-abortion activists have a known habit of doing it no less, starting with Roe Vs WadeGem fr (talk) 18:46, 24 June 2019 (UTC)

What "figures" is he talking about in your fifth item? ←Baseball Bugs ^{What's up, Doc?} carrots→ 01:12, 25 June 2019 (UTC)

He's referring to NARAL's figures for how many women died each year as a result of not having access to safe, legal abortion. Tbh, I'm not sure if he's referring there to deaths from unsafe abortions or death from complications of pregnancy. I believe you could arrive at the figure they give (5-10 thousand per year) by making the naive assumption that all maternal deaths would have been prevented by abortion. Someguy1221 (talk) 02:51, 25 June 2019 (UTC)

If it's the latter, then it's indeed deceptive, although mothers dying after giving birth don't seem to be nearly as frequent as they were a century or more ago. ←Baseball Bugs ^{What's up, Doc?} carrots→ 04:32, 25 June 2019 (UTC)

• @Richard L. Peterson: A search of pubmed for abortion and uterine cancer, and abortion and cervical cancer found no relevant reviews linking abortion to a higher risk of those cancers. A search for abortion and ovarian cancer found one relevant article, which concluded A systematic review of 37 previous studies of the topic confirmed our findings that a history of incomplete pregnancy does not influence a woman's risk of epithelial ovarian cancer. It is hard to prove a negative but these results would suggest that the linking of abortion to increased risk of uterine, cervical, or ovarian cancer is an old-wives tale, or as Someguy1221 notes, intentionally spread misinformation.

PS: A note about my methodology: to weed out older articles, case-reports and weaker studies, I restricted my scrutiny of the pubmed hits to review articles published over the last decade. Given the quantity and pace of medical publishing this is unlikely to have caused me to miss any established links. Abecedare (talk) 20:28, 23 June 2019 (UTC)

Actually, in the case of causes of medical conditions it is often easier to prove a negative than a positive. That is, if no correlation is found between A and B, then A can't cause B. On the other hand, if there is a correlation, that doesn't necessarily mean that A causes B, as discussed above. SinisterLefty (talk) 17:52, 24 June 2019 (UTC)

In medicine, most published research findings are false anyway Gem fr (talk) 18:27, 24 June 2019 (UTC)

'Research findings" includes an enormous amount of exploratory research on small samples. And yes they should be redone with larger tests with careful checks on the design if the results are to be actually applied in real life, that is one reason tests of new medicines are expensive. That is quite a different matter from saying most reviews of a number of better done tests give wrong results. Dmcq (talk) 20:24, 25 June 2019 (UTC)

You seems to believe we can avoid that just redoing test on large sample. Well, no. just like shit, P hacking happens; it could even be more happening with big data and very large sample. For scientific disaster, you just need a very large listing of possible symptoms: just like in a classroom chances are 2 pupils are born the same day, you WILL find statistically highly significant correlations between some symptoms and literally anything, and you WONT be able to disprove it, and boom, you are toasted. You cannot detect which one are false, you just know that most of them are. Gem fr (talk) 21:11, 25 June 2019 (UTC)

Assuming you've actually read Ioannidis, you'd know that he stated this was primarily a problem with preclinical and early-clinical research, and he also discusses statistical methods for detecting and dealing with these issues regardless of where they occur. Using his findings to argue that we can basically ignore any results we don't like, because "most research is false", is simply using the ecological fallacy to avoid having to think. Someguy1221 (talk) 04:30, 26 June 2019 (UTC)

Sure, except your ecological fallacy claim is a straw man fallacy. And your use of "we" is also a fallacy: those who are doing research and lay people must not look the same way at a positive result. Scientists need them as a first step, while lay people turn into tin-folding lunatics when they don't ignore them. Gem fr (talk) 12:56, 26 June 2019 (UTC)

It is only well established that HPV cause almost all cervical cancers. However these is no evidence that HPV cause uterine, breast or ovarian cancer. Ruslik_Zero 20:30, 23 June 2019 (UTC)