User:WBTtheFROG/sandbox/t32q4/TalkAutism

Cutting edge research

(Moving this from my talk page as it's about the Autism article.) Eubulides (talk) 21:43, 16 December 2009 (UTC)

Hi,

I edited the autism page with some comments that are from cutting edge researchers, but, as far as I know, unpublished research, at least, I have not seen it published.

While I respect the need to not put junk in Wikipedia, the comments edited were out of date and somewhat general in themselves. Also, the links used for support of the edited comments were not fully accessible, in that they lead to science journal articles, (generally good) but do not allow non-subscribers to view more than abstracts of those articles,(bad) and the abstracts do not directly support the comments that cite them.

The information I am getting is probably top of the line, cutting edge stuff, that, unlike some things published in Wikipedia, should not be denied the readers just because it's not yet out in the literature yet. Why hobble Wikipedia and it's readers? —Preceding unsigned comment added by 76.232.9.94 (talk) 20:55, 16 December 2009 (UTC)

Unfortunately Wikipedia cannot accept cutting edge, unpublished research; we need reliable published sources to support claims in an encyclopedia. (See the verifiability policy.) It's standard practice on Wikipedia to cite the contents of science journal articles, even when they (unfortunately) are not freely readable. Review articles are best, and for autism there are plenty of recent reviews, so there should not be a problem finding good sources. Please see Wikipedia:Reliable sources (medicine-related articles) for the sort of sources we are looking for here. Eubulides (talk) 21:02, 16 December 2009 (UTC)

There are several things you should know if you don't already: First, there is a tremendous amount of research going on in autism, neurology, and immunology. Comments such as "such and such has not been found" are easily made obsolete by the new research. Also, not all findings are published. The findings I reported were made by the leader of a major reserch study, in a formal presentation to study participants. But, the change to comments was not on some currently controversial topic either. I am not sure which change you are responding to, but if you actually followed research, even published research, in this area you would know that biomarkers for subsets of autistics have been found. Not all the biomarkers I mentioned, but enough to show that the stuff I edited was obsolete. —Preceding unsigned comment added by 76.232.9.94 (talk) 21:18, 16 December 2009 (UTC)

I agree that there's a tremendous amount of research. However, I'm afraid that many of the genetic findings are controversial within the research community as a whole, and in the past a lot of them haven't panned out. (Some of the biomarkers are obviously bogus; some haven't been confirmed, etc.) An encyclopedia article cannot appear to endorse a theory or proposal that has never been published or peer-reviewed, even if the theory is correct; this is a core policy of Wikipedia. The text that's currently in Autism, "Research into causes has been hampered by the inability to identify biologically meaningful subpopulations", is supported by a high-quality 2008 review (Altevogt et al. 2008, PMID 18519493; this is freely readable), and this text is still accurate today, as there is no consensus in the wider research community about what these subpopulations are. The Autism paragraph goes on to say that there is a proposal to classify autism using genetic subpopulations, and gives a specific example of one of these proposed subpopulations, so it's not like the article is ignoring the issue. I'm not saying the article can't be improved: far from it! but improvements must be supported by reliable biomedical sources. Eubulides (talk) 21:43, 16 December 2009 (UTC)

The phrase you reverted to by removing my edit is really misleading. "Research into causes has been hampered by the inability to identify biologically meaningful subpopulations". Volumetric MRI is finding subpopulations of autistics with abnormally large brains, (that was already in the article) but also some with abnormally small brains, (and increased IL10 in the blood) along with brain autoantibodies not found, or in some casex, almost never found, in typically developing children. You are really doing a disservice to readers by keeping this stuff out, especially since you leave in the "no biologically meaningful subpopulations" comment, which is kind of ambiguous when you think of it, because what does "biologically meaningful" mean exactly? I guess I can not stop you but if you actually want to generate good content, you should go to PubMed and look at the latest research and then think about the old phrasing and how it can be made better. We had a similar discussion when I was trying to edit a section about brain autoantibodies in the "Causes of Autism" article some time back, and showed you that the same authors who you relied on to keep reverting my edits had actually found brain autoantibodies in autistics since they published the article that you used as a citation to challenge my changes. It seemed like a huge waste of time frankly, but finally, you let the obvious and non-controversial changes I had made remain. —Preceding unsigned comment added by 76.232.9.94 (talk) 22:08, 16 December 2009 (UTC)

• If you'll recall from the discussion about that earlier change to Causes of autism, the wording you originally installed was not "obvious and non-controversial"; on the contrary, it contained claims like "many researchers report abnormalities of some type of immune function to be far more prevalent in autistics than controls" that presented a one-sided view of a controversial topic, claims that were not supported by a reliable source. After some back and forth, I did the legwork to come up with a reliable review in this area, Stigler et al. 2009 (doi:10.1016/j.rasd.2009.01.007), and as a result a better text was drafted with much better-supported wording such as "Results from immune studies have been contradictory", wording that was eventually installed. No doubt a similar process could occur here, but the first step is to find a reliable medical source: we simply cannot put stuff into a Wikipedia article based on an anonymous editor's recollection of an unpublished presentation.
• One part of the change you just now installed into Autism has been kept, in a reworded form, as the text no longer claims that all autistic children have faster-than-usual brain growth just after birth. (There is a pattern of faster-than-usual growth, but it's not known whether this is universal; note that this is different from the issue of brain size.) The parts that were not kept, because we don't have reliable sources, are:

• Reduced blood levels of melatonin is a notable neurotransmitter abnormality in autism.
• The claim that we are now able to identify biologically meaningful subpopulations of autism.
• The following text:

"However, recent research has indicated many distinct biomarkers for autism not found in healthy control subjects, including greater brain volume, lesser brain volume, abnormal brain response to stimuli, abnormal cytokine profiles, and abnormal brain autoantibodies in subsets of autistics."

The issues of greater brain volume, abnormal brain response to stimuli, abnormal brain autoantibodies are already covered by the article, in the Other symptoms and Pathophysiology sections, and it's not clear why they need to be repeated. The issues of melatonin and cytokine profiles would also be good to cover, along with lots of other biomarker hypotheses being investigated (such as oxidative stress, glutamate, GABA, dopamine, lymphocytes, etc). But this is not something that a Wikipedia editor should research! We should be relying on reliable sources on this topic, such as we found for Causes of autism.

Eubulides (talk) 00:45, 17 December 2009 (UTC)

A paper on reduced melatonin in blood of some autistics was published by Thomas Bergeron of the Pasteur Institute. It's on Pubmed. As to the increased brain volume, I am only saying that it's not good writing to mention increased brain volume as though this is a universal finding when some researchers doing volumetric MRI, while confirming increased brain volume in some autistics, have found decreased brain volume in others. Well, I am sorry if I got snippy but I feel very strongly that parents of autistics should understand there ARE findings of biologically meaningful subpopulations. If enough of them come to understand that, It think they will make specific demands of the political/research funding process, demands that can only help, ie, find out why some brains are small, some big. But besides the importance to the parents and society, I just don't think the article as now written is accurate, as in the my last disputed edit, because the research has made this statement, made only a year ago now pretty clearly obsolete. I guess you could say it's partly a matter of semantics, maybe the authors of that review would say, "biologically relevant" does not apply to findings of different brain size, if you can't tell us what is means, but I think most people would not agree with that as written. Again, at the very least it's misleading. A very large number of researchers are talking now about the heterogeneity of autism, in other words, they already believe very strongly that the etiologies are different in different autistics because they find abnormalities in some but not all. Brain size just happens to be one of those things that jump out at you, at least to most people, that it's at least very likely, when the brain sizes of autistics are found outside the norms, that this is "biologically meaningful" Well, I've got no more time for it, but this autism article is just plain terrible as written, it's so misleading in so many ways, so out of line with current research, that it should not even be in the dictionary, it's doing more harm than good. At the very least, it should contain dozens of caveats, ("brain size is increased in SOME autistics" is just one example. I'm not going to fight it all out, but the readers are really getting a bad deal here. —Preceding unsigned comment added by 76.232.9.94 (talk) 06:40, 17 December 2009 (UTC)

And, not to beat it to death, but to me, the sense of the article from which the quote came puts it in a different light altogether. Here is the entire paragraph, and, by the way, the review article does not really seem to be a literature-wide review of published studies, it's more the notes of a symposium held by this organization. But, anyway, here is the quote:

"The diversity of symptoms and symptom clusters attributed to autism emphasizes that there is not just 1 type of autism but rather a collection of different variants, each with its own set of characteristics and, presumably, etiologies. The inability to identify and stratify biologically meaningful subpopulations has limited research into the causes of autism. Thus, several workshop participants highlighted the need to develop autism phenotypes that may be used to identify clinically meaningful subtypes, looking at onset, clinical presentations, the associated symptoms, and the clinical course. Participants identified several potential strategies that could be used to investigate potential subtypes, including but not limited to analysis of genetic data, analysis of biological samples (eg, blood, urine), metabolic profiling, and stratification and analysis on the basis of comorbidities. ""

To me, the entire paragraph, showing as it does the widespread belief of many different etiologies for the condition, and the subsequent findings of very clear abnormalities in some subsets of autistics, turns the quote made from the article completely on it's head. However it's to be done, I think this should be incorporated in the article, so that at the least people know the thinking of a large number of researchers. —Preceding unsigned comment added by 76.232.9.94 (talk) 06:59, 17 December 2009 (UTC)

And, last point, the rewording in Causes of Autism to "Results from immune studies have been contradictory" is terribly misleading. I can give you many studies showing immune system abnormalities that have not been refuted at all. Not in peer reviewed journals, not in any journals covered by PubMed at all. So really, there was nothing controversial about the statement of immune system abnormalities in some autistics. If someone disputes the findings of immune system abnormalities, and get's their dispute published, then there will be controversy, and, unlike the immune system studies, you have that with the identical twin studies which are quoted without any caveats, even though peer-reviewed journals have had articles disputing the meaning of the high identical twin concordance rate, ie, that in utero effects may explain a lot of it, rather than purely genetics. —Preceding unsigned comment added by 76.232.9.94 (talk) 07:12, 17 December 2009 (UTC)

• The wording in Causes of autism should be discussed in Talk:Causes of autism.
• I'm aware of the Institut Pasteur paper on melatonin and ASD (Melke et al. 2008, PMID 17505466), and it's very interesting work that is part of an active research area, but it is a primary source and we need reliable reviews to cover research like this (see WP:MEDRS). Bourgeron's own 2009 review of the area (PMID 19545994) says only "other modulators such as serotonin and/or melatonin may play crucial roles in the onset and severity of ASD" which isn't much to go on vis a vis melatonin. Melke's own review has not, as far as I know, been published yet, so we can't cite it; and at any rate it'd be better to use someone else's review of the work of the Melke–Bourgeron–etc group. (The most recent high-quality review of autism that I know of, Levy et al. 2009 (PMID 19819542), does not mention melatonin at all, which suggests that melatonin's significance may not seem as important to researchers outside that group.)
• I am unfamiliar with reviews covering the topic of decreased brain volume associated with autism; the reviews I've seen (e.g., Casanova 2007, PMID 17919128) talk only about increased brain volume. If you could mention some sources that would be helpful.
• The heterogeneity of autism is discussed in Autism, which contains statements such as "there is increasing suspicion that autism is instead a complex disorder whose core aspects have distinct causes that often co-occur" and "autism does not have a clear unifying mechanism at either the molecular, cellular, or systems level; it is not known whether autism is a few disorders caused by mutations converging on a few common molecular pathways, or is (like intellectual disability) a large set of disorders with diverse mechanisms". As far as I can see, this pretty much covers everything that's in that symposium quote about causes, though no doubt it could be worded better.
• The key notion here is still sources. We need good sources to make claims about biomarkers or melatonin or whatever. Autism has lots of primary sources, covering all sorts of things (melatonin is just one small corner); we can't go picking and choosing among the 20,000 or so primary studies that have been published on the subject, and we have to rely on recent reliable reviews to help us report what's significant and skip what's not.

Eubulides (talk) 08:00, 17 December 2009 (UTC)

Although the link you've given me says review articles in medical journals are considered very good sources, it does not seem to forbid citations to primary research published in medical journals, and in fact, you find those citations all over wikipedia. And yet, you've taken away the mention of the low melatonin findings of the folks at Pasteur, even though you know about them yourself. Does that make sense? I am not saying they found something of great importance, but it is something that has not been refuted, so it seems to me it should be in the article. Really, the decision what to put in there, and what to take out, seems quite arbitrary on your part. Recent findings no good, only reviews good? That ensures the article will not be up to date, in other words ensuring it will be inadequate. And you say we should use "recent" reviews, but you have citations to reviews from 2007. A very long time ago. With several papers on abnormalities in autistics nervous, immune and metabolic systems, and therefore possible etiology, coming out every week at this point in time, a review from 2007 is probably about 1,000 peer-reviewed papers behind the state of research. Instead of debating your application of policy, which I really think is biased towards removing solid information, I will point out just one of the many flaws in the article: First, inadequate emphasis on the heterogeneity of autism, for example, difference between an autistic who is clearly intellectually above average and one who is clearly mentally retarded. I am not saying the article does not mention this, but it's all so poorly written that it gets lost. The stilted over-use of jargon is most of the problem there. Well, I really don't have time for this. I think you are probably doing more to ensure the public is uninformed or misinformed on this topic than all but some small number of out and out charlatans and crackpots. You keep out solid new information, while keeping in old statements proven false. This is exactly what you did before in regards to auto-antibodies, insisting on keeping in a quote that said there were no definite findings of abnormal autoantibodies, from an article by the folks at UC Davis Med School, when the up to date information showed those same people had since discovered and published findings of autoantibody abnormalities. Please reconsider the way your are applying the policy. At the least, findings from Institut Pasteur could be mentioned as, "one group at the Pasteur Institute published findings of abnormally low levels of melatonin". That is an objective fact, and everyone has been warned it's not necessarily universally accepted by the "one group" wording. Using that wording, a lot of stuff could go in without being misleading. Well, that's all I've got, but I will tell you, you are missing a huge amount of good information from me that could be attributed to "unpublished studies" or such. —Preceding unsigned comment added by 76.232.9.94 (talk) 13:55, 17 December 2009 (UTC)

I surely do not want to open up a can of worms but I do find that the page could do with more, let's call it "less confirmed" or "less rigorous" research. Calling it "cutting edge" or "new" imparts a misleading sense of it being better/more modern than what's gone before. Calling it fringe is unfair in the sense that something that is fringe, with NO published research shouldn't be included in this category. I don't contend it's an easy line to walk, keeping the page from being cluttered with secretin, dolphin play, music therapy but at the same time, more information on newer published research is valuable. People can make their own decisions on the value/investigate further. The advantage wikipedia has over a print encyclopedia is the ability to integrate new information very quickly. I think there can be a strongly worded header that what follows are new directions that while published, are not confirmed yet by the larger scientific community. If a study is particularly poor with say 6 patients, with a biased selection criteria and written in an off-axis journal, I have no problem with leaving that off the page because it's of little value. But I do think there are some other areas that can be covered that aren't being covered. HarmonicSeries (talk) 14:23, 17 December 2009 (UTC)

Please see WP:MEDRS, WP:UNDUE and WP:RECENTISM; Wikipedia is not the place for publishing original research or "less confirmed" or "less rigorous" research, nor the place for "integrat[ing] new information very quickly" if that info is "Recentism" (Wiki is not a newspaper), and a feautued medical article is based on giving due weight to published secondary reviews. When research is confirmed and finds its way into secondary reviews, it can be included on Wiki; what other Wiki articles (that are not featured and haven't been rigorously reviewed) do or don't do isn't relevant. Please see WP:OTHERSTUFFEXISTS. Unless the original poster has strong sources supporting proposed content, this talk page shouldn't be used as a WP:SOAPBOX. Causes can be explored in the Causes sub-article. SandyGeorgia (Talk) 14:38, 17 December 2009 (UTC)

Just to be clear, I am not proposing that all new stuff be let in, rather that old stuff that is clearly out of date should be taken out when it can be shown to be out of date, and that new stuff should be let in under less strict criteria than Eubulides is using. The out of date stuff I am referring to were things like the "no autoantibodies found in autism" which was clearly untrue at the time I got in a dispute over it. Or the findings on low melatonin, which Eubulides kept out because he said it had not been in a literature review. Although there is a guideline stating that review articles are the preferred source of medical information, I have seen nothing prohibiting citation of any journal article reporting a study's findings, and it seems to me that excluding all study findings on the basis they are not included in review articles is really a bad idea in a field as active as autism. As I pointed out above, review articles from 2007 are about 1,000 papers behind the latest research. I hope people will reconsider their application of rules which require judgment calls, as this one does. I am not on a soapbox except to get accurate information out, which means removing old no longer accurate statments. If I went too far with reporting unpublished information, sorry about but as for mentioning research without citations for everything, the fact is the article does that repeatedly as it stands, and the editor knew about the finding of low melatonin independently, so, why remove it, unless you are going to remove all uncited claims, even when you know they are true? Could be done but seems very silly. —Preceding unsigned comment added by 76.232.9.94 (talk) 15:35, 17 December 2009 (UTC)

Please sign your entries by typing four tildes (~~~~) after your posts. Adding every bit of recent published info from primary sources would amount to WP:UNDUE and WP:RECENTISM; I concur with Eubulides. SandyGeorgia (Talk) 15:40, 17 December 2009 (UTC)

I went to the links you gave on recentism and undue and I do not agree with your statement. I do not agree that citations to articles in peer-reviewed science journals should be disallowed under the rules as given. Is that what you are claiming? Please clarify. I see countless such citations in Wikipedia.76.232.9.94 (talk) 15:55, 17 December 2009 (UTC)

I think that constructive comments are trying to be made here. There is nothing in wiki guidance that says new peer-reviewed research can not be included without first being reconfirmed by later studies. None of what has been said here is inconsistent with WP:UNDUE and WP:RECENTISM either. More collaboration among editors is surely a good thing. HarmonicSeries (talk) 15:57, 17 December 2009 (UTC)

Once again, please see WP:OTHERSTUFFEXISTS and WP:MEDRS; what may or may not be in any other Wiki article (on the website "anyone can edit") isn't relevant to a thoroughly reviewed and vetted featured article. Imagine if we included every piece of primary research in every article: the article would become a lengthy, unreliable, example of newspaper recentism. Due weight is given to high-quality sources and research over unvetted Recentism. It might be helpful to reread Eubulides response above, which does focus the discussion on sources as they are applied on Wiki. SandyGeorgia (Talk) 16:14, 17 December 2009 (UTC)

It is not helpful to continually suggest that the addition of anything new is recentism or the tip of the iceberg to a horrible page. Your opinion is that the hurdle for inclusion is very high. Other editors have a different but equally valid opinion that the hurdle could be lower and still consistent with wiki guidance and those editors believe that a better page would result. You do not believe that. We understand that. HarmonicSeries (talk) 16:23, 17 December 2009 (UTC)

I agree that constructive comments are being made here; the article has already been improved a bit as a result, and I hope for more improvements. Still, we have to remember that the article is already full already, and it does not contain nearly enough room to talk about all the lines of research into causes, pathophysiology, treatment, etc. If we're going to talk about melatonin, how can we justify excluding naltrexone, carnitine, acetylcholinesterase inhibitors, oxytocin, tetrahydrobiopterin, alpha-2 adrenergic agonists, tetrahydrobiopterin, music therapy, hyperbaric oxygen, etc., etc.? All these treatments have their adherents and have some promising evidence published as primary studies and I wouldn't call any of them purely WP:FRINGE; but none are mainstream by any reasonable measure. There simply isn't space here to talk about them all; the article's already too big and really needs to be put on a diet (it takes more than 30 seconds for me to load it!). I suggest discussing more-cutting-edge therapies in Autism therapies, which has more room for it. And, above all, we need good sources. Primary studies are not good sources for autism: as I've said before, there are tens of thousands of published and peer-reviewed autism papers, and we should not be placed into the position of deciding which 1% to cite and which 99% to omit. That's the job of qualified expert medical reviewers. This field has no shortage of high-quality reviews! We should be using them, not redoing them. Eubulides (talk) 18:17, 17 December 2009 (UTC)

Just like to add my voice to support Eubulides and SandyGeorgia wrt the importance of using reviews. As Eubulides points out, we have no shortage of them in this important topic area and his primary studies point about "deciding which 1% to cite and which 99% to omit" should probably be framed as a banner in WP:MEDRS. It is healthy to point out when some of our sources are getting out-of-date, note which facts may now be questioned and to discuss potential facts for inclusion. But we must be patient when it comes to writing encyclopaedic material. If editors are frustrated that what they regard as important and significant research isn't being covered by recent medical reviews, it is not Wikipedia's job to right that wrong. As Eubulides says, adding new material to daughter articles may be easier as they have more room. Then the facts can bubble up and be condensed for this parent article. Colin°^Talk 20:01, 17 December 2009 (UTC)

I want to be clear, I only mentioned a low level of melatonin because a high level of seratonin is mentioned in the article, and melatonin is synthesized from seratonin in the body. So, it seemed logical to mention this in that context, not as a therapy of some sort. The low melatonin and high seratonin may be related, in that if less serotoin is coverted into melatonin there will be exactly the situation that is widely report, high serotonin, low melatonin. Got to go now, but please someone explain to me why research papers in peer-reviewed articles, especially the findings of a researcher at the Pasteur Institute, are not good enough. I just can not see this. —Preceding unsigned comment added by 76.232.9.94 (talk) 00:42, 18 December 2009 (UTC)

It's hard to know what else to offer you, when Colin, Eubulides and I have all explained in several ways and several times: I'm at a loss for what else we can add to help you understand. SandyGeorgia (Talk) 01:24, 18 December 2009 (UTC)

Well, I will respond in what I hope is a constructive way. I think primary research papers are kind of open to abuse, because someone may have a minority viewpoint they want to push, but I think this should not mean all primary research is banned. For one thing, I think the problem of "deciding which 1% to include and 99% to omit" is the traditional job of an editor. But, if you don't have that paradigm, then at the least, under the theory of user-contribted content, shouldn't it be the case that studies are not deleted if they are at least cited to peer-reviewed articles? Because, to at least some extent, deletions are being made, at least as I understand it, in other words, someone is making these decisions on a case by case basis. Even the concept of only using "review" articles is a bit circular, after all, the reviewe are made of the original papers by searching pubmed.gov.

In my view, the true task, which requires substantial understanding of the underlying content, is to first separate wheat from chaff, then separate high-quality wheat from low-quality wheat. Not to beat the melatonin thing to death, but if you look at the most respected and cited institutions, people, and journals I think you will find Pasteur Institute is near the top. Now, that would not mean much, if for example the paper was on say one "autistic" patient with some extremely rare genetic defect that probably had no implications for most other austistics. Not meaningful. But if you've got a finding of low melatonin in a very large percentage of the autistic population, then at the very least, parents who want to give their autistics kids melatonin, not for autism, but just to make them sleep, (half the kids have sleep problems according to another part of the same article)would have some indication it's likely, in most cases, to only be bringing them back to a typically developing child's level. In other words, with just a little background in the subject, the melatonin information becomes of much greater interest to the laymen/parent readers of the article than many other biomarker findings. So, my conclusion is, it's just not possible to edit effectively without all the background. Perhaps my change to include the melatonin findings did not explain enough background. My fault for that. —Preceding unsigned comment added by 76.232.9.94 (talk) 03:06, 18 December 2009 (UTC)

We're not saying primary research is bad source material in general but it is bad for Wikipedia. This is wholly due to our "anyone can edit" (anonymously) policy. If Wikipedia's Autism article was written by a single or group of well-regarded experts, then I'm sure they would cite primary research papers much as they do in review articles. The job of deciding whether a primary study was done (a) by a well regarded group (b) to a good scientific standard (c) is actually clinically relevant or just a stepping stone for further study and (d) is notable when compared to all the other research ... is something that requires an expert opinion. Wikipedia doesn't trust its editors to be experts even if they are, so we need instead to cite experts' literature reviews. And if we cite their material in high-quality journals, we can be sure those experts have been vetted and their material checked. Our policies do limit at times what Wikipedia can say or how up-to-date is can be. But if you are prepared to work within those restrictions, then some pretty decent content can be produced.

Wrt melatonin, we can only indicate that this might be useful to some children if we can cite reviews or clinical guidelines that make such a recommendation. Just because melatonin levels are low in some children doesn't automatically mean that orally consuming the stuff will help or be safe. Colin°^Talk 09:20, 18 December 2009 (UTC)

Also please see #Levy et al. below. Eubulides (talk) 00:02, 19 December 2009 (UTC)

Levy et al.

I went through Levy et al. 2009 (PMID 19819542), a fine review of autism published last month in The Lancet, and used it to update and modernize the article a bit. Among other things, this change introduces a brief discussion of melatonin-related issues, as follows:

"The relationship of neurochemicals to autism is not well understood; several have been investigated, with the most evidence for the role of serotonin and of genetic differences in its transport."

This is as close as Levy et al. get to talking about melatonin. I hope it addresses part of the comments above, at least the parts about being based on older reviews and on ignoring the possible role of melatonin in autism pathophysiology.

Some responses to previous comments:

• "shouldn't it be the case that studies are not deleted if they are at least cited to peer-reviewed articles" No, because there are 20,000 peer-reviewed articles on autism, and the article cannot possibly contain material that cites them all. In an area where little work is done, and where no reviews are available, it may be necessary to cite a primary study with all due caution. But in autism there are reviews galore, and there is no reason for us to be citing primary studies ourselves. We should rely on reviews.
• "Even the concept of only using "review" articles is a bit circular" No, because the reviews are written by acknowledged experts in the field, and are themselves peer-reviewed. In contrast, Wikipedia editors are anonymous, with no known qualifications. The only quality control we have, really, is our sources. If we refuse to rely on reviews, we'll have an article in which editors are free to promote all the latest fads: not just melatonin, but also hyperbaric oxygen and the rest. And there simply is not enough space in this article for all that stuff.
• "if you look at the most respected and cited institutions, people, and journals I think you will find Pasteur Institute is near the top" Yes, they're quite well-respected, but it's just one research group at one place. It's standard practice in science for people to replicate promising results before relying on them; in medicine, the results need to be not only replicated, but its effectiveness and safety have to be tested clinically. Besides, even if we limited ourselves to "top" institutions, that would cut the number of primary studies that we could cite down only to 5,000 or 10,000, hardly enough to help us decide which of those studies to cite.
• "parents who want to give their autistics kids melatonin, not for autism, but just to make them sleep, (half the kids have sleep problems according to another part of the same article)would have some indication it's likely, in most cases, to only be bringing them back to a typically developing child's level" This is exactly the sort of original research that Wikipedia articles cannot do. They can't even imply this sort of thing, not without a reliable source. See Wikipedia:No original research.

Eubulides (talk) 00:02, 19 December 2009 (UTC)

I have to ask, do you see 20,000 citations? If not, then maybe it's not a big problem. Obviously we've beat this to death but I would ask, what are the rules for removing other people's edits exactly? —Preceding unsigned comment added by 76.232.9.94 (talk) 03:30, 19 December 2009 (UTC)

We could easily see thousands of citations if every citation ever added to Autism had to be kept. Perhaps you're thinking about the Wikipedia:Consensus rule? Eubulides (talk) 03:38, 19 December 2009 (UTC)

I don't know the Wikipedia consense rule, I am asking. Though I've seen the rule favoring review articles, I've seen nothing forbidding citations to original research articles in journals, and in fact, I see those kinds of citations all over Wikipedia, countless times. So, what are the rules? —Preceding unsigned comment added by 76.232.9.94 (talk) 06:30, 19 December 2009 (UTC)

I do want to point out one thing: the claim made here, "primary sources can not be used" is false according to the link given above. In fact, the link gives extensive guidance on how primary sources may be used. Since the information in the link seems to be routinely misrepresented, I will go and get the relevant sections from the link. —Preceding unsigned comment added by 76.232.9.94 (talk) 06:46, 19 December 2009 (UTC)

So, here is one relevant section from the guidlines on medical articles "Reliable primary sources can add greatly to an article, but must be used with care because of the potential for misuse. For that reason, edits that rely on primary sources should only describe the conclusions of the source, and should describe these findings clearly so the edit can be checked by editors with no specialist knowledge. In particular, this description should follow closely to the interpretation of the data given by the authors, or by other reliable secondary sources. Primary sources should not be cited in support of a conclusion that is not clearly made by the authors or by reliable secondary sources, as defined above".

In other words, you can use a primary source but not as though it was a review article secondary source. But you can use it to say "Dr. Bourgeron found abnormally low melatonin in the blood of 55% of autistics". This is citing the source as a primary source, perfectly allowed under this section, particularly if it is not presented as a consensus belief among researchers, or given undue weight in some other way. I am upset that I was told primary sourced are diallowed when that does not seem to be the rule at all. Please someone, tell me the basis for the statement primary sources are not allowed? 76.232.9.94 (talk) 06:53, 19 December 2009 (UTC)

I am quoting another section of the rules on primary sources which clearly shows they are not forbidden if presented properly:

"Individual primary sources should not be cited or juxtaposed so as to "debunk" or contradict the conclusions of reliable secondary sources, unless the primary source itself directly makes such a claim (see Wikipedia:No original synthesis that advances a position). Controversies or areas of uncertainty in medicine should be illustrated with reliable secondary sources describing the varying viewpoints. The use and presentation of primary sources should also respect Wikipedia's policies on undue weight; that is, primary sources favoring a minority opinion should not be aggregated or presented devoid of context in such a way as to undermine proportionate representation of expert opinion in a field.

If an important scientific result is so new that no reliable reviews have been published on it, it may be helpful to cite the primary source that reported the result. Although popular-press news articles and press releases often tout the latest phase II clinical trial, such trials are rarely important enough to mention in an encyclopedia. Any such results should be described as being from a single study, for example:

"A 2009 U.S. study found the average age of formal autism spectrum diagnosis was 5.7 years." (citing PMID 19318992) After enough time has passed for a review to be published in the area, the review should be cited in preference to the primary study. If no review is published in a reasonable amount of time, the primary source should be removed as not reporting an important result. When in doubt, omit mention of the primary study, as per WP:RECENTISM.

So, this section calls for some judgment calls on how long is long enough, but what it does not say is that one review that does not mention a topic is adequate reason to delete the primary study reference. On the contrary, the relevant phrase is "If NO review is published in a reasonable amount of time". So, you've got to wait a "reasonable amount of time" at the very least, before deleting any primary source reference. In other words, the very latest journal papers can be cited as long as it is done correctly.

It seems to me you've badly misunderstood and applied these rules quoted above. The right thing to do would be to check all new postings to see if they comply with these rules given above, and only remove them if they don't, or if enough time has passed and no review articles mention them. Note to Eubulides, not if the Lancet review article fails to mention them, that is not good enough, it has to be that NO review articles mention the finding in a "reasonable amount of time". So, you had it exactly the opposite of the rule. New is not forbidden, in fact, it's implicitly protected until it's been proven to be not regarded as important by consensus, which takes time to develop.

Please don't delete my postings again under the claim that new is forbidden. I will try very hard to obey the rules as given in the section, and any others you relay to me, but you must obey them as well please. 76.232.9.94 (talk) 07:18, 19 December 2009 (UTC)

Nobody said primary sources are not allowed. What's not allowed is to use primary sources to dispute secondary sources, or to promote primary sources out of proportion to the prominence that secondary sources give them; see WP:PSTS. After this thread began I checked a recent secondary source on this topic, namely Levy et al. 2009 (PMID 19819542), published in The Lancet last month. Here's what Levy et al. says about neurochemicals in general (and this includes melatonin):

"Neurochemical investigations with animal models and empirical drug studies remain inconclusive. Serotonin and genetic differences in serotonin transport seem to have the most empirical evidence for a role in autism spectrum disorder, whereas data lending support to the roles of dopaminergic and glutaminergic systems are presently less robust, but are evolving. Study of the role of the dopaminergic and cholinergic system, oxytocin, and aminoacid neurotransmitters shows promise."

In other words, we simply don't know what roles neurochemicals play in autism: they may have causative roles in the disorder, or they may just be largely-irrelevant side effects, or they may even be artifacts of the primary studies in question. In light of this review it would be greatly misleading and a violation of WP:WEIGHT for Autism to trumpet one particular melatonin primary study over all the other work done in this area. (By the way, I urge you to read Wikipedia:Consensus, if you haven't already, and reflect about what Wikipedia:Consensus would say about this very thread.) Eubulides (talk) 07:25, 19 December 2009 (UTC)

WP:MEDRS offers guidelines that apply through the whole spectrum of medical issues, from drugs to rare diseases to common disorders to minor ailments. The sort of sources one uses for one topic will vary from those used in another. For example, an article that describes an extremely rare disease may be forced to cite a few case studies. A topic that is as well studied and as comprehensively and frequently reviewed as autism has little need to go scrabbling around in the primary research cupboard to find source material. The MEDRS example was for the situation where a major important study had just been published, and for which it would be expected for Wikipedia to have content. Material that is too new to have been reviewed yet has to be exceptionally important justify including. BTW: the editor you accuse of having "badly misunderstood and applied these rules" played a primary role in drafting those very words. Colin°^Talk 12:30, 19 December 2009 (UTC)

This is getting tiring and WP:TLDR. IP76, I agree with the suggestion that you should carefully read and understand WP:CONSENSUS, and if you disagree with WP:MEDRS, the place for that discussion is WT:MEDRS, where possibly more editors will help you understand the problems with the position you are taking here. I have seen no indication that melatonin should be included in this article, although based on the research I'm aware of, it likely does have a place in some sleep disturbance article. SandyGeorgia (Talk) 14:05, 19 December 2009 (UTC)

I agree, it is tiring. Please, can someone refer me to someone who has authority to resolve such disputes? Because I really can't see how I am wrong, I feel I've quoted the guidelines extensively and they seem to very widely support my viewpoint, regardless of who wrote them. 76.232.9.94 (talk) 14:37, 19 December 2009 (UTC)

Ok, so I read up on consensus, and still think I am not in the wrong to criticize the deletion of primary material when that primary material was presented properly. But, I also see the mention of the value of discussion. I will put out some things here for discussion as to inclusion:

1. The article's wording seems to be very unclear as to expressing the fact that there are many different causes which will lead to the diagnosis of autism. I believe this should be more clearly reworded. I say it's a "fact" because one can find countless papers indicating separate genetic abnormalities, Fragile X and much rarer problems, that have been firmly tied to autism, and yet, one can not find them in the majority of autistics. I do not believe the current wording is wrong, just confusing. I also think it should be made clear that the phenotypes vary so greatly that there is not much commonality, necessarily, between two people diagnosed as "autistic". In other words, the fact that the diagnosis is very flawed, and there is controversy over even using it, should be brought to the fore. This is not really a controversial move if you look at the literature.

2. Primary papers from Kenney Krieger, MIND, and Johns Hopkins, and for that matter, many other sources indicate autoimmune or inflammatory abnormalities in some autistics. This should be mentioned and made clear, while avoiding undue weight of course. There was a study by researchers at Johns Hopkins a few years back indicating long term inflammation in brains of autistics who had undergone regression. I believe this is kind of a landmark study in this regard, and should be mentioned with all the caveats given primary research, thought it's hard for me to believe it has not been mentioned in review articles, (don't have easy access to full articles from my home computer, get them from time to time by going to University bio library, but can't always get there)

3. The identical twin studies are almost always mentioned to show very high genetic determination in autism. I do not dispute this, but think that a careful edit should be made to indicate limitations of twin studies. For example, that most of the twins are not separated at birth, they have grown up in the same enviroments, and had the same in utero blood supply, possibly same exposure to maternal immune system, same passive immunity from the mother, etc. This need not be in great detail, but quoting "90% genetic" without explanation could be misleading.

4. Related to point 3 above, there should be more discussion of controversy over increase in rate of autism, (implying environmental cause) vs. "90% genetics" implying just the opposite. Some short indications of the difficulties, in the real world, of separating genetic from environmental factors is appropriate, again without undue weight and without too much time taken up on it. I don't think all the current wording is bad, just that it's a little bit imbalanced towards the genetics without explaining the other side.

5. I don't have information on this handy, but as far as causes are concerned, I do not believe it is wrong or misleading to research the way current studies are being done, particularly large-scale longitudinal studies, because these give implicitly an indication of what actual researchers are thinking will be productive. Obviously, this must be done carefully, but it can be done and not in any prohibited way. In any case, many people are very interested in knowing where research is going. Kids waking up and stopping me, but is there any objection to these changes generally? —Preceding unsigned comment added by 76.232.9.94 (talk) 15:37, 19 December 2009 (UTC)

Responding to your first question: there is no one person or group in charge. If you want to discuss a medical issue, with editors who are knowledgeable and have access to journal sources, then I suggest a post to the medical wikiproject. If you think our guidelines are wrong or aren't clear or helpful then Sandy's suggestion of WT:MEDRS could be used, but that page is not for discussing specific article content disputes. There are other suggestions at Wikipedia:Dispute resolution. It would really help if you could edit your posts for brevity. Also, it will help any third-parties if you could outline clearly the article text you wish to add or change and the sources you are using that back up that change. You raise lots of points. Perhaps we could deal with them one at a time. Could you start a new section with one issue. In addition to the general points you make above, could you suggest what actual changes to the text you want to see and what sources those changes might use? Colin°^Talk 15:54, 19 December 2009 (UTC)

I'd like to second Colin's suggestion to make specific proposals for changes to article text. I share 76.232.9.94's concern that the article is a bit dated and not worded as clearly as it could be, and hope that this discussion can continue to improve the article, as it has already improved it. Please bear in mind, though, that we're pretty ruthless in checking sources and making sure that the text agrees with medical consensus, not only about the topic material, but about how much weight to give each topic. Proposals that cite recent reviews are far more likely to work, but even if you can't supply sources I'm willing to go check what recent reviews say, as I have access to many (though not all) medical journals. (Please don't expect immediate turnaround on this; we're all busy.) Eubulides (talk) 18:45, 19 December 2009 (UTC)

I just wanted to say, I have not reread the entire article, but I think the latest edits are huge improvements over what was present before, it's much clearer I think and much less likely to mislead anyone. Thank you for the rewrites. 76.232.9.94 (talk) 20:05, 20 December 2009 (UTC)

Cutting edge research

Hi, I added a sentence about recently found biomarkers for autism, including abnormal cytokine profiles, reduced melatonin in blood, and abnormal brain autoantibodies. These are from top-of the line, cutting edge researchers. –ZY

Unfortunately Wikipedia cannot accept cutting edge research even from top-line labs; we need reliable published sources to support claims in an encyclopedia. (See the the verifiability policy.) We need review articles, especially by people outside that research group. For autism there should not be a problem finding good sources, such as the one that states “Research into causes has been hampered by the inability to identify biologically meaningful subpopulations.” –EB

What I edited was obsolete, and what I put in has not yet been refuted. Biomarkers for subsets of autistics, even if not all the ones I listed, have been published in recent research. Citations directly to medical journals are common on Wikipedia; relying only on reviews ensure the article will always be out of date. You are really doing a disservice to readers (especially parents of autistic kids) by keeping this stuff out. –ZY

Judging the methods and usefullness of a primary study requires an expert opinion. Wikipedia doesn't trust its editors to be experts even if they are, so we need instead to cite experts' literature reviews. –EB

I still think primary material should be includable if presented properly. I will put out some things here for discussion as to inclusion, in a few days. -ZY

OK. I agree that the article is a bit dated but keep in mind that we’re pretty ruthless in checking sources. –EB

Cutting edge research

Biomarkers for subsets of autistics have recently been published by some top-of the line, cutting edge researchers at the Pasteur Institute. Shouldn't those be noted here? –ZY

Unfortunately Wikipedia cannot accept cutting edge research even from top-line labs. Instead, we need review articles, especially by people outside that research group. For autism there should not be a problem finding good sources, such as the one that states “Research into causes has been hampered by the inability to identify biologically meaningful subpopulations.” –EB

That source is obsolete, and the new markers have not yet been refuted. Relying only on reviews ensure the article will always be out of date. You are really doing a disservice to readers (especially parents of autistic kids) by keeping this stuff out. –ZY

Judging the methods and usefullness of a primary study requires an expert opinion. Wikipedia doesn't trust its editors to be experts even if they are, so we need instead to cite experts' literature reviews. –EB

I still think primary material should be includable if presented properly. I will put out some things here for discussion as to inclusion, in a few days. -ZY

OK. I agree that the article is a bit dated but keep in mind that we’re pretty ruthless in checking sources. –EB

I've made some changes that add in the research I found. -ZY

I removed some of your changes that aren't supported by verifiable reviews, and helped clean up around some of your other changes, which were good. –EB

I have not reread the entire article, but I think the latest edits are huge improvements over what was present before, it's much clearer I think and much less likely to mislead anyone. -ZY

Cutting edge research

This article is inaccurate. Biomarkers for subsets of autistics have recently been published by some top-of the line, cutting edge researchers at the Pasteur Institute. Shouldn't those be noted here? –ZY

Unfortunately Wikipedia cannot accept cutting edge research even from top-line labs. Instead, we need review articles, especially by people outside that research group. For autism there should not be a problem finding good sources, such as the one that states “Research into causes has been hampered by the inability to identify biologically meaningful subpopulations.” –EB

That source is obsolete, and the new markers have not yet been refuted. Relying only on reviews ensure the article will always be out of date. You are really doing a disservice to readers (especially parents of autistic kids) by keeping this stuff out. –ZY

Judging the methods and usefullness of a primary study requires an expert opinion. Wikipedia doesn't trust its editors to be experts even if they are, so we need instead to cite experts' literature reviews. –EB

I still think primary material should be includable if presented properly. I will put out some things here for discussion as to inclusion, in a few days, because this article definitely needs correction. -ZY

OK. I agree that the article is a bit dated but keep in mind that we’re pretty ruthless in checking sources. –EB

I've made some changes that add in the research I found. -ZY

I removed some of your changes that aren't supported by verifiable reviews, and helped clean up around some of your other changes, which were good. –EB

I have not reread the entire article, but I think the latest edits are huge improvements over what was present before, it's much clearer I think and much less likely to mislead anyone. -ZY

Cutting edge research

I just don't think the article as now written is accurate, because new cutting edge research has found some biologically meaningful subpopulations and markers; the statement that they don’t exist is obsolete. I've got no more time for it, but this autism article is just plain terrible as written, it's so misleading in so many ways, so out of line with current research, that it should not even be in the encyclopedia, it's doing more harm than good. I'm not going to fight it all out, but the readers are really getting a bad deal here. -ZY

I am unfamiliar with reviews covering the specific markers you mention. We need good sources to make claims about biomarkers or melatonin or whatever. There are a lot of primary studies on autism, covering all sorts of things (serotonin and melatonin and brain volume among others); we can't go picking and choosing among the 20,000 or so primary studies that have been published on the subject, and we have to rely on recent reliable reviews to help us report what's significant and skip what's not. -EB

Really, the decision what to put in there, and what to take out, seems quite arbitrary on your part. Recent findings no good, only reviews good? That ensures the article will not be up to date, in other words ensuring it will be inadequate. And you say we should use "recent" reviews, but you have citations to reviews from 2007. A very long time ago. If this is how you’re going to apply the policy, and this is what the article’s going to be, I don’t want any part of it. But I will tell you, you are missing a huge amount of good information from me that could be attributed to "unpublished studies" or such. -ZY

Cutting edge research

“Hi, I added a sentence about recently found biomarkers for autism, including abnormal cytokine profiles, reduced melatonin in blood, and abnormal brain autoantibodies. These are from top-of the line, cutting edge researchers.” -ZY

“Unfortunately Wikipedia cannot accept cutting edge research even from top-line labs; we need reliable published sources to support claims in an encyclopedia. (See the verifiability policy.) We need review articles, especially by people outside that research group. For autism there should not be a problem finding good sources, such as the one that states “Research into causes has been hampered by the inability to identify biologically meaningful subpopulations.” -EB

“That last line is just no longer true. Biologically meaningful subpopulations have been found in recent research, even if a review hasn’t been published on it yet. I think you are probably doing more to ensure the public is uninformed or misinformed on this topic than all but some small number of out and out charlatans and crackpots. You keep out solid new information, while keeping in old statements proven false. This is exactly what you did before in regards to auto-antibodies, refusing to incorporate recent research. At the least, the new findings should be mentioned. You have no idea what you’re doing, don’t know the research, and are just misinforming the public.” -ZY

“I’m sorry you feel that way, but you obviously don’t know how Wikipedia works, and that there’s a reason why we only like review articles. Read the policy.” -EB

Cutting edge research

Hi, I added a sentence about recently found biomarkers for autism, including abnormal cytokine profiles, reduced melatonin in blood, and abnormal brain autoantibodies. These are from top-of the line, cutting edge researchers. -ZY

These aren't verifiable sources, and what you put in isn't fully supported. You obviously don't know how to read a medical research paper. The studies aren't enough to contradict that “Research into causes has been hampered by the inability to identify biologically meaningful subpopulations.” We rely on review papers, so idiots like you don't fill Wikipedia with quack science and mere speculation. -EB

That quote is just no longer true. Biologically meaningful subpopulations have been found in recent research, even if a review hasn’t been published on it yet. I think you are probably doing more to ensure the public is uninformed or misinformed on this topic than all but some small number of out and out charlatans and crackpots. You keep out solid new information, while keeping in old statements proven false. This is exactly what you did before in regards to auto-antibodies, refusing to incorporate recent research. At the least, the new findings should be mentioned. You have no idea what you’re doing, don’t know the research, and are just misinforming the public. -ZY

I’m sorry you feel that way, but you obviously don’t know how Wikipedia works, and that there’s a reason why we only like review articles. Would you like some blocks to play with, like the kid in the picture? -EB

Cutting edge research

“This article is missing important information about recently found biomarkers for autism, including abnormal cytokine profiles, reduced melatonin in blood, and abnormal brain autoantibodies. These findings are from top-of the line, cutting edge researchers and you are really doing a disservice to readers by keeping this stuff out. Where the article says that biologically meaningful subpopulations have not been found, that statement is pretty clearly obsolete, even if a review hasn’t been published on the new findings yet. The reviews cited here date way back to 2007, a very long time ago. This autism article is just plain terrible as written, it's so misleading in so many ways, so out of line with current research, so poorly written, that it should not even be in the encyclopedia, it's doing more harm than good. It is important to fix this for parents, society, and readers. I think you are probably doing more to ensure the public is uninformed or misinformed on this topic than all but some small number of out and out charlatans and crackpots. You keep out solid new information, while keeping in old statements proven false. At the least, the new findings should be mentioned.” -ZY

translation to french proposal

Hi there, I'd like to get in charge of translating this purpose that interests me that much. To keep our use efficient, i'm just trying to brief. By the way, I'm not aware of the process to figure out my project. Anyway, I'm up to. Regards ;) — Preceding unsigned comment added by Gambi 13 (talk • contribs) 23:35, 20 January 2011 (UTC)

I'm afraid I haven't understood what you just said. --Anthonyhcole (talk) 12:13, 21 January 2011 (UTC)

I believe he was offering to make a French page correlating to this one. He didn't want to make the entire thing, just key parts. -- Homo Logica (talk) 20:36, 21 January 2011 (UTC)