MedicalFilter April 26, 2016 7:32 AM   Subscribe

This post is based on is sensationalised and scaremongering reporting and should be deleted. And then the ground should be salted so that no other post like it can ever grow again.

The inclusion of the NHS news blog and the link to the paper are appreciated, but there seems to be very little comment on them, and a lot of responses to the first article.

I work in a medical field that's quite different to this one, so I may have made some mistakes, but I’ve gone through the paper and the points that I thought should be highlighted are:

• This is only relevant for people taking anticholinergic medications (ACs) after they are 65.
• There is only an association between AC and dementia at the highest dosage level that was included. There was no association found otherwise.
• There was no comment on the relative benefits of using ACs vs this risk.
• Association does not mean causation. Association does not mean causation. Association does not mean causation.

On this basis, I would say that a vanishingly small proportion of people should be concerned. But that's not how it was interpreted, with a bunch of people talking about changing their medications, offering potentially dangerous advice etc.

In general, I feel that posts like this on public health are always going to amplify the more sensationalist aspects of the story. They’re not going to promote good debate, and they could encourage people to make bad decisions about their health. This kind of story is only one level above ‘burnt toast is a cancer risk!!!’. I don't think it should be on metafilter.
posted by Ned G to MetaFilter-Related at 7:32 AM (30 comments total) 19 users marked this as a favorite

Now I can't eat burnt toast? GODDAMMIT.
posted by grumpybear69 at 7:35 AM on April 26, 2016 [9 favorites]


Association does not mean causation. Association does not mean causation. Association does not mean causation.

Oh, great. Now you've summoned the Association does not mean causation ghost.
posted by bondcliff at 7:36 AM on April 26, 2016 [46 favorites]


what if i just burn it on one side?
posted by pyramid termite at 7:37 AM on April 26, 2016


CTRL-F "ned g": zero results.

Perhaps you could be saying this in the thread?
posted by Etrigan at 7:37 AM on April 26, 2016 [15 favorites]


I did not love this post or the "Hey spanking is actually bad for you" post (which got deleted, thank you). I've been following this topic (the benadryl = alzheimers one) closely for a while and while I'd love to see a good post on this topic, this was not it. Nothing personal Sleeper (eponysterical!)
posted by jessamyn (retired) at 7:39 AM on April 26, 2016


No, we can't say for certain the ghost will pop up if you say that three times. The link has never been proved clinically.

But, yeah, if the post had more substance to it, I would be happier. I don't like single links to games, either, or single obit links. Tell me why it's interesting and important. Give me context.
posted by blnkfrnk at 7:39 AM on April 26, 2016 [7 favorites]


And then the ground should be salted so that no other post like it can ever grow again.

Do you want to give our FPPs high blood pressure?
posted by phunniemee at 7:40 AM on April 26, 2016 [7 favorites]


Perhaps you could be saying this in the thread?

Yeah, if I'd got there early I might have tried to. But there is a point where I think a post is bad enough not to exist, rather than bad enough to have me disagree with it. I think this post crossed that boundary, though I'd be interested to see if other people agree.
posted by Ned G at 7:41 AM on April 26, 2016 [6 favorites]


Oh, great. Now you've summoned the Association does not mean causation ghost.

How do you know that?
posted by Ned G at 7:42 AM on April 26, 2016 [36 favorites]


I don't carry any water for that post, but I thought that the (at least former) consensus on this type of stuff was that Metafilter opts for more speech rather than less. Bad medical answers aren't deleted, the expectation is that other folks will come along and correct them in thread.

Maybe this has changed, or maybe I've misunderstood this in the past, but that would seem to argue against deletion.
posted by OmieWise at 7:42 AM on April 26, 2016 [5 favorites]


"I’ve gone through the paper and the points that I thought should be highlighted are:

• This is only relevant for people taking anticholinergic medications (ACs) after they are 65.
• There is only an association between AC and dementia at the highest dosage level that was included. There was no association found otherwise.
• There was no comment on the relative benefits of using ACs vs this risk."


I think it would be helpful if you put each of those points into the thread. Correlation and causation were pretty well covered. When it comes to a medical study, it is often hard to understand or impossible to access the entire text, so people are only reading an article that summarizes the study. In this case the study is paywalled. Remove the thread from your Recent Activity after posting the comment so you don't have to see the counterarguments if you don't have the energy to respond to all of them.

None of that is to say you shouldn't/couldn't also post this, but it would provide some balance assuming the thread is not deleted.
posted by soelo at 7:55 AM on April 26, 2016 [4 favorites]


How do you know that?

Hee hee!
posted by the quidnunc kid at 8:54 AM on April 26, 2016 [5 favorites]


I disagree, in that the actual research paper is not your typical "broccoli associated with 1% increase in cancer" paper. I did not read the popular press accounts, and I don't have a dog in the fight as to whether the metafilter post was good or not, so I won't comment on those issue.

But I will re-post my comments here as I think they are relevant in making the point that the original research paper itself is not a terrible one, and could be breaking new ground scientifically.

I just skimmed the article, and although it's a house built on a lot of correlations, they tell a story that's consistent. In particular, the one that stands out to me: the relative risk ratio of cognitive decline was 2.5 for those taking AC meds, but 7.7 for those with AC meds who are also amyloid beta positive. While this is still just correlational, this are correlations that hint at a plausible mechanism.

As midmarch says, it's not clear how to interpret cognitive testing in people currently using anticholinergics, however the 96 month followup including progression to MCI or Dementia suggests this is not solely a "day of test" phenomenon. If it were, then the experiment is easy: take the AC+ folks off their meds for 72 hours and retest them, and they should all bounce back to normal.

The biggest "hmm" things to me in the study are that the AC+ folks have statistically significantly higher rates of depression and anxiety and overall medication use, and also higher rates of other disease conditions that may not be statistically significant. Although the authors tried to control for these statistically, that's usually not a guarantee that these factors are weeded out.

posted by soylent00FF00 at 10:11 AM on April 26, 2016 [2 favorites]


Also nothing in the post that is publicly available even bothers to list what the anticholinergic medications are and a lot of people in the thread were jumping to assumptions about compounds that don't qualify. Particularly as Benadryl is a brand name used for more than one active compound (in several countries at least), so using that word as if it means something useful is misleading at best. Which specific active compounds the post is even about is vital to even begin to understand the story.

I haven't read the linked paper yet because I was too busy at work today so I don't know if it's outlined in there. But most readers here won't have access to it anyway so it's kind of moot. Missing that information is what pushes the whole post over into scaremongering for me. We are better off without this level of bad science reporting here in my opinion, despite that I've largely given up on caring how shitty science reporting is in the world in general.

(I'm a research biomedical scientist in case that matters at all)
posted by shelleycat at 11:19 AM on April 26, 2016 [7 favorites]


The public also learns something useful about how our field works from review and criticism. This is not the first Bad or Corrupt Science post that has shown up here, and it likely won't be the last, but so long as fact-based responses are allowed to stand, perhaps immediate deletion is not always merited.
posted by a lungful of dragon at 12:05 PM on April 26, 2016 [2 favorites]


This got confusing because people were bringing in other studies. The original link - which was not good - had that scare quote about cumulative dose so people found the paper that was (misleadingly) derived from to debunk it and that conversation is going on while other people are reading the pull quote and just dropping in to say oh no.

Also I could swear I did find the full text of both of these papers even though the original link was paywalled and I don't have a subscription to anything. Did they not make it into the thread? Am I mistaken about having read the primary cited one yesterday?
posted by atoxyl at 12:08 PM on April 26, 2016

And then the ground should be salted so that no other post like it can ever grow again.
Do you want to give our FPPs high blood pressure?
It's the electrolytes FPPs crave!
posted by maryr at 3:26 PM on April 26, 2016 [4 favorites]


what if i just burn it on one side?

then you just get cancer on one side, duh, do you not even science?!??
posted by escabeche at 4:48 PM on April 26, 2016 [5 favorites]


If we start deleting FPPs because Mefites are likely to stupidly misinterpret and overreact to them, I f:5(?dc;Yu!&€^~{\:(6}_.
posted by Joseph Gurl at 5:42 PM on April 26, 2016 [2 favorites]


MeFi can't bear the burden of fact-checking every article posted, medical or otherwise. I'm squarely in the camp of letting it stand (not because it's a great, well-contextualized post, but because it doesn't break any guidelines), because I believe it's better for me and other people to see this here, and see it carefully broken down and debunked by people with a larger perspective and more expertise, than to have it come up elsewhere where I don't have access to that sort of review and context. If it's going to show up on Facebook, I appreciate having encountered it here and knowing better how to place it.
posted by Miko at 5:49 PM on April 26, 2016 [7 favorites]


Yeah, I don't think metafilter can or should have to fact check posts. But as there is more potential for harm from misinformation here than in most other places, I think that the bar for this type of post should be far higher. I'm not sure that the debunking argument is a great one either, reading through the post there seems to be a fair amount of confusion, and a lot of opinions flying around, and it's difficult to sort the wheat from the chaff. I think that an open forum of non expert readers is fundamentally not a good place to disseminate and discuss technical information. In most threads, I'm happy to eye-roll and move on, but here, it seems like there is a potential cost in allowing it to stand.

There's also an approach to journalism that attempts to induce as much fear as possible in your readership, notwithstanding any facts (see: the daily mail). This post set those alarm bells ringing, which is why I started looking at it.
posted by Ned G at 4:25 AM on April 27, 2016 [2 favorites]


I only looked pretty briefly at this post and the accompanying discussion, but my impression was the same at Nat G's—this was a fairly speculative report with a narrow scope and some obvious unaddressed caveats, and which was drawing a lot of criticism in its field. I was pretty surprised to see how seriously it was being taken by my fellow MeFites, with people apparently using it to make medical decisions in their own lives based on what looked to me like some incredibly shaky evidence that probably didn't even apply to most of them. Surprised and a little disappointed, to be honest. I wasn't feeling like I wanted to really engage with this (partly because, like Nat G said, a lot of the damage had already been done by the time I saw it) but I felt like it was a poorly-framed post and was generating some pretty poor discussion. It was really sensationalized and the community appears to have bought it pretty uncritically.

I would be fine with it being deleted, if only to protect future information-seekers from running across it and using it to make malinformed medical decisions. I know that that's not normal MeFi policy, but I think posts and discussions like that one have a potential to cause tangible, ongoing harm just by existing and so I think an exception would be reasonable in this case. I don't think the mods should be responsible for vetting every medical post that comes along (that would be unreasonable) and I wish that someone had seen it earlier and alerted them of the potential problem before the discussion went as far as it did, but apparently nobody did and now here we are. In cases like this one I think the post should be deleted regardless of the fact that it's already many comments in.
posted by Anticipation Of A New Lover's Arrival, The at 7:00 AM on April 27, 2016


I would definitely be in favor of instituting some guidelines on new medical and social science research papers for FPPs. Arguably, posters who cannot read the paper and comment intelligently on aspects such as those raised by Ned G should not be posting FPPs on the subject. I don't think Metafilter FPPs should just reprise crappy science journalism, basically.
posted by yarly at 8:25 AM on April 27, 2016 [3 favorites]


there is more potential for harm from misinformation here than in most other places

Why would that be the case?
posted by Miko at 9:46 AM on April 27, 2016 [2 favorites]


Just came in to second yarly; was surprised at the pushback on what I thought was quite a sensible MetaTalk suggestion.

At any rate, if the community doesn't agree, it doesn't agree. But +1 to Ned G for raising this issue.
posted by joseph conrad is fully awesome at 10:28 AM on April 27, 2016


I agree with the problem but don't agree on the solution of deleting; I think the pushback in thread is well done (thanks for adding your comments, NedG). Another issue is that people who are qualified to comment professionally may not be able to post something, because they could easily know the study authors. Research fields can be tiny.
posted by nat at 4:16 PM on April 27, 2016 [1 favorite]


there is more potential for harm from misinformation here than in most other places

Why would that be the case?


Sorry, I wasn't very clear, I meant that if a post on a medical topic is misleading then there's more potential for harm than if a post on almost any other topic is misleading. I think that's a reasonable position.
posted by Ned G at 2:31 AM on April 28, 2016


Well I'm here now, but I swear I was just in the area, and my presence has very little to do with the, er, Summoning.
posted by the Association does not mean causation ghost at 3:04 AM on April 28, 2016 [4 favorites]


I meant that if a post on a medical topic is misleading then there's more potential for harm than if a post on almost any other topic is misleading.

I would agree with this stance if the media universe were confined to MetaFilter. But since people are immersed in messages from waking to sleeping, and are likely to encounter bad medical messages everywhere (especially on Facebook, but also on TV and via email), I appreciate that they can be aired here with some critical examination likely to follow. I am a broken record on this, but the solution to the existence of bad information is not prohibition on communication, but more information, which is likely to include better information. In addition to the ways that can help here, I'm sure you've noticed that when people Google a topic, MeFi is often a first-page result, so the pushback that happens here when bad or ambiguous medical information appears can extend to be helpful beyond the site.
posted by Miko at 7:41 AM on April 28, 2016 [1 favorite]


Dormitator Delenda Est!
posted by y2karl at 8:16 AM on April 29, 2016


« Older MealFilter   |   Happy birthday Cortex! Newer »

You are not logged in, either login or create an account to post comments