« Quote of the Day | Main | Chart Interpretation for Bloomberg Dummies »

October 20, 2014

OMG!! THOUSANDS OF CENTRAL AMERICANS HAVE ENTERVIRUS D68!!!!!11!!

This kind of idiocy (from Alabama Rep. Mo Brooks) begs to be fisked within an inch of its miserable life:

“Immigration is part of Ebola, is a part of this new virus – I say ‘new’ in quotations marks because it’s relatively new to the degree we’ve seen it in the United States of America that taking the lives of American children, that is causing partial or complete paralysis of American children. All of this is related to immigration because some of these diseases are coming from abroad. By way of example, there was a study in 2013 – I think it was called the enterovirus that is causing the paralysis and death of young children in America – that thousands of residents of Central American countries were found with this illness over a year ago in 2013.

We touched on this topic last week, but we're still seeing volunteer clowns everywhere we turn. Unfortunately for them, the study doesn't say anything even close to what Brooks (and far too many other bloggers) claim it does. Let's walk through Brooks' claim step by step:

(1) Step 1: "...thousands of residents of Central American countries"...

The total number of patients studied (taken from 8 different countries, only 2 of which are in Central America) was 3375. Only 246 of them came from Central America:

central_amer.png

Now unless Rep. Brooks and every blogger who has cited this study are using some kind of newfangled math (perhaps Common Core?), it is impossible to get "thousands" from a sample of 246, not all of whom even tested positive for any of the 100-odd types of enterovirus. Thus, simple examination of the first part of his claim is sufficient to show that this man has absolutely no idea what the study he's flogging actually says. But hey, this is fun. Let's keep going.

(2) Step 2: "...were found with this illness". Again, the study says nothing of the kind. Or even anything close. "This illness" is one of about a hundred known strains of human enterovirus (HEV): Enterovirus D68. So, how many TOTAL cases of human enterovirus (of ANY kind) were found? From the study results:

Overall, HRVs and HEVs were identified in 16% (548 samples) and 3% (84 samples) of the ILI cases, respectively.

That 3% (or 84 cases) is comprised of all types of enterovirus, not "the specific strain currently spreading in the United States. And the 3% was taken from 8 different countries (not just Central America). So once again, clearly the people citing this study either haven't bothered to read the [short] summary or can't interpret a chart to save their own lives. It's right there in black and white. It's even in English.

Now let's move to the chart I keep seeing. Due to the extremely graphic [pun fully intended] nature of this chart, I have hidden it below the fold so as not to send the unwary reader into sudden and fatal cardiac arrest:

HEV-D68.png

What does this chart tell us? Well, from a total of 3375 patients, only 632 of whom tested positive for any of the viruses being studied, the study found precisely 10 cases of HEV D68. That's right: 10. By way of contrast, here are the numbers of cases found during the last 6 global outbreaks of Enterovirus D68 (fully half of which occurred right here in the United States):

2008
Philippines - 21 cases

2009
Georgia, USA - 6 cases
Pennsylvania, USA - 28 cases

2010
Netherlands - 24 cases
Japan - over 120 cases
Arizona, US - 5 cases

Notably, the United States is the ONLY country represented more than once in the list of global enterovirus D68 outbreaks. But enough of all these distressing references to actual study findings. Let's move to the final element of Brooks' claim:

(3) [the study found THOUSANDS OF CASES OF THIS ILLNESS IN CENTRAL AMERICA].... in 2013.

Wrong. Again, reading is fundamental. The study was published in 2013. The samples (according to this chart) were taken in 2010 and 2011 (for those of you without home calculators or anyone who has cited this study as "proof" that HEV D68 came from illegal Central American immigrant children), that's 3-4 years ago.

For Pete's sake, if you're going to cite this study (or worse, post and then completely mischaracterize charts taken from it) read the study. This isn't rocket science. This is precisely the kind of thing bloggers rightly criticize the NY Times for doing.

What's our excuse?

Posted by Cassandra at October 20, 2014 08:45 AM

Trackback Pings

TrackBack URL for this entry:
http://www.villainouscompany.com/mt/mt-tb.cgi/5352

Comments

Nice analysis and explanation of the numbers.

Posted by: Frank Karl at October 21, 2014 10:11 AM

Thanks, Frank :)

Posted by: Cass at October 21, 2014 10:48 AM

Two thoughts: First, any large-scale mixing of populations will spread diseases. That happens, for instance, at universities, when new students come in from all over.

The shipping of tens of thousands of immigrant children all over the United States must have helped spread cases of disease to the native population -- and to the children. Almost all of them will be minor, colds and the like, but a few might not be.

Second, past outbreaks of Enterovirus 68 in the United States have been localized. This one seems to be popping up all over.

Now I can think of a number of posible explanations for that. For instance, it is possible that EV 68 has been around all over for a long time, but we just weren't seeing it, because it is so similar to colds. If that were true then it would get diagnosed only when you had one of those rare serious, or even tragic, cases. And maybe not even then.

(That there have been more outbreaks in United States is almost certainly explained by better diagnostic procedures here. I doubt whether there are even dozen nations that can diagnose the disease. This new test will give you an idea about how the disease is detected.)

But I do think it possible that it was spread by that surge of children, and I do think that the widespread outbreak deserves investigation.

Your two posts leave me wondering whether we agree on those two points.

(For the record: Of course I agree with you that some have taken the argument far beyond what the evidence allows.)

Posted by: Jim Miller at October 21, 2014 01:05 PM

You make some great points, Jim. Allow me to respond:

Two thoughts: First, any large-scale mixing of populations will spread diseases. That happens, for instance, at universities, when new students come in from all over.

Absolutely. It also happens every September when the kids start school. Or when kids are put in day care and bring home all sorts of germs and viruses that infect the rest of the family.

The shipping of tens of thousands of immigrant children all over the United States must have helped spread cases of disease to the native population -- and to the children. Almost all of them will be minor, colds and the like, but a few might not be.

I think that's quite likely, but since we're not seeing large scale epidemics of previously unknown diseases, this remains just a theory. Keep in mind that these kids weren't all moved at the same time. And the states with the early outbreaks were a loosely contiguous group mostly located in the middle of the country.

Second, past outbreaks of Enterovirus 68 in the United States have been localized. This one seems to be popping up all over.

Yep.

Now I can think of a number of posible explanations for that. For instance, it is possible that EV 68 has been around all over for a long time, but we just weren't seeing it, because it is so similar to colds. If that were true then it would get diagnosed only when you had one of those rare serious, or even tragic, cases. And maybe not even then.

Bingo. A lot of the articles I've seen have pointed out that this virus is extremely common, but most people never even realize they had it. In that way, it's similar to the [related] coxsackie strain that put one of my little nieces in the hospital a few years ago.

(That there have been more outbreaks in United States is almost certainly explained by better diagnostic procedures here. I doubt whether there are even dozen nations that can diagnose the disease. This new test will give you an idea about how the disease is detected.)

This is the best point you made (and one that seems to follow logically from the study). But that begs an important question: the sample from the study were kids who were *more* likely to have EV68 than the general population because they're already in the hospital with respiratory problems. And yet EV68 was described as "we also found small amounts..."

Which is very problematic if you're claiming that this study *proves* it came from Central America, and that it's common there! This study does nothing of the sort.

But I do think it possible that it was spread by that surge of children, and I do think that the widespread outbreak deserves investigation.

I think that's one possibility among many. Another is that there's some other vector we aren't seeing.

Another is that the strain mutated and became more virulent (so IOW, the same or slightly more kids got it, but because the virus was more dangerous this year, more kids had serious complications). IOW, did more kids *get* the virus in the first place this year? Or are we just seeing more kids with serious symptoms?

Your two posts leave me wondering whether we agree on those two points.

Well, I'm not sure why. My first post carefully pointed out that nothing I said proved EV68 *didn't* come from/wasn't spread by immigrant kids. I only said that the evidence presented so far as "proof" was pretty laughable.

For example: if (as was claimed) EV68 was common in Central America and infected kids came here and then were disbursed to other states, shouldn't we see more cases in states that received more children?

That's not what happened, though. The first outbreaks were in states that received relatively few children. I can think of explanations for that. But more tellingly, the states that received the most children don't have the most cases, AND they got it later than states that got far fewer kids.

That doesn't disprove the theory that these kids spread the disease (the traditional method is to look for patients 0, 1, 2... in each outbreak - a practice that was inexplicably ignored in the posts I've read). But it *does* cast considerable doubt on an already pitifully weak argument that relies on almost willfully mischaracterizing the cited study.

(For the record: Of course I agree with you that some have taken the argument far beyond what the evidence allows.)

That is the ONLY point I'm trying to make. Though I did hope to suggest a further point: that crying wolf with such a poorly supported claim is a really bad idea that could well cause people to discount the more-reasonable point you made (mixing populations *does* increase the chances of spreading disease).

Had these folks gone out with that argument, you'd have likely found me agreeing wholeheartedly. Instead, though, they threw out something that sounds like a paranoid conspiracy theory, supported by 'evidence' that - when you bother to read it - doesn't even say what they claim it does.

Somewhat ironically, my initial interest in this story was due to my thinking there might well be something to it! It's just that when I started to think about it, it just didn't stand up to even casual inspection.


Posted by: Cass at October 21, 2014 02:10 PM

One more point, Jim.

If the kids are the contagion vector, then why didn't the earliest outbreaks occur in the places where the kids were being kept?

Also, we saw tons of articles earlier this year about TB, dengue fever, scabies, even leprosy among these kids. But none about EV68??? Not even any about a strange respiratory illness.

Doesn't that seem more than a bit strange to you?

Posted by: Cass at October 21, 2014 02:36 PM

One more thought: if the children are the vector, what explains the 3 outbreaks in British Columbia, Alberta, and Ontario?

Posted by: Cass at October 21, 2014 02:51 PM

Well, one possibility is that there are multiple vectors acting simultaneously. It could be that better diagnosis nor the surge of children explain the outbreak fully by themselves.

I'm buried deep in my own model documentation to read the docs, but some generic avenues for investigation would be:

A study of 246 patients, that were hospitalized, certainly increases the chances of infection with D68 and detection. However, even with vastly lower infection rates, once you multiply that (vastly lower) probability by the population size, you could still get a rather large number of infected people. Increase that probability by those who carry the disease, but don't actually get sick, and you could get a non-trivial proportion of carriers.

However, even if the total number of infected people is non-trivial (well over "thousands") the likely number of people infected in any given 40,000 sample may still just be 1 or 2 people. I have no idea if the children involved in illegal immigration would be more or less likely than the general population to carry the virus. I could make an plausable hypothesis in both directions: Rich kids with a history of good medical care don't seem particularly likely to become illegal immigrants implying that these kids are disproportionately selected for poor medical care and higher disease rates. At the same time, the kids had to have come from some level of resources. The didn't walk all the way here stealing bread along the way. While they are certainly here to improve their situation, they could still be from the middle-upper socio-economics ranks from their home countries.

But if infection rates are 1 or 2 individuals, where they show up after being distrubuted is essentially random and wouldn't necessarily follow the largest groups. It could just be dumb luck that one showed up in Kansas and another in Kentucky. It then spread as any other disease: schools->Parents->business/vacation travel, etc.

Combine that with a concurrent outbreak domestically and untangling the two effects could be impossible.

Frankly, I'm more worried about those kids distributing a different strain of flu or other more common diseases. Ones which our immune systems (and/or vaccines) aren't prepared for.

Posted by: Yu-Ain Gonnano at October 22, 2014 09:56 AM

From what I've read, this strain usually generates few noticeable symptoms in the vast majority of folks infected with it.

That's a statement that's just as true of people who are already here in the US as it is of people from anywhere else.

I read somewhere this morning that the strain that killed the 7 or so children has been determined to be essentially the same as the one in previous outbreaks. I will try to find the link tomorrow and also re-read (I was in a bit of a rush and may well be "dismembering" what I read).

But if infection rates are 1 or 2 individuals, where they show up after being distrubuted is essentially random and wouldn't necessarily follow the largest groups. It could just be dumb luck that one showed up in Kansas and another in Kentucky. It then spread as any other disease: schools->Parents->business/vacation travel, etc.

Combine that with a concurrent outbreak domestically and untangling the two effects could be impossible.

True, and that's something that occurred to me whilst writing my first post. But keep in mind that the thesis I was addressing was that EV68 was widespread in Central America (which it may well be - from what I've read it's actually pretty widespread here! - it's just that the vast majority of folks who get it just think they have a cold and that's the end of it).

Frankly, I'm more worried about those kids distributing a different strain of flu or other more common diseases. Ones which our immune systems (and/or vaccines) aren't prepared for.

Me too. In the study, the largest group of patients who were hospitalized were under 5. Adults had antibodies b/c they had already had it once. I've read the same thing here several places.

I'll try to post the info I read last night later - it was interesting stuff.

Posted by: Cass at October 22, 2014 08:36 PM