Archive | Epidemiology RSS feed for this section

Anti vaccine activists are angry about a new study…and they didn’t even bother to read it

14 Mar

This is a big piece of what the “vaccines-cause-autism” idea is built on. Really bad analyses. Another study shows up showing that vaccines don’t cause autism. People immediately jump to give talking points to their community: “ignore this study! In fact, it’s just another conspiracy to defraud you!!!!!”*

But they know from history, their community doesn’t check. They don’t test whether the talking points hold up. Either that, or they really are this bad at simple math.

One reason I slowed down a lot on writing debunks of the “vaccines cause autism” community is that it’s very repetitive. It only takes a few minutes to see where they make their mistakes. I admit, I have the training and the time to do that checking, but it leaves me wondering. As in, it’s hard to not ask: “are they really this bad at science? Are they so biased they don’t know they are wrong? Or, are they lying?” I don’t know. What matters most is they are wrong. I am grateful that I had the time to spend checking their claims and was able to not get sucked into that world.

That intro aside, there was a recent study that, once again, shows that the MMR vaccine doesn’t increase autism risk. 21 years since Wakefield’s junk study and we are still spending time and money countering his misinformation. When you look at what the autism community needs, it’s just sad to see resources used this way. But I get it. The anti-vaccine community is doing real harm (note the rise in measles in the US and elsewhere). But, dang, I’d like more work done to help my kid.

So, what’s this new study? Measles, Mumps, Rubella Vaccination and Autism: A Nationwide Cohort Study. With the conclusion:

The study strongly supports that MMR vaccination does not increase the risk for autism, does not trigger autism in susceptible children, and is not associated with clustering of autism cases after vaccination. It adds to previous studies through significant additional statistical power and by addressing hypotheses of susceptible subgroups and clustering of cases.

Thanks for doing this, but, again, the results are not surprising in the least.

Naturally this brought about a big response from the anti-vaccine community. The go-to place for defense of “vaccines cause autism” used to be the Age of Autism blog. I went there and found they just recycled a piece from a guy named James Lyons Wieler. Apparently he was once a scientist, but he now seems to be a guy asking autism parents for money to write bad articles about how vaccines cause autism. In this case has found the “Smoking Gun” for why the MMR study is so wrong.

(1) The smoking gun is the study-wide autism rate of 0.9-1%. The rate of ASD in Denmark is 1.65%. Where are the missing cases of ASD? Given past allegations of this group’s malfeasance and fraud, the rest of the study cannot be accepted based on this disparity alone: the study group is not representative of the population being studied.

Clearly they must be fudging the data!!!!!

This appears to have become the talking point that the anti-vaccine community is passing around. It was picked up by Brian Hooker. He wrote “A Scientist’s Rebuttal to the Danish Cohort Study“:

1. Children were notably missing from the study sample:

First and foremost is the underascertainment of autism cases within their data sample. The study authors used Denmark population registries of children born in Denmark of Danish-born mothers which should reflect the current reported autism incidence in Denmark at 1.65% (Schendel et al. 2018, JAMA). However, the autism incidence within the sample of the Hviid et al. paper is 0.98%, meaning that approximately 4,400 autistic children are missing from this study. The authors do not discuss the discrepancy in the number of cases.

Again, his number 1 point, the missing children!

Even JB Handley (remember him? Yes, he’s back) wrote about this. Focusing his whole piece on this “missing” group in the MMR study: New Danish MMR study shows autism rate of 1 in 100—CDC should rush to Denmark!

2. The most compelling data in the study will never get covered: why is the autism rate in this study only 1 in 100?

Here in the U.S. we’re at 1 in 36! Shouldn’t CDC researchers rush to Denmark to figure out why their autism rate is so much lower than ours? For every 1,000 Danish kids, only 10 have autism. But here in the U.S., we have 28 per 1,000, that’s 177% more autism! I thought Paul Offit wanted everyone to believe the autism rate was the same everywhere? What gives?

Hold on to your seat, Handley. It’s about to get discussed.

I had thought I’d take the easy route and just email the study author for an explanation. That could have answered the big question for Handley, Weiler, and Hooker. But that would take a day or two to get a response from Denmark. Why not just, you know, read the paper? Or, just the introduction?

Under the “Abstract” section of the MMR paper, which has the “missing” autistic kids and a lower prevalence rate:

Participants: 657 461 children born in Denmark from 1999 through 31 December 2010, with follow-up from 1 year of age and through 31 August 2013.

Under the methods section for the earlier Denmark paper, which has the higher prevalence rate:

All live births in Denmark between 1980 and 2012 were identified in the Central Person Register and followed through 2016 for an ASD diagnosis

It’s not that hard to compare the two studies.

One study looked at Danes born from 1999 to 2010. And took data from 2013. This is the MMR study.

The other study looked at Danes born from 1980 and 2012. And took data in 2016.

Apples, meet Oranges.

Seriously, people are surprised that they came to different answers as to the overall prevalence? I mean, this is your “smoking gun”? This is the best the “vaccines cause autism” community can do? If nothing else, one study took data later than the other. You are the “it’s an epidemic!” team, surely you accept that the autism rate is higher in the later dataset.

But, hey, this didn’t take the full 5 minutes I allocated to check the claims of this “smoking gun” against this new study. I still had 4.5 minutes.

So, let’s see if the data really are compatible. Can we take the data from the prevalence study and get the same number as in the MMR study? Yes, I’m a geek and this is what I do. But we just saw that 2 Ph.D.’s (Hooker and Wieler) and a business guy (Handley) didn’t think to do that. Is it really that hard? (I do wonder how Handley made money. Seems like he must have relied on someone else to do the numbers.)

I just wrote about the autism prevalence study: Yes, there are a lot more adult autistics than commonly thought. The real question is what we do with this information. I have the graph from the prevalence study, so I ran the numbers quickly. If we limit ourselves to the autistics in the MMR study (born 1999 to 2010) and take data in 2013, we get a prevalence value of 1.02%.

1.02% using the prevalence study. Compared to 1% in the MMR study.

They are the same. No “malfeasance”. No “fraud”. No “discrepancy”. And, Mr. Handley, no evidence you can use to blame the HepB vaccine for autism.

Now for the dull part. Here’s my math.

Step 1: I digitized the graph. The red points are where I took prevalence data from the graphs. Each line represents 2 birth years, so I took points where for the age of the average kid in each cohort in 2013.

Here’s the summary table from those data points.

I did this fast. Let me know if I made a mistake. That’s why I’m showing my work. It’s not precise because, well, it’s done by hand. Also, there’s the fact that the MMR paper was for kids born from 1999 to 2010. The prevalence study has kids grouped by 2 years. So I have data for 1998-99 where I only really want 1999. It’s good enough. The “age in 2013” is what the digitizer gave me for the datapoint positions I chose. I can’t get exactly, say, 10.5. But, again, it’s good enough.

Anyway, there’s no “smoking gun” as James Lyons Weiler says. There aren’t children “notably missing” as Brian Hooker claims. And the “most compelling data” according to JB Handley is just that he can’t read a scientific paper.

This is a big piece of what the “vaccines-cause-autism” idea is built on. Really bad analyses. Another study shows up showing that vaccines don’t cause autism. People immediately jump to give talking points to their community: “ignore this study! In fact, it’s just another conspiracy to defraud you!!!!!”*

But they know from history, their community doesn’t check. They don’t test whether the talking points hold up. Either that, or they (people like Handley, Hooker and the rest) really are this bad at simple math.

—
By Matt Carey.

*Yeah, you guys are pushing a conspiracy theory. I know you like to say “oh, he called us conspiracy theorists, therefore we don’t have to listen to him.” Surprise me. Grow a backbone and defend the points in this commentary rather than either (a) ignoring it or (b) dismissing it because I pointed out that you are claiming scientists conspired to fool the world.

Yes, there are a lot more adult autistics than commonly thought. The real question is what we do with this information.

13 Mar

What if I told you that there are a lot more autistics born in the 1980’s than 1 in 10,000. And this isn’t me saying “it’s almost certainly true”. This is fact. I’d hope that one of your first reactions would be, “what should we do with this information to make life better for autistic people?”

Hang around online autism discussions, especially those involving causation, and you will undoubtedly read someone claiming “back in the 1980’s, the autism rate was 1 in 10,000.” People base this on studies done back in the 1980’s and they assume or assert, “nothing has changed since then”. Which is odd, since a whole lot has changes in autism understanding since the 1980’s.

What if I told you that there are a lot more autistics born in the 1980’s than 1 in 10,000. And this isn’t me saying “it’s almost certainly true”. This is fact. I’d hope that one of your first reactions would be, “what should we do with this information to make life better for autistic people?”

If your first reaction was, “this can’t be true. There’s an autism epidemic because of vaccines or some other pet causation theory”, you have your priorities in the wrong place. If you think you are an autism advocate, you aren’t.

Back to the recent study. When I read a recent study from Denmark, I was amazed. No exaggeration, I was amazed to see data so clearly showing that the autism prevalence for autistics born in the 1980’s is much higher than previously reported. I say this with apologies to the study authors, because there’s a lot more in this study. That said, there is also the trend of prevalence* with time. It’s, well, amazing. The study is Cumulative Incidence of Autism Into Adulthood for Birth Cohorts in Denmark, 1980-2012. Basically, the authors mined the Danish Central Person Register for people with autism diagnoses. Here’s the graph from that study (click to enlarge):

There is a phrase for a certain type of graph in my work: spaghetti plots. This isn’t quite that complicated, but it is complicated enough to take some time to walk through. So let’s take that time. Basically, each line shows show data for people born in a given 2 year period as a function of their age. The bottom line, for example, shows data for people born in 1980-81. The data is basically what the autism prevalence would be if someone had reported it in the year for that age.

Let’s take a few examples. Again, for that bottom line (people born in 1980-81). At age 10, the autism prevalence is about 0.02%. So, if you had asked in 1990 (when these autistics were 10), “what’s the autism prevalence for 10 year olds in Denmark”, you’d have been told 0.02%. 2 in 10,000. And if you never looked again, that’s what you’d think the autism prevalence is for people born in 1980. Which is exactly why you hear, “the autism prevalence back in the day was 1 in 10,000.” No one looked again.

But now we have more data. Let’s look at, say, age 20, which would be about year 2000. By this point, the autism prevalence has risen to about 0.13% (about 1 in 800). That’s 6 times larger than when this group was 10. This doesn’t mean that more people “became” autistic. No, it means that people previously undiagnosed were identified. By the time this group was 35 years old (about 2015), the autism prevalence was 0.33% (about one in 300). That’s 16 times larger than when they were 10.

I digitized the data for the 1980-81 group and plotted it. If I get time I’ll make a better graph. It isn’t 100% precise, but here’s that graph (click to enlarge):

Let’s say again what we see here:

1) the autism ‘rate’ for Danes born in 1980-81 is not 1 in 10,000, it’s about 1 in 300. Getting comparable to the autism ‘rates’ reported today.
2) the majority of these autistics were not identified until they were older than 10 years old.
3) the autism ‘rate’ is still climbing today. Yes, they are still diagnosing people well into adulthood.

Now, consider some of the other data in the study. For example, Danes born in 2000-01 had an autism prevalence in 2016 of 2.8%. With no sign of a plateau. When those Danes were 8 years old, their autism prevalence was about 0.77%. Yes, it climbed by a factor of 3+ from age 8 to age 16.

Another way to look at this is: people have been diagnosed throughout the lifespan. And this is still going on.

We’ve only looked at a small subset of the data in this study, but we have to ask ourselves what do we do with these results? I know what I think we need to do:

First, we have to accept that many autistics were not diagnosed when they were children. The autism prevalence in the adult population is much higher than the old data would suggest. It may be (probably is, in my view) close to or the same as the prevalence in children.

Second, we have to accept that the autism prevalence in children is likely higher than what is being reported today. The 2.8% reported for one birth cohort in Denmark is basically the same as that reported in the highest estimates in the U.S.. And there was no sign that this Danish value is the maximum value that will be seen for that cohort, or any other.

Third, and here’s where we need actual action, not just a change in beliefs. If we believe that the supports, services and therapies autistics deserve are different than those targeting other disabilities (which I firmly believe), we should be working harder to diagnose those as yet undiagnosed. This goes for young children as well as adults.

Third, and here’s where we need actual action, not just a change in beliefs. If we believe that the supports, services and therapies autistics deserve are different than those targeting other disabilities (which I firmly believe), we should be working harder to diagnose those as yet undiagnosed. This goes for young children as well as adults.

I am the parent of an autistic teenager. There’s a lot of advice (good and bad) on how to support autistic children. But there’s very little on autism specific supports for adults. And this hasn’t really changed in the past decade or more. One way to get real information is to take a look at what has worked and what has failed for autistic adults. Besides helping those adults and other autistic adults (a good cause in itself), it helps pave the way for the autistic kids coming of age. Well, you know what? It’s hard to do those sorts of studies if we don’t even identify who is and who isn’t autistic in the adult population.

As a final aside, consider the CDC autism prevalence estimates. They use a very different methodology than in this Danish study, which will introduce some differences in the results. That said, the CDC focuses on children 8 years old. The CDC team had good reasons to choose 8 years old, but let’s look at some of the data from this recent Danish study and ask what this tells us about the CDC data. For people born in 2000-2001, the Danish autism prevalence is be about 0.77% at age 8. One might say, “Hey, the CDC autism prevalence for 8 year olds born in year 2000 is 1.13%. Autism is more prevalent in the U.S. than in Denmark.” People making this sort of comparison often then try to fit this into their own favorite causation theory. As in, “Denmark gives fewer vaccines, so vaccines cause autism!”. But, wait. The autism prevalence for Danes born in 2000 is actually 2.8%, well over double the CDC estimate for US kids born in 2000. Makes the comparison of CDC numbers to others a bit of an apples and oranges story.

More to the point, consider graphs of CDC prevalence vs. time. It goes up and up, doesn’t it? Kids born in 2004 have a higher autism prevalence than those born in 2000, right? But if you’ve missed a bunch of the kids because you looked at 8 year olds and many kids aren’t identified by then, can you really say that the number of autistic kids has gone up?

Another way to say this: the CDC data are good for what they are. What they aren’t is an actual census of the fraction of kids who are autistic in any given birth year. Trying to say, “there’s an epidemic” from these data just can’t be done.

But that’s getting off the topic. As I noted above, the real question with these data are not “how do they fit into the failed idea that [exposure x] causes autism”. The real question is, what do we do with the knowledge that there are a lot of autistic adults? That there are a lot of kids, younger and older, who aren’t diagnosed accurately? That autism gets diagnosed through the lifespan? I’ve pointed out what I think above. We act on the data. We do what we should be doing: try to use this information to make life better for all autistics, be they young or old, diagnosed or not.

—
By Matt Carey

*note–I am not an epidemiologist. Rather than try to keep bouncing between “incidence”, “cumulative incidence” and “prevalence”, I will use the terms more loosely than an epidemiologist would.

Here’s one part of Brian Hooker’s “reanalysis” that shows just how cynical the anti vaccine movement is

18 Dec

When my kid was diagnosed autistic I started reading research papers (I am a Ph.D. researcher by profession) and the raw data. One thing that struck me immediately was the fact that minority children are much less likely to get a diagnosis than white kids. And minority kids are diagnosed later.

This inequity really bothers me. Accurately identifying the needs of a disabled child can focus the appropriate therapies and supports on that child. The need to rectify this inequity is 100% accepted within the autism advocacy and research communities.

This inequity poses a problem to people who claim that autism is an “epidemic”. If we are not identifying all the autistics in any given group (we aren’t), autism prevalence numbers are inaccurate. Being inaccurate, how does one compare, say, one CDC prevalence number with one 2 years later and claim a “real” increase?

One can’t. Plain and simple.

So, for years, groups like those promoting the idea that autism is caused by vaccines have not only ignored this inequity, they have actively denied it. They are stuck between accepting that the data can’t show an epidemic, or accepting that minorities have some sort of protection from this supposed “autism as vaccine injury”.

When was the last time you read something from, say, the Age of Autism blog or Andrew Wakefield calling for efforts to end this inequity? You haven’t. They don’t do it. When have you heard from someone like Brian Hooker that we should study minority populations to see what “protects” them from “vaccine injury”? You haven’t.

Who is Brian Hooker? Brian Hooker is a parent of an autistic child. Brian Hooker strongly believes that vaccines cause autism. He can back this up with his observations of his child’s development. Observations which are contradicted by his child’s medical records. I discussed this before as Double checking Brian Hooker’s story in VAXXED. A Special Master (a judge in the vaccine court) put it very strongly:

After studying the extensive evidence in this case, I am convinced that the opinions provided by Petitioners’ experts in this case, advising the Hooker family that there is a causal connection between SRH’s vaccinations and either the initial causation or aggravation of SRH’s ASD, were quite wrong.

In the original, the Special Master emphasized “quite wrong“.

So, we have someone who believes vaccines cause autism to the point of ignoring the facts in front of him.

A few years ago Mr. Hooker “reanalyzed” some data from an old CDC study, suggesting that evidence showed that the MMR vaccine might increase risk in African American boys. That was discussed in great detail here and elsewhere. (for example: Brian Hooker proves Andrew Wakefield wrong about vaccines and autism and MMR, the CDC and Brian Hooker: A Guide for Parents and the Media).

Mr. Hooker’s study was retracted. In the research world thats a big deal. As in, embarrassingly bad.

Recently, as in 4 years after his original study, Mr. Hooker republished his “reanalysis”. In the Journal of American Physicians and Surgeons. I’d be completely embarrassed to have a paper in that journal, to be blunt. A lengthy discussion of this reappearance of the study can be found at Respectful Insolence as Brian Hooker’s antivaccine pseudoscience has risen from the dead to threaten children again.

Let me just focus on how Mr. Hooker, in my view cynically, abuses the African American community in order to attack vaccines. From the website of an organization Mr. Hooker belongs to (the ironically–to be polite–named “Children’s Health Defense”), we read this:

Main Points from Reanalysis:

The rate of autism diagnoses has increased alarmingly in the U.S., and is about 25 percent higher in black children. Boys are far more likely than girls to receive this diagnosis.

…

This is not only wrong, it’s wrong in a way that points to incredible dishonesty.

This first point is that autism is about 25% higher in black children. A “main point from the reanalysis”.

Tell me, when you read that did you think, “this study found that autism is more prevalent in African American children”? If so, you were misled. The 25% higher prevalence is from a different study than Hookers. And that other study says something completely different.

From the Hooker study:

However, one study showed that prevalence of autism in African-Americans was approximately 25% higher than that of whites when the data were adjusted for socioeconomic factors[7].

Reference [7] is Socioeconomic inequality in the prevalence of autism spectrum disorder: evidence from a U.S. cross-sectional study.

The Socioeconomic Equality study states:

Also notable is that, although the overall ASD prevalence was higher among non-Hispanic White and Asian children than among non-Hispanic Black or African-America and Hispanic children, when the results were stratified by SES, we saw that the racial/ethnic differences in prevalence varied by SES (Table 3). The lower prevalence among non-Hispanic Black or African-American and Hispanic children was seen only in the low SES category, and the fact that more non-Hispanic Black or African-American and Hispanic children live in poverty contributed to the lower overall prevalence among these groups.

Emphasis mine. Overall ASD prevalence was lower for African American children. Not 25% higher. This lower prevalence was due to lower socioeconomic status. I.e. poverty.

Want to see this a different way? Here’s a figure from the paper (click to enlarge):

The overall prevalence in White non-Hispanic kids was 6.9/1000. For black non-Hispanic kids it was 5.7/1000. About 20% lower. Not 25% higher as Mr. Hooker claims.

Fewer African American kids are getting autism services. Not because they aren’t autistic, but because their poverty keeps them from getting a diagnosis.

This is something we should be working towards fixing. No question. But don’t look to the anti-vaccine community to care or act. It’s an inconvenient fact for their epidemic story.

I guess he has such a low level of respect for the people in his own community (those who believe vaccines cause autism), that he thought no one would check this.

By the way, this paper isn’t the only one that shows a lower autism prevalence among African Americans. You know those CDC autism prevalence reports that come out every two years? Every single one has reported a lower prevalence among African Americans. Every one.

Here’s a line from the latest report:

Previous reports from the ADDM Network estimated ASD prevalence among white children to exceed that among black children by approximately 30% in 2002, 2006, and 2010, and by approximately 20% in 2008 and 2012.

When I saw the claim on Hooker’s organization’s website I figured he must have cherry picked a study that shows what he needed to make his story work. It’s just such common knowledge in the autism community that African Americans get diagnosed less frequently. It’s in every CDC report. I didn’t know he wasn’t cherry picking, he was just misrepresenting the study entirely.

I discuss this as a scientist. He “misrepresented the study”. My father had a word for that sort of behavior: lying.

OK, Brian. You’ve read the studies and decided to do nothing about the fact that many autistic African American kids aren’t getting identified and getting appropriate services. I get that, you have your own cause. But, really, is that community so much of a nothing to you that you can just use them like this? I ask rhetorically. You and your community have always acted with callous disregard.

I once had hope that as it because completely obvious that you and your community were wrong (and that was many years ago), you’d join the actual autism community and put your advocacy to use. I now know that will never happen. And, frankly, we don’t need dishonest people.

—
By Matt Carey

Why autism at 2.5% isn’t surprising.

29 Nov

Let’s get one thing out right away–autism prevalence studies undercount. Not all autistics are diagnosed. That’s just a fact. Consider the recent CDC study. They look at school and medical records. In many cases, they find children are autistic based on their records–but the schools and doctors hadn’t diagnosed those children.

Combining data from all 11 sites, 81% of boys had a previous ASD classification on record, compared with 75% of girls (OR = 1.4; p<0.01).

Yeah, more than 20% of the kids counted in their prevalence had no diagnosis. They and their families didn’t know.

And, if there isn’t enough in the records to show a kid is autistic? That kid gets uncounted altogether.

So, when people look at the CDC prevalence estimates from over the years and cry “epidemic”, well, there’s a reason why those people usually have some causation theory that they believe in. The irony is that they are usually wrong that their theory needs an epidemic to support it. But, heavily biased people are not usually the best sources of reliable analyses.

What would be a better method of counting how many autistics are in a population? Sounds obvious–test all the kids in a given population. Equally obvious–this is a much more expensive and difficult task. One such study was published in 2011. Yes, 7 years ago. In Prevalence of autism spectrum disorders in a total population sample, the autism prevalence in Korea was found to be 2.63%. A study performed in South Carolina and reported at IMFAR last year found a prevalence of 3.62%.

This all said, we had another autism prevalence come out this week–The Prevalence of Parent-Reported Autism Spectrum Disorder Among US Children. This study found a prevalence of 2.5%.

Now here’s a nice thing about this recent study–OK, two nice things. First, they don’t just look at kids of one age. Second, you can obtain the data. Which I did. Let’s look at the autism prevalence broken down by birth year.

Do you see autism prevalence increasing with birth year? I don’t. I see some scatter, but in general the autism prevalence is about 2.5% from birth year 2000 to 2010. For what it’s worth–the scatter is due to the small numbers of kids in each year making the value uncertain. It’s statistical noise.

If you are wondering about how the autism prevalence drops off above birth year 2010, keep this in mind: kids aren’t diagnosed at birth. In the CDC study, 1/2 of the kids didn’t get diagnosed until after 52 months (4 years 4 months) of age. For this type of reason, the more recent study didn’t count kids under 3 years of age.

People are very fond of graphing autism prevalence data from various years and claiming these are accurate, full counts of autism prevalence (they aren’t) and, from that, claiming an epidemic. Here are the CDC data:

The numbers go up. Steadily up. I’d have to be a total denialist to not see that as evidence of an epidemic, right?

Consider this–the CDC autism prevalence for birth year 2000 is 1.1%. The study just out gives an autism prevalence more than double that (2.9%). For the same birth year. Both are good studies, for what they are. Both are limited. But, for one thing, the CDC study was performed in 2008. 10 years ago. Since then a lot has changed. For one thing, the kids got older and had more chances to get diagnosed. They didn’t just suddenly become autistic in the past 10 years.

So, yeah, we have an autism prevalence estimate of 2.5%. I’m not surprised and I’m not taking this as evidence of an epidemic.

The unfortunate thing in this discussion is that with all the work in this study, all the potential for advocacy, the only number that usually gets discussed is the overall prevalence. Watch the video abstract (which I can’t get to embed). One of the authors goes into a lot of detail about the other findings. Findings I hope to discuss soon.

//players.brightcove.net/1327978102001/rkA3rSifl_default/index.html?videoId=5839990273001
—
By Matt Carey

One line from the CDC Autism Prevalence report you will likely never see quoted

28 Apr

The CDC came out with an autism prevalence estimate a few days ago. There have been a number of news stories on the subject and the usual attempts by credulous websites to use this to claim that vaccines cause autism.

It’s right there at the top, in the interpretations section of the abstract:

Because the ADDM sites do not provide a representative sample of the entire United States, the combined prevalence estimates presented in this report cannot be generalized to all children aged 8 years in the United States

The Age of Autism blog (as noted already, always a good place to look for people getting it wrong on autism) ran a piece “Breaking News: 1 in 59 children Born in 2006 have Autism, 1 in 36 between the ages of 3 and 17. What’s going on?” Because, you know, claiming an epidemic is in their mission statement.

SafeMinds, another organization promoting the failed “vaccine-induced-epidemic” idea of autism wrote:

Baltimore, MD, April 26, 2018 – SafeMinds, along with other national autism advocacy organizations, sent a letter today to Department of Health and Human Services (HHS) officials demanding a meeting within the next 30 days to discuss the creation of a Federal Autism Strategic Plan to address the nation’s autism crisis. The urgent letter follows the release of a report this afternoon by the U.S. Centers for Disease Control (CDC). The CDC report found that autism is now diagnosed in one in every 59 American children, representing 2 ½ times more autism in 12 years and a 15 percent increase in just two years.

First, there is an Autism Strategic Plan. A member of SafeMinds helped craft it with HHS while she was on the Interagency Autism Coordinating Committee. She voluntarily decided not to pursue another term on the IACC.

Second, your reason is that ” The CDC report found that autism is now diagnosed in one in every 59 American children”. The CDC told you explicitly in the first few paragraphs that can’t be said.

But I do appreciate that you are implying no epidemic among children. You clearly state that the rate is the same for all American children, not just the 8-year old children of the CDC study. Or did you miss that important point?

So, good luck with that letter. I’m sure your readership will not notice the problems with your logic, but HHS will.

There are more examples, but these make the point.

—
By Matt Carey

Did autism prevalence increase by 20%? (answer: no)

28 Apr

The CDC came out with a new autism prevalence estimate yesterday. Their estimated autism prevalence is 1.68%.  That’s up from the estimate from 2 years ago (1.46%), but lower than a different recent study (2.76%).

I’m going to discuss some minor-league shenanigans.  It’s no surprise that some groups abuse facts and cherry pick data to make political points.  In this case it was useless.  there was no need to cherry pick.  I’ve done a lot of exposing the abuse of facts by some so-called autism advocacy groups.  I don’t get any joy from noting that people in my community are both dishonest and ignorant of science.  But much as that bothers me, it pales in comparison to the lost opportunities.  Millions of dollars were spent on this prevalence estimate alone, but all some groups do is dig for reasons to justify their “epidemic” story and push the long-ago failed idea that vaccines are to blame.

Over the years, news organizations and autism groups have jumped at the chance to put their spin on each new CDC autism prevalence estimate. Over the years I’ve spent a lot of time trying to unspin the takes of groups like the Age of Autism blog. AoA is, well…if you want to see autism done wrong, read the Age of Autism blog. Science, medicine, human rights and dignity, they can be counted upon to get things wrong.

With that in mind I decided the CDC announcement was a time to break my current trend of ignoring AoA. I literally haven’t read it in months. What I saw was both surprising and typical for them. Surprising in that they didn’t even bother to write their own pieces for the new prevalence numbers. Seriously–they just copied a couple of articles from other sites and left it at that.

By the way–this new low-level interest in autism prevalence appears to me to be a general trend–beyond just AoA. Autism is losing it’s status as having intense focus from the US media and public.  The chance to leverage the public’s interest into meaningul change is waning. That’s a far more interesting topic and far more important than this and I want to come back to it in the future.

So, what did AoA post? Big League Politics BREAKING: CDC Reports 20 Percent Autism Increase In Children and Breaking News: 1 in 59 children Born in 2006 have Autism, 1 in 36 between the ages of 3 and 17. What’s going on?. That first one caught my eye. First because in absolute terms, the autism prevalence increased by 0.22%. In relative terms, that’s about a 15% increase.  They could have said 15% and made their point, but why do that when 20% is even more scary?  More to the point–digging just a little into this 20% figure gives understanding that a true advocacy organization could use.

So how did they come to the idea that the autism prevalence increased by 20%? They skipped to page 13 and took one part of one paragraph out of context and, well, cherry picked. Completely unsurprising. They skipped over pages of data showing that we are failing to identify–and, therefore provide adequate services for–autistic minority children.

Sadly, AoA and their allies have spent over a decade denying this huge issue.   The reason is obvious: it doesn’t fit into the “vaccines cause autism” narrative.  So rather than push for better identification and better services for minorities, they’ve sacrificed these communities in for their political message.

It’s disgusting.

They didn’t dig any deeper into these numbers, even though the data were right in front of them. Also completely unsurprising. If they were the sort of people who dig into data and question, they wouldn’t be pushing the idea that vaccines cause autism.   Seriously.

Here’s the section they are relying upon:

Among the six ADDM sites completing both the 2012 and 2014 studies for the same geographic area, all six showed higher ASD prevalence estimates for 2012 compared to 2014, with a nearly 10% higher prevalence in Georgia (p = 0.06) and Maryland (p = 0.35), 19% in New Jersey (p<0.01), 22% in Missouri (p = 0.01), 29% in Colorado (p<0.01), and 31% in Wisconsin (p<0.01). When combining data from these six sites, ASD prevalence estimates for 2014 were 20% higher for 2014 compared to 2012 (p<0.01).

So, if you only pick the states where there were data in both this study (called 2014 because that’s when the data were collected to be analyzed) and the previous (2012) study, you get about a 20% increase (click to enlarge).

But what happens if you ask “why?”  As in, why do these states show a larger increase than the entire group?

I put some numbers in red and bolded them for emphasis. Those are states with lower than average autism prevalences. Take a moment to think about that–the states with low autism prevalences are showing larger prevalence increases than other states. Start with a low number, and you are bound to get bigger percentage increases.

Missouri, Colorado and Wisconsin started out with very low identification rates.  They’ve improved their identification rates. What if someone were to ask, “how did they do that?  Can we use that elsewhere to serve more unidentified autistic children?”

I’m sure the cherry-picking “let’s find a reason to fit this into the vaccines-cause-autism story” groups have already focused on New Jersey in the above figure.  New Jersey stands out–their autism rate increased by about 19% and they have a much higher autism prevalence rate. New Jersey also stands out for another reason: they aren’t lagging in identifying black or Hispanic autistic children:

In New Jersey, there was almost no difference in ASD prevalence estimates among white, black, and Hispanic children. Estimates for Asian/Pacific Islander children ranged from 7.9 per 1,000 (Colorado) to 19.2 per 1,000 (New Jersey) with notably wide CIs.

Wouldn’t it be great if someone were to say, “how did New Jersey close the gap in identifying some minority groups?”

Groups advocating for a better future for disabled people would do that.  People trying to keep the failed “it’s all vaccines fault” idea alive won’t.

Some people will be confused, and understandably so, at this point.  These are “autism prevalence” numbers.  Doesn’t that mean that this is a complete and accurate count of the number of autistic kids in each location?  No, they aren’t.  The people at the CDC are working hard and doing a good job with the information they have.  But this is a review of existing records, not a test of each individual kid.  The CDC have access to medical and/or school records (it varies by location what sort of records they have).

Keep in mind–the CDC autism prevalence numbers aren’t “these are the absolute accurate numbers for how many kids are autistic”. We are missing identifying kids. Doctors are missing autistic kids.  Schools are missing autistic kids. The CDC tries to make up for this by reviewing the records to determine which kids are autistic (i.e. they don’t just count existing diagnoses in the records–they “diagnose” from the records).  But they still miss kids in their counts. We need to get better. The fact that New Jersey isn’t leaving minority kids behind is huge.  The fact that lagging states are catching up is huge.

Doing things right–checking on numbers is hard. It takes time. AoA not only doesn’t take the time to be careful (which is typical for the “vaccines cause autism” groups), they know that it’s the simple message that scares people. “Autism increased by 20%”!!!! It may only take 5 minutes to actually dig and see what’s going on. But that’s five minutes most people don’t have time for.

Which is the long winded way to say:

A lie can travel halfway around the world before the truth can get its boots on

Groups like AoA live and breathe on taking advantage of that.  They were fake news long before the term was coined.

That said, I knew with the new autism prevalence numbers people would misuse them. It’s pretty obvious.  They’ve used that misinformation to drag parents into a world of guilt and shame for being part of vaccine injuries that, in at least the vast majority of cases, didn’t happen.  Even more, the autism-is-a-vaccine epidemic community had the resources for over a decade to make a real difference in the lives of autsitic people. As we watch autism fade a bit from the public’s view, it’s difficult to not cry a bit for the lost opportunity. It’s that, not their failed logic or twisted facts that is why I wrote the above.

—

By Matt Carey

Tags: ,

Americans are still failing to identify and serve minority autistic children

27 Apr

The CDC recently published another autism prevalence study. It’s 23 pages long and has 26 authors, took 2 years to put together and no doubt cost millions of dollars. Out of that, the one fact from it that will be quoted is simply–the autism rate is now at 1.68%, or 1 in 59.

There’s so much more. But sometimes focusing on one simple message makes more impact than a lengthy analysis. So I’ll pick my own simple message (of my own):

we are failing to identify minority autistic children. And with that, we are failing to provide them the appropriate services and supports they deserve as citizens and residents of the U.S.

We can and we should do better.

Here is table 3 from the report:

The estimated autism prevalence for Hispanics is 1.4%. For Whites, it is 1.7%. Thousands of Hispanics and other minorities are being missed. Overall, thousands of autistic children, and many, many more adults, are being missed. But that’s another discussion.

—

By Matt Carey

Tags: , ,

If you are using California data to claim an autism epidemic, you’re doing it wrong. Or:The great anti-epidemic of intellectual disability in California.

22 May

If you’ve been reading about autism online, you have almost certainly read that autism “rates” are on the rise. But what if I told you that here in California intellectual disability has been dropping for over 20 years?

For many years the mainstay of the “autism is an epidemic” idea was the California Department of Developmental Services data. The CDDS keeps track of how many Californians are getting support under a number of specific disability categories. These data are publicly available (although not as easily available in the past), which makes them an easy source of data.

It’s easy to take a cursory look at the CDDS data and think “these are official data. Look at how much autism has increased!” Or claim “the CDDS only serves “severe” autism, there’s no way they were missed in the past.” You can even find a few publications to cite to back up these observations.

About a year ago I asked CDDS for some data. I hadn’t checked in a while and I wanted to see what trends are ongoing. Coincidentally, the Autism Society of San Francisco put out a report shortly after that: Autism Rising, A Report on the Increasing Autism Rates in California. So I was not alone in asking for data.  The Autism Society of San Francisco made the argument that the CDDS data are accurate and show an epidemic.

The Autism Society of San Francisco graph the data in many different ways, but the one that was closest to the way I was looking at the data was in Figure 5 (click to enlarge):

AS-SF Autism Rising Figure 5

and here is the caption for Figure 5:

Births of individuals later deemed to have DDS-eligible autism have been increasing sharply every year since the early 1980s. Typically intake into the system occurs between 2 and 7 years of age. The data reflects about 200 DDS autism births per year into the 1980s, but now the system is reflecting nearly 5,000 such births per year. The drop off in cases after birth year 2008 is likely attributable to usual delay in cases entering the system, and likely does not represent an actual decrease in DDS-eligible autism cases.

You can stop there and support your argument. And that’s just what most people do. Or you can question–how can I test if this is a “real” autism increase? For example, is the autism rate the same among different races? The answer is no. Is the autism rate the same in, say, San Francisco, Los Angeles, and Kern County? The answer is no. And there are many more questions one can ask of these data and over and over, the answer is no.

Either we aren’t counting all the autistics in our state, or there is something much more complex going on than vaccine, toxins, epigenetics, or whatever the claimed causes of the rise are. And I’ve gone through many of these discussions over the years. Let’s make this simple then. If one claims that the CDDS counts everyone within each disability category accurately and that the definitions they use aren’t changing with time, why is intellectual disability (mental retardation) dropping so fast in California?

You see I also graphed intellectual disability. I got autism counts, intellectual disability counts and “unduplicated” (total, each disabled person counted once) by birth year. I also got census data by birth year. And I graphed them. And anyone claiming CDDS data show an autism epidemic needs to do the same and to explain this graph, complete with the sharp peak for birth year 1993. (click to enlarge):

CDDS including ID

Intellectual disability has dropped. Off about 40% of the peak value.

If you think your idea for the rise in autism is correct, let’s take the failed vaccine idea as an example, you need to also explain how that resulted in far fewer people with intellectual disability. Plain and simple. And none of these claimed causes of an “epidemic” can explain the drop in ID.

Why bother challenging the people claiming an autism epidemic? Because it denies the existence of undiagnosed autistic adults. We have very little effort to identify those who were missed in past generations. And the likelihood is that these people–our people–are not being supported appropriately because of their misdiagnoses. And not only are we abandoning the misdiagnosed, we are failing to learn. What worked for past generations, the adults of today? What failed? What are the appropriate supports for the various needs of autistic adults? We don’t know today. And are unlikely to know by the time my kid is an adult, especially if we aren’t even looking at autistic adult needs today.

And then there’s the whole autism causation question. People spending their time trying to correlate CDDS data–data clearly confounded by numerous social influences–are unlikely to ever find a real answer.

But, for those who want to keep trying, include all the data. Give an explanation for this and you may be on to something.

CDDS including ID
—
By Matt Carey

U.S. announces 2% autism rate. Again. And it’s dependent on a change in the way a survey is worded. Again.

17 Nov

Last week a study was released showing an autism “rate” in the U.S. of about 2% (an estimated prevalence of 2.24% to be exact). Luckily one of the best science journalists out there focuses a lot of her attention on autism and covered this story. I’m writing of course about Emily Willingham and her article Increase In Autism Diagnoses Not An Increase In Autism.

The study in question is Estimated Prevalence of Autism and Other Developmental Disabilities Following Questionnaire Changes in the 2014 National Health Interview Survey. The abstract is at the bottom for those interested.

While other articles are discussing the large “increase in autism”. Or, as in this Reuters article, US autism numbers soar in which we read what must be a confusing message to many:

The results reflect a near doubling in autism rates over the past three years, but the US Centers for Disease Control, which released the data last week, says the shift is largely due to a change in the way the survey was worded.

Here’s the thing, before I saw what the study was, I actually thought that the discussion was about this study: Changes in Prevalence of Parent-reported Autism Spectrum Disorder in School-aged U.S. Children: 2007 to 2011–2012. In that study, released in 2013, a 2% autism “rate” was announced. And that was a significant change from the previous survey data (using the National Survey of Children’s Health). And like previous estimates based on survey’s the change is at least due in large part to a change in the way the survey was presented to parents (I’m looking for a link to the discussion of the change in survey questions that affected the 2013 estimate, but there was a shift and it had a big impact).

Thankfully much of the media, and even Autism Speaks, are pointing out how this 2% figure doesn’t represent a “real” change in autism prevalence, but is largely dependent on the study methods. Others can be counted on to shout “epidemic” and ignore some of the key reasons why this doesn’t reflect a real increase.

Number 1 reason–the autism prevalence is basically the same for kids who were 3-10 and kids who were 11-17:

AutPrevalence

A real increase would show up as a higher prevalence for younger kids.

One thing we see is a decrease in disparity for autism prevalence by race/ethnicity. African Americans have a slightly lower autism prevalence (it’s unclear whether this is statistically significant or not), and Hispanics have a decidedly lower autism prevalence (1.49% vs. 2.55% for Caucasians). Which means we still have far to go to identify and bring services to all autistics. And it also means that the autism prevalence estimates will continue to rise as we do identify more in under diagnosed populations.

AutPrevalence race-ethnicity

But let’s focus on the important part of that–we still have a large under-diagnosed/under-served population and it is largely among Hispanics. We really need to be focusing more attention on remedying that situation.

Here is the abstract for the current study:

Objectives—The developmental disabilities questions in the 2014 National Health Interview Survey (NHIS) were changed from previous years, including
question reordering and a new approach to asking about autism spectrum disorder (ASD). This report examines survey-based estimates of the lifetime
prevalence of ASD, intellectual disability (ID), and any other developmental delay (other DD) following the inclusion of a standalone ASD question, the
inclusion of specific diagnoses in the ASD question, and the ASD question preceding the other DD question, and compares them with estimates from
previous years.

Methods—In NHIS, one child is randomly selected from each family to be the subject of detailed questions on health conditions, functional limitations, and
health care utilization. Parents are asked if a doctor or health professional had ever told them that their child had each of a series of developmental disabilities.
Prevalence estimates of ASD, ID, and other DD for children aged 3–17 years were calculated using data collected in 2011–2014.

Results—The estimated prevalence of ASD based on 2014 data was 2.24%, a significant increase from the estimated annualized prevalence of 1.25% based on
2011–2013 data. In contrast, the prevalence of other DD declined significantly from 4.84% based on 2011–2013 data to 3.57% based on 2014 data. The prevalence of ID did not significantly change from 2011–2013 (1.27%) to 2014 (1.10%). The prevalence of having any of the three conditions was constant across survey years.

Conclusions—The revised question ordering and new approach to asking about developmental disabilities in the 2014 NHIS likely affected the prevalence estimates of these conditions. In previous years, it is likely that some parents of children diagnosed with ASD reported this developmental disability as other DD instead of, or in addition to, ASD. Following these changes, the 2014 ASD estimate was more similar to ASD prevalence estimates from other sources.

—
By Matt Carey

New study on inflammatory bowel disease and autism: Prevalence of Inflammatory Bowel Disease Among Patients with Autism Spectrum Disorders.

12 Aug

People with developmental disabilities often have additional medical issues at rates higher than the general population. For example, heart problems are more common in the Down Syndrome population and Timothy Syndrome. Hip dislocation is common among those in the Fragile X community. Mental health conditions and neurological disorders are very common in autistics (but somehow those are rarely mentioned in discussions of autism and comorbidities).

When it comes to autism parents online, perhaps the most talked about autism comorbidity is gastrointestinal disease. And, in specific, inflammatory bowel disease. This is a lasting legacy of Andrew Wakefield’s attempt to link the MMR vaccine and autism (an effort which set back work on autism and GI disease by a decade or more–see Blame Wakefield For Missed Autism-Gut Connection).

Mr. Wakefield’s assertion was that the MMR vaccine leads to a unique form of IBD (he dubbed it autistic enterocolitis, a condition which doesn’t appear to exist) and this somehow leads to autism. The model also asserts that autism rates have climbed with the introduction of the MMR in the UK (an argument that fails when when considers when the MMR was introduced in the U.S., but I digress). Given the Wakefield model, including the claim that the MMR has played a major role in the “autism epidemic”, we would expect a large fraction of autistics should have IBD.

With apologies to autistics with IBD for taking so long on this introduction–this all begs the question of what is the prevalence of IBD in the autistic population? Well, a recent study discusses this:

Prevalence of Inflammatory Bowel Disease Among Patients with Autism Spectrum Disorders.

Before we get to the prevalence let’s consider the important points. First–IBD does exist in autistics. Given communication issues and sensory issues, any medical condition is serious in the autistic population. Second–IBD is more prevalent in the autistic population. What this may say about the biology of autistics and the developmental trajectory is not discussed in the abstract of this study.

Finally let’s ask how big is the prevalence of IBD in the autistic population? The study looked at two sample populations. In one population 7 out of 2728 (0.26%) autistics had IBD. For another, 16 of 7201 (0.22%). Just because the prevalence is small doesn’t mean this isn’t an important issue for the autism communities. But, let’s face it, the claims of high and rising IBD prevalence in the autism community–the claims by Mr. Wakefield to support his attack on the MMR vaccine–are just not true. And, yes, this also means that people who think that all or most autistic kids should be treated for IBD are also not doing a service. Yes, treat people with IBD. But no, don’t assume autism = person with IBD.

The fact that IBD is not that common in autistics is not really that new. I recall the press conference for the MMR/autism study by Hornig et al.. One thing that slowed the study was the fact that there weren’t that many autistic kids whose symptoms really indicated the need for a colonoscopy. Contrary to some practitioners who seem to believe that all autistics should be ‘scoped.

Here’s the abstract from the study:

Background:
The objective of this study was to measure the prevalence of inflammatory bowel disease (IBD) among patients with autism spectrum disorders (ASD), which has not been well described previously.

METHODS:
The rates of IBD among patients with and without ASD were measured in 4 study populations with distinct modes of ascertainment: a health care benefits company, 2 pediatric tertiary care centers, and a national ASD repository. The rates of IBD (established through International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] codes) were compared with respective controls and combined using a Stouffer meta-analysis. Clinical charts were also reviewed for IBD among patients with ICD-9-CM codes for both IBD and ASD at one of the pediatric tertiary care centers. This expert-verified rate was compared with the rate in the repository study population (where IBD diagnoses were established by expert review) and in nationally reported rates for pediatric IBD.

RESULTS:
In all of case-control study populations, the rates of IBD-related ICD-9-CM codes for patients with ASD were significantly higher than that of their respective controls (Stouffer meta-analysis, P < 0.001). Expert-verified rates of IBD among patients with ASD were 7 of 2728 patients in one study population and 16 of 7201 in a second study population. The age-adjusted prevalence of IBD among patients with ASD was higher than their respective controls and nationally reported rates of pediatric IBD.

CONCLUSIONS:
Across each population with different kinds of ascertainment, there was a consistent and statistically significant increased prevalance of IBD in patients with ASD than their respective controls and nationally reported rates for pediatric IBD.

—
By Matt Carey

← Older Entries