The Somali Minnesotan Autism Epidemic

13 Apr

One of the most intriguing scenarios to pop up from the so-called autism epidemic happened in Minnesota in the US. It was noted that there were what seemed to be disproportionately high numbers of second generation Somali children in Minnesotan schoolrooms. Of course, this was immediately latched on to by the usual suspects, despite the caution issues by health authorities that until proper epidemiological studies had been done it would be impossible to say whether this really was a cluster worth investigation or just a coincidence.

On the 31st March, the Minnesota Dept Health released the study. Its a fairly substantial read. One thing immediately struck me about this:

Administrative prevalence of Somali children, ages 3 and 4, who participated in the MPS ECSE ASD programs was significantly higher than for children of other races or ethnic backgrounds

OK so this is what the families were saying. But there’s an extreme note of caution that should be noted here. This is basically CDDS all over again. Just like some believed that an increasing number of reports to CDDS meant that there was an autism epidemic and just as CDDS said there reports really shouldn’t be used to study these things, MDH are also saying:

Because of the study’s limitations, it is not proof that more Somali children have autism than other children…

This is a vital point. Back in 2005, James Laidler made the clear point that Department of Education data on autism are not reliable for tracking autism prevalence.

Sadly, the new Age of Autism editor, Abdulkadir Khalif, either misunderstood or elected to ignore this issue when he said:

It was obvious from the numbers that the issue of prevalence has finally been settled, and that there definitely is a cluster of autism in Minneapolis

It is clear from the report that Khalif has grossly overstated the case. Firstly, the issue of prevalence is far from settled. MDH seem to be solely using educational data which, as pointed out by Laidler, is not reliable for tracking autism prevalence. Indeed the phrase ‘administrative prevalence’ used by MDH reflects this. ‘Administrative prevalence’ refers solely to numbers of kids in educational programs. This is a clear distinction from ‘prevalence’ which is the proportion of individuals in a population who suffer from a defined disorder. Using only educational data gives a distorted picture.

As has been shown, the USDE data on autism are at odds with studies of autism prevalence, largely because the criteria used by the school districts (the source of the USDE data) to categorize children as autistic are neither rigorous nor consistent. They are inconsistent over time, as are the medical criteria, and are inconsistent from region to region. The USDE data are not reliable for tracking the prevalence of autism, and they in fact never were meant to fill this need.

Secondly Khalif uses the word ‘cluster’ whereas the MDH report does not use it at all. And it is not a word that should be used in such a throwaway fashion – it has a distinct epidemiological meaning.

So clearly, contrary to Khalif’s assertion that the issue of prevalence has been settled, it has not. Contrary to his statement that there is a cluster, there has been no such epidemiological assessment or statement.

Here are some more quotes from the MNH report that Abdulkadir Khalif either chose to ignore or never actually read:

The fact that a child is participating in an ASD early childhood program is an indicator of educational need, but that child may or may not have a medically diagnosed ASD.

Further, Minnesota’s public school open enrollment policy allows children to attend special education programs in school districts where they are not residents. This raised the question of whether participation rates for Somali children might appear higher than the participation rates for non-Somali children because of an influx of Somali children who are not residents of the Minneapolis school district attending MPS ECSE programs for ASD.

Data on variability of ASD prevalence by race, ethnicity, and SES is limited and inconclusive, and apparent differences between racial and ethnic populations may largely be due to differences in case finding and service provision.

Across all assumptions and ASD program types, administrative ASD prevalence estimates for Somali children were uniformly higher than the U.S. parental reported ASD prevalence, but most of the 95% confidence intervals corresponding to the administrative prevalence estimates for Somali children contained the value of the U.S. parental reported ASD prevalence estimate – suggesting that the 2005-2006 administrative ASD prevalence for Somali children might be no different from what would be expected in the U.S. population of children ages 3 and 4 based on parental report.

So what does this mean?

It means that there are no firm answers and that Khalif is simply wrong to assert that there are.

Its always been one of the great puzzles to me that a section of (mostly) parents who demand accurate answers fast cannot seem to understand that there _are_ no accurate answers until the science – proper science – has been done. And that takes time. What legacy do these parents want to leave the autism community? Fast inaccurate mistakes? Or well planned and rigorous science that helps build the growing knowledge we already have?

5 Responses to “The Somali Minnesotan Autism Epidemic”

  1. Joseph April 13, 2009 at 22:21 #

    Once we’re in agreement this is only administrative prevalence, though, it’s still valid to ask why the administrative prevalence of autism in Somali children might be different, isn’t it? It’s similar to the question of why the administrative prevalence of autism in Hispanic children is much lower than it should be in some parts of the US (something Sullivan mentions from time to time.)

    Here’s the thing. The prevalence shown in the report (1.0% to 1.5% just about) is not high compared to the IDEA prevalence of autism in Minnesota. Minnesota probably has the highest administrative prevalence of autism in the US for children 6 or older. Their prevalence of MR is about half that of autism.

    What’s unusual is that these are 3-4 year old children. Normally, you shouldn’t have that high a prevalence in this cohort. The consensus prevalence they are referring to is not based on surveys of 3-4 year old children.

    It would be interesting to see if this prevalence increases once these same children are 7 or 8. If it doesn’t, the explanation is pretty clear: they were recognized earlier than most.

  2. Joseph April 13, 2009 at 22:22 #

    I thought the new login system would help reduce the number of comments caught by the spam trap. I guess it didn’t 🙂

  3. Kev April 14, 2009 at 10:20 #

    Once we’re in agreement this is only administrative prevalence, though, it’s still valid to ask why the administrative prevalence of autism in Somali children might be different, isn’t it?

    Absolutely – couldn’t agree more 🙂

  4. Amanda December 21, 2010 at 15:40 #

    I am trying to reach abdulkadir khalif. If someone has the email contact for him please contact me.


  1. Thanks for the measles yet again, Andy – Respectful Insolence - May 1, 2017

    […] sure whether actual autism prevalence was higher, not the least of which was that as was noted in Left Brain, Right Brain, Department of Education data are not reliable for tracking autism. Jim Laidler made the same point […]

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: