1) Back in July, Minneapolis Public School officials had furnished data on autism in Somali and non-Somali speaking students for the 2007-2008 school year to journalists, parents, and the Minnesota Department of Health (HERE). 

2) Officials at the CDC were scheduled to listen to Dr. Punyko present “findings” on a possible “Somali autism cluster” at the CDC in Atlanta on Tuesday, November 18. It seemed odd to me that Punyko would be discussing her findings at the CDC, but did not mention that fact three days earlier at the Somali autism forum in Minneapolis. It also seemed odd that she would be presenting “findings” in Atlanta, when she told us that MPS would not be furnishing any data to her team until December.

And so, I wrote to the Minnesota Department of Health and to the CDC to see if I could clarify any of this.

Initially, I got this reply from Doug Shultz, of the MDH Communications office:

“Dr. Punyko was invited by CDC to participate in a grantee meeting of other states that receive CDC grants to implement autism surveillance systems (although MN is not a grantee).  This was an opportunity for Judy to learn more about what other states are doing to address how best to gather data about autism. She will not be presenting any data from Minneapolis. Please let me know if I can be of further assistance.”

This was not consistent with information I was getting from sources within HHS, so I persisted, and wrote back for further clarification.

I am very happy to report that both agencies -- MDH and CDC -- responded with notable courtesy, respect, promptness and, I feel, thoroughness. And for that I am very grateful to them.

Because they obviously took the time and trouble to answer my questions thoughtfully, and in detail, I have decided to publish the exchanges, verbatim, below, with just a few key observations. (Thank you for your patience as you wade through it – I think that it provides everyone with invaluable insight into the thinking and working of government epidemiologists):

A) The state and federal governments clearly take this situation, and the concerns of the Somali parents, seriously. They are not trying to sweep this under the rug or make it just go away. They know they have a situation on their hands – one that could potentially be hugely significant – but they are proceeding with caution.

B) Even so, public health officials are not at all convinced that they’re dealing with a Somali autism “cluster” in Minnesota.

C) It is difficult to assess autism prevalence in such a “fluid” immigrant community as Somalis, according to state and federal officials.

D) It is also difficult to know exactly how many Somali students in MPS actually have an autism diagnosis, as school services are based on educational evaluations, and not medical analyses.

F) It’s possible that the number of Somali students with autsm is artificially high. For example, there are no Somali students in Rochester, MN – did their families migrate to Minneapolis-St. Paul for better services after receiving a diagnosis?

G) It’s equally possible that the number of Somali students with autism is artificially low. For example, some Somali families listed English as the primary language spoken in their homes, meaning their children with autism would not be counted as Somali. There is also evidence to suggest that Somali parents of children with milder cases of autism (where the child is speaking) will reject an ASD classification or services for their child.

H) Even if autism rates are the same as non-Somalis, they are exponentially higher than autism rates among children in Africa (See Mark Blaxill's piece HERE.)

I) Finally, the CDC response includes one of the most extraordinary statements on autism that I have ever seen from a federal agency, and it would appear to shut the door on the old 20th Century viewpoint that autism is “strictly genetic,” and that increased numbers could therefore only be due to better diagnostics and greater awareness. To wit:

“There are likely multiple causes of the autism spectrum of disorders. Most scientists agree that today’s research will show that a person’s genetic profile may make them more or less susceptible to ASDs as a result of any number of factors such as infections, the physical environment, chemical exposures, or psychosocial components.”

THE RESPONSE FROM MDH

Here are my questions to MDH Officials, and the answers that were provided by Buddy Ferguson, Risk Communication Specialist, who wrote:

"The questions that you e-mailed on Wednesday to Doug Schultz raise a number of important issues, so we wanted to respond to them individually, and also offer some broader observations. I have been asked to respond on behalf of Doug and the MDH Community and Family Health Division."

1) Why did Dr. Punyko say at the autism forum on Saturday that she had not received data about Somalis with autism in the Minneapolis schools, when this apparently was not the case?

2) What other data is Dr. Punyko waiting upon from local public education officials (figures she said she would not have until sometime in December?)

Items #1 and #2: In terms of Dr. Punkyo’s presentation at the forum on Saturday, she did not mean to imply that she had received no information from the schools, but simply that her information is incomplete. When she presented at the forum, she was still awaiting information about the place of residence of the students who are receiving special services from the school system, and the classification of students who receive those services. For reasons I will describe below, this information is important if we are to develop a clear picture of the issues we are facing with regard to the Somali community and Autism Spectrum Disorder.

3) What possible explanation is there for the high rates of autism among Somalis in M/SP, but zero cases reported in Rochester schools?

Regarding the difference in reported autism cases in Rochester and Minneapolis/St. Paul, right now we do not have the kind of statewide data that would allow us to make that comparison. We do not have a statewide autism surveillance system in Minnesota, and setting up such a system is a complex undertaking. Simply relying on data regarding which and how many students are receiving special services is not a substitute for a good surveillance system * again, see below for a fuller explanation.

4) Why didn’t Dr. Punyko mention at the forum that fact that she was travelling to Atlanta to take part in a grantee meeting for autism surveillance in a number of states? (This would have been pertinent and welcomed news - it seems odd she would not have told us about it).

If Dr. Punkyo neglected to mention her pending trip to CDC, it was because her participation in the national meeting was simply part of the ongoing, informal conversation she has been having with CDC regarding the issue of autism in the Somali community.

This was a meeting held primarily for the 14 states that have received CDC grants to develop autism surveillance systems. Those states do not include Minnesota. Judy was simply extended an informal invitation to attend for one day of a three day meeting, so she could benefit from the discussion, get advice from the other states, and briefly share what she knew about the situation in our state. She was not providing them with any information not already available to forum participants (a copy of her PowerPoint is attached), so there was little reason to share her travel plans.

5) I could find nothing on the CDC website or elsewhere about the November 18 meeting in Atlanta. What information can you provide to me regarding the autism grantee meeting at the CDC on November 18, 2008. Specifically, can you provide me with any of the following?:

A) Invitation to attend the meeting that was received by Dr. Punyko.
B) An agenda for the meeting.
C) A list of attendees and states represented.
D) Minutes from the meeting.
E) Copies of Dr. Punyko’s notes from the meeting.

Item #5: In additional to the PowerPoint, we are also sending you a copy of the e-mail messages sent and received regarding Dr. Punkyo’s participation in the CDC conference. Again, this was very informal - the invitation was handled via e-mail. (HERE) - I am also attaching a copy of the conference agenda and Dr. Punkyo’s notes, in a single PowerPoint file HERE. We do not have meeting minutes or participant list. You will need to request them from CDC, although my understanding is that CDC may not have minutes available for this type of event.

(NOTE: The meeting agenda is HERE Dr. Punyko was scheduled to speak at 1:10 PM on “ASD among Somalis in Minnesota.”

6) It seems logical that Dr. Punyko would have shared the Somali autism data (attached) with authorities at the CDC at some point by now. Did she discuss this information with anyone at CDC prior to the meeting on November 18, and/or did she discuss autism numbers among Somalis in Minnesota with CDC officials at any point during her trip to Atlanta this week -- either during the official proceedings, or at some other time during her visit?

As previously noted, Dr. Punkyo has been sharing information about autism and the Somali community with CDC, on an informal basis, for several months now. Her PowerPoint should give you a sense of what she discussed at the conference session on Tuesday. However, she didn’t necessarily share information about the special education classification of Somali students in precisely the same format that you provided it to us.

The critical thing to understand is that this information doesn’t necessarily provide a clear picture of autism in the Somali community - again, for reasons discussed below.

7) If it is determined that there is a Somali autism “cluster” in Minnesota, what happens next? Is the CDC required by law to investigate?

There is no specific legal “trigger” that would require a particular action in response to an identified problem regarding autism in the Somali community. We will continue to investigate this issue with the resources we have available - which are, unfortunately, limited. We will continue to work closely with CDC, seeking their assistance when appropriate, as we proceed with our investigation.

Identifying the Problem.

In addition to answering your specific questions, we also want to offer a couple of larger observations about the task that now faces us. What we have right now is simply classification data used by the public schools in assigning children to receive special education services. This is not the same as diagnostic data: The children who have been classified in that way may or may not have received a medical diagnosis identifying them as autistic.

In fact, some of the children so identified in the summary you provided may not be autistic, although they may be facing other kinds of challenges. There may also be autistic children who do not show up in that data because they do not receive special education services through the schools, but are receiving services from some other source. 

Even if we did have a complete and comprehensive picture regarding the number of autism cases in the Somali community, it would still be difficult to calculate an “autism prevalence rate” for this population. The Somali population in Minnesota is large and highly fluid. Large numbers of Somalis are migrating to the state all the time. In fact, we have the highest “secondary immigration rate” in the country - that is, we are the top destination for Somalis who first settled somewhere else when they initially entered the country. Estimates from the state demographer’s office have placed Minnesota’s Somali population at anywhere from 15 to 40 thousand.

In short, in terms of calculating the actual prevalence of autism in Somali children, we have neither a reliable numerator nor a reliable denominator. While we appreciate your diligence in attempting to calculate observed versus expected autism rates based on the school classification data, those calculations probably obscure more than they reveal.

The Larger Picture

It should be emphasized that we are not presuming to second-guess parents and others in the Somali community regarding the seriousness of this problem. The situation they are dealing with is very real. We know that this is frustrating. We wish that we could provide quick and easy answers - but we lack basic knowledge about autism, about how common it really is, and about what causes it. We recognize how frustrating this is for a concerned parent - it’s frustrating for us as well. However, before we can answer the big questions, it’s important to get the basics right. That’s what we’re working to do right now.

THE RESPONSE FROM CDC

Here are my questions to CDC officials, and the answers that were provided by CDC Director’s Office of Enterprise Communication (OEC):

1) I have been sent a copy of autism prevalence data generated by the Minnesota Department of Education (LINK #1). This information, by several accounts, was given to Dr. Punyko in July. Has she ever shared this information with anyone at CDC in the past?

Keep in mind that the Minnesota Department of Education data reflect an educational system classification that provides information on the number of students enrolled in special education programs under an autism eligibility and are not complete prevalence data.  Certainly, it is concerning for all families if their child is identified with autism, either for special education purposes or through a formal diagnosis, and it is important to determine if one group is disproportionately affected.  In order to understand if children from Somali families in Minneapolis are disproportionately affected by autism, it would be necessary to evaluate the available data and to determine how complete these data are and what would be needed to have the most complete and accurate accounting of children with autism in the Somali and other comparison groups of children.  It is our understanding that the MN Department of Health is first working to verify the education data, including trying to evaluate birth place and potential moving districts.  This is an important first step.

With respect to your question, to date, CDC scientists have only provided informal support to the Minnesota Department of Health and Dr. Punyko.  This has involved answering questions related to measuring autism prevalence and helping connect Dr. Punyko with experts in other states.  These informal conversations began earlier this year when Dr. Punyko contacted CDC to inquire about methods that could be used to set up a state-wide autism surveillance or tracking system in MN based on wanting to get a better understanding of who is affected with autism in the state more generally. 

To date, Dr. Punyko has shared some of the basic educational system reporting information from the Minnesota Department of Education with Dr. Catherine Rice, (Behavioral Scientist/Epidemiologist) and other scientists working on developmental disabilities prevalence projects in CDC’s National Center on Birth Defects and Developmental Disabilities.  More generally, over the past year, Dr. Rice has provided information on CDC’s Autism and Developmental Disabilities Monitoring (ADDM) Network (http://www.cdc.gov/ncbddd/autism/addm.htm), which is the CDC autism surveillance and tracking program.  Dr. Punyko has shared that MN is working on preliminary data analysis and sent a letter to Dr. Coleen Boyle last week requesting CDC's input on the report when it was prepared and providing input on the next steps, including evaluating whether more formal assistance from CDC would be helpful. 

Following a July media report about the number of Somali children enrolled in autism special education programs in the Minneapolis public school system, Dr. Rice and other CDC scientists have provided information about research methodology challenges involved in identifying and measuring autism prevalence, and have provided suggestions on people to talk with in other states about conducting autism prevalence studies.  In line with this, Dr. Rice invited Dr. Punkyo to a November meeting of CDC grantees involved in autism surveillance.  This annual meeting brings together researchers/principal investigators from various past and present CDC-funded sites to share information and coordinate efforts related to autism measurement and surveillance.

Dr. Rice has also provided information on how states can seek formal assistance from CDC in investigating public health issues.

3) It seems logical that Dr. Punyko would have shared the Somali autism data with authorities at the CDC. Did she discuss autism numbers among Somalis in Minnesota with meeting participants and/or CDC officials at any point during her trip to Atlanta this week -- either during the official proceedings, or at some other time during her visit?   

State and local educational classification data are not routinely shared with people at CDC.  There are also no legal or formal requirements that education system classification data, even if related to autism or the provision of educational services to who have been diagnosed or placed in autism-related programs, be provided to CDC.  To the extent there are legal or formal reporting requirements for education data, it is to the Federal Department of Education.  For states, these educational data are available on the public website (www.ideadata.org).  CDC experts may become aware of information or reports through informal contacts, formal requests or agreements for technical or scientific assistance or presentations at meetings or conferences.

Dr. Punyko was invited to attend the ADDM meeting this week in Atlanta as a way to help connect her with other state-based investigators who have been working to determine the prevalence of autism or on projects designed to help better define the characteristics of affected children.  She did provide a brief overview of the initial education data reported in the media indicating a concern about autism in Somali children and the general outline of the work that has been done to follow-up on these concerns in MN to date.  She did not present any additional data. Dr. Punyko did share that there are particular challenges with trying to determine accurately what the prevalence of ASD is in young Somali children (as well as for any group of children born to a potentially mobile population) in Minnesota.

4) Does CDC have any comment at all on the MPS provided data?

Dr. Catherine Rice and Dr. Marshayln Yeargin-Allsopp both agree the educational tracking system information that has been shared with them merits further assessment.  It is important, for instance, to determine if all children in need of specialized educational services are being identified and that identified children are getting the services they need. It is our understanding the MN Department of Health is working within the state to follow-up on these concerns.

5)  If it is determined that there is a Somali autism “cluster” in Minnesota, what happens next? Is the CDC required by law to investigate (perhaps because of EPA Superfund requirements)?   

Much work will need to be done to determine actual autism prevalence among children in Minnesota, children in the Minneapolis and St. Paul school districts and among Somali populations in Minneapolis, St. Paul, and Minnesota.  Without a statewide autism surveillance system, it will be very difficult to determine whether a group of children or a school district has an unusually high autism prevalence rate.  It is should be noted that setting up a statewide autism surveillance system is a complex undertaking.

With respect to formal CDC assistance, such involvement requires a state or local health department to send a formal written request.  As requested by Dr. Judy Punyko to date, CDC plans to provide feedback on the preliminary report by the MN Department of Health, when available, and to input on needed follow-up.

6) If it is determined that Somali children in Minnesota do in fact have higher rates of autism than non-Somali children in Minnesota, and that they also have higher rates than Somali children in Somalia, will CDC officially concur that autism, at least in these cases, must necessarily have an environmental component?

As noted in #5 above, determining actual autism prevalence rates, whether in a community, state or country, requires a valid autism surveillance system.  In this case, such systems do not exist in Minnesota or Somalia.

It is also important to note that even if one has valid autism prevalence estimates, those estimates do not provide much information or insight into the causes of autism.  For example, information about the number of children in a state who have autism does not tell you very much about the potential cause, or more likely, multiple causes —particularly since children move in and out of states as well as in and out of school districts.  A higher than expected rate in a school district may be caused, for instance, by the availability of a strong program for autistic children.  Also, other children in other groups, especially young children, may not yet be identified for special education services under autism so they are not counted accurately as a comparison group.

In an effort to better understand autism and autism prevalence, CDC has been working to better understand how the Autism Spectrum Disorders (ASDs) affect children in the United States trough the ADDM Network and Early ASD Surveillance Projects (www.cdc.gov/autism).  Our initial efforts have shown that, autism is being identified more often than in the past and there are an increasing number of efforts to identify affected children as early as possible so that interventions can begin.

In addition, CDC has worked with Autism Speaks to form the International Autism Epidemiology network (IAEN) (http://www.autismepidemiology.net).  In many other countries, including Somalia, awareness of autism is in the early stages.  We are not aware of any efforts to measure autism prevalence in Somalia and have not found any published baseline data on autism prevalence in Somali children from Somalia.  Without baseline data, it is not possible to make a comparison of autism prevalence between Somali children in the U.S. and Somali children in Somalia. 

Similarly, without an established autism tracking system in Minnesota, it is not possible to accurately compare rates of autism for Somali children in Minnesota versus non-Somali children in Minnesota.

Finally, while it is important to understand if autism is affecting any group of children disproportionately, it is also important to keep in mind that there are likely multiple causes of the autism spectrum of disorders.  Most scientists agree that today’s research will show that a person’s genetic profile may make them more or less susceptible to ASDs as a result of any number of factors such as infections, the physical environment, chemical exposures, or psychosocial components.  CDC researchers are currently working on one of the largest U.S. studies to date, called the Study to Explore Early Development (SEED).  The project is examining numerous risk factors for autism such as genetics, environmental exposures, pregnancy factors, and behavioral factors.  CDC is also supportive of the coordination of research efforts of the multiple government and non-government organizations involved in autism research organized through the Interagency Autism Coordinating Committee - IACC (http://iacc.hhs.gov).

We hope this information is useful to you in your effort to report on this complex issue.

David Kirby is author of Evidence of Harm and a contributor to Age of Autism.