Do we feel lucky? Autism and the sibling question

We’ve had a few days to absorb the ramifications of the latest study regarding autism incidence rates amongst siblings, and now we’re all wondering whether to have another child (luckily for me, I had another child before these horrendous data were published - sometimes ignorance is, indeed, bliss).

Based on this new study in
Pediatrics, a couple needs to ask each other: Do we feel lucky? According to these researchers, your odds are not good! In my original blog, I made it clear that the study’s findings have not been replicated; however, these results are so scary that I can imagine parents not wanting to take the plunge, yet they may not have the time to wait for a replication. If researchers find that these results are severely flawed, then they may have mislead 1000’s of parents who choose not to increase their family size based on bad data (which brings me back to my mantra: bad data is worse than no data!). On the other hand, if this study is accurate, it is a sad day for every family with autism...

Now that the study is actually out, we are better able to evaluate the study design. One very good aspect of this study is that they followed the siblings of children with autism from eight months of age, to three years of age and then diagnosed the children. At three years of age, an autism diagnosis is quite reliable. These researchers found that of the 664 siblings, 54 were diagnosed with autism (which is an 8% recurrence). When they combine an autism diagnosis with a diagnosis of pervasive developmental disorder not otherwise specified (PDD-NOS), that’s how they come up with the much higher number of approximately 19% (and here I’m referring only to families with one child on the spectrum). Here’s the lead researcher discussing her study:

Here are the questions I have for the researchers (I couldn’t source the answers in the original article):
1) How can the researchers guarantee that the families in their study were representative of families of children with autism in general? How did they correct for volunteer bias? Remember, if they are making generalizations to the entire population of families with children with autism, the sample must be 100% representative of that population.
2) The study claims to have excluded those children with other neurologic or genetic conditions such as fragile X syndrome. Did they do the genetic testing required on each and every one of them to make sure that fragile X was not skewing their results upward?
3) Was the study sample of autistic children representative of the spectrum, or did these children have autism, as opposed to Asperger’s syndrome? If children with Asperger’s syndrome, and PDD-NOS were excluded or under-represented, this would also skew their results.

I can’t wait for the debate about these results to begin in the next issue of

Until then, we are all going to have to hold our breath!