By Matt Carey
December 22, 2013
Some children are identified as autistic and later found to be non autistic. How much does this represent “recovery” from autism and how much does this represent something else?
The question became very big after the 2007 National Survey of Children’s Health NSCH) results were published. The survey asked parents if they had ever been told by a health care provider that their child was autistic. (Note that this is different from confirming that a child actually had a diagnosis). They then asked if the child is still autistic and about 40% said, no, I was told my kid was autistic before, but he/she is not autistic now.
This raised a lot of questions. Are these kids “recovering” from autism? Were they autistic to begin with? These and more questions just couldn’t be answered in by the data collected.
Consider the 2007 dataset. Here is a list of raw data from the 2007 NSCH. 1427 parents, or 1.56% of parents answered yes to “Has a doctor or other health care provider ever told you that [S.C.] had Autism, Asperger’s Disorder, Pervasive Developmental Disorder, or other Autism Spectrum Disorder? ” [S.C] is the code for the child being discussed (selected child or something to that effect).
(As with all figures in this article, click to enlarge).
Again, 1427 parents who said that some medical professional had stated the child was autistic in the past. Of these 459 answered no to “Does [S.C.] currently have Autism or ASD?”
What’s going on there? Again, are these kids recovering? It’s a question certainly worth looking in to. The researchers felt it needed more attention and in the 2011/12 survey, discussed below, the researchers did ask more questions about autism and this subgroup.
The 2011-12 NCSH was published this past year (March of 2013). Once again they asked autism-related questions. The prevalence estimate from this survey was about 2% (1 in 50). In all, 2.13% answered yes to, “Has a doctor or other health care provider ever told you that [S.C.] had autism, Asperger’s disorder, pervasive developmental disorder, or other autism spectrum disorder? ” [S.C] is the code for the child being discussed (selected child or something to that effect).
They then asked, “Does [S.C.] currently have autism or autism spectrum disorder?” and 0.36% said no. I.e. out of the 2.13%, 1.7% said that, yes, they report that their child is autistic. After controlling for some factors, an estimated prevalence of 1 in 50 (2%) was reported.
So, what about the 343 kids who were previously identified as autistic but who aren’t now. 343 kids are about 17% of all those ever identified as autistic. But in 2007, about 33% of parents answered “no” to “Does [S.C.] currently have autism or autism spectrum disorder?” I.e. the fraction of these potentially recovered kids went down by 1/2. In 4 years. If those kids are a measure of autism recovery, something dramatic is happening. As in, autism recovery rates are dropping fast.
Luckily we can test whether these kids do represent autism recovery. The authors of the NSCH added new questions to the 2011-12 survey. First they asked the obvious and important question “To the best of your knowledge, did [S.C.] ever have autism or autism spectrum disorder? ” 228 parents, 0.24% of the total population surveyed, said “no”. Only 97 said “yes”.
That’s a lot of numbers, so let’s recap. In the 2011/12 survey:
- 95,677 parents were surveyed
- 2,041 answered that at some time they were told their child was autistic
- 343 of those said their child is not currently autistic
- 228 of those said that their child never was autistic.
Most of the kids who “lost” their autism label were never autistic to begin with.
Of the total kids in the “ever identified autistic” group, 97/2041 or about 5% said that they believed their child was autistic at one time in the past, but was not in the present. Not 20%, not 30% as some have suggested. 5%. Still worth investigating, but not the high numbers I sometimes hear people quote.
The survey authors asked two follow up questions to the parents who reported that their child is not currently autistic, but was in the past: “Treatment helped the condition go away” and “The condition seemed to go away on its own”.
For those looking for support that some therapy or combination of therapies is recovering kids: 69 parents out of 2043 reported that their kid lost the autism label, and that treatment was the primary factor in the loss.
69/2043: we are talking about roughly 3%.
They also asked parents to comment on whether “The behaviors or symptoms changed” or “A doctor or health care provider changed the diagnosis”.
There are a few other questions on autism. For example, “With more information, the diagnosis was changed” (158 parents said yes). And “A doctor or health care provider changed the diagnosis” (46 parents said yes).
The last two autism questions are very important: “The diagnosis was given so that [S.C.] could receive needed services,” and, “You disagree with the doctor or other health provider about his or her opinion that [S.C.] had autism or autism spectrum disorder”.
Out of the 343 children who “lost” their diagnoses, 102 (30%) parents say the diagnosis was given to obtain services. 122 (36%) of parents say they disagree with the original diagnosis.
The parents report that these kids were never autistic.
To summarize–Yes, a significant fraction of the children in this survey reported as once holding an autism label are not currently autistic. About 17%, to put a number on it. And, of that 17%, many have social factors involved in their “loss” of an autism diagnosis: incorrect diagnoses, the search for services, etc.. About 5% of autistic kids are reported by their parents as once really being autistic but not presently autistic. Are these kids a subgroup of actual recovery? It’s hard to tell.
One can drill down further into the data and get more insight about this group, but that will wait for another article.
The bottom line is simple, though. The National Survey of Children’s Health does not support the idea that 20-40% of autistic children are recovering. Maybe a few percent are, and with small numbers that will make studying this subgroup very difficult.