There is no question that the rates of autism diagnosis have been increasing in the US and the Western world. The latest Centers for Disease Control (CDC) estimates indicate that 1 in 88 children have been diagnosed as being on the autism spectrum (1/54 boys and 1/252 girls). This is up from just two years ago when the estimate was 1/100 children, and much higher that the 1990s when estimates were around 1/250.  

It certainly looks like we are in the middle of an autism epidemic, but experts are not so sure, and for good reason.  

As a general rule in any scientific discipline when a measured quantity is changing over time you have to consider the possibility that the method of measurement is changing, rather than the thing being measured itself. In medicine this often relates to the method of diagnosis. To be precise we often say the rate of diagnosis is increasing (or decreasing), rather than the rate of the disease - unless all diagnostic artifacts have been reasonably ruled out.  

Autism is particularly susceptible to diagnostic artifacts. Various factors have been identified that could, and probably are, affecting the rate at which the autism diagnosis is made.  

The first is an expansion in the diagnosis itself. In fact autism is part of what is now called autism spectrum disorder, because we recognize that it exists along a fairly broad spectrum. If you cast a wider net, you're going to catch more fish - that doesn't mean there are more fish.  

Related to this issue is the fact that autism is a clinical diagnosis - it's based on recognizing a set of signs and symptoms. At present there are no biological markers that can be used to reliably diagnose autism. There is therefore no laboratory confirmation or objective result that can be used to compare prevalence over time. Clinical judgment is required, and that judgment can drift over time.  

Awareness and familiarity with autism can therefore also lead to more diagnoses. Teachers may be more likely to consider the diagnosis and raise the issue with parents, and doctors may be more likely to make the diagnosis.  

Yet another related issue is so-called diagnostic substitution. There are several related diagnoses the may be more likely to be diagnosed as autism today. Two or three decades ago, however, the same children might have been diagnosed with a non-specific developmental disorder, or a language disorder.  

Surveillance is another issue - the effort being made by school systems and others to identify children who have special needs because they fit on the autism spectrum.

The scientific evidence strongly supports a major role for all these various factors in explaining the increase in autism diagnosis. Studies have shown that there is diagnostic substitution (as autism diagnosis goes up, other related diagnoses go down), there is increased surveillance, and there is wider recognition of autism. The chance of having a child diagnosed with autism strongly corresponds to living near other families with a child on the spectrum, and being in the same school system. Therefore social contact with others with the diagnosis increases the chance of being diagnosed.  

Several studies have shown that when the same methods are used to compare different cohorts of children born at different times, the autism prevalence is the same. Further, the prevalence of autism in different age groups (when the same surveillance and diagnostic methods are used) appears to be the same. If the true incidence of autism were increasing then younger age groups would have a higher prevalence.  

It is possible, therefore, from these various factors that the apparent rise in autism is entirely an artifact of diagnosis and surveillance. It is difficult to prove this, however, and so there is always a certain amount of uncertainty. It is also possible that there is a real increase in the incidence of autism over time, but there is no data that establishes that there is a real increase beyond the factors described above.  

Of course there are some groups that are invested in the notion that autism is truly increasing and represents an epidemic. Most notable is the anti-vaccine movement, who over the last decade have been blaming the increase in autism on vaccines. Initially they blamed autism on the MMR vaccine, but the scientific data did not support that claim, and the credibility of the originator of this fear, Andrew Wakefield, has since crashed and burned.  

After the MMR hypothesis failed they next turned to mercury in the form of thimerosal in some vaccines. They confused correlation with causation by arguing that autism rates were increasing as the vaccine schedule also increased. By the end of 2002, however, thimerosal was removed from the routine vaccine schedule in the US, and therefore the amount of exposure to thimerosal plummeted. Advocates of the thimerosal hypothesis (such as David Kirby) predicted that autism rates would also plummet.  

They were correct in that, if thimerosal were a significant contributor to autism then the rates should drop considerably once thimerosal was largely removed from childhood vaccines (it was still present in some flu vaccines). Since 2002, however, autism diagnosis rates have continued to increase at the same rate. We are now 10 years later, and there is simply no justification for clinging to the thimerosal hypothesis any further (of course this hasn't stopped antivaccinationists).  

Ironically the antivaxers are now using the continued increase in autism diagnosis to argue that vaccines cause autism, even though that increase contradicts their prior predictions.  

The data clearly shows that much, if not all, of the increase in autism is due to expanded diagnosis and surveillance. It is still possible that there is a real increase, but more research would be needed to establish that.  

However, regardless of cause, we now know that autism is very common. It therefore deserves a proper level of attention and resources, both for research and services for those on the spectrum.


Steven Novella, M.D. is the JREF's Senior Fellow and Director of the JREF’s Science-Based Medicine project.

Dr. Novella is an academic clinical neurologist at Yale University School of Medicine. He is the president and co-founder of the New England Skeptical Society and the host and producer of the popular weekly science show, The Skeptics’ Guide to the Universe. He also authors the NeuroLogica Blog.