There is no question that the incidence and prevalence of autism are on the rise. Starting in the early 1990s and continuing to today, there has been a steady rise in the number of children diagnosed with autism. Prior to 1990 the estimates of autism prevalence were about 3 per 10,000. The most recent estimates from the CDC and elsewhere now have the number at about 100 per 10,000, or 1%.

The burning question is – why are the rates increasing steadily? There are those, particularly in the anti-vaccine community, who conclude that the increase in prevalence is a real biological effect – an epidemic – and is evidence for an environmental cause (which they believe is vaccines, even though the scientific evidence does not support this position). However, the evidence strongly suggests that the rising prevalence of autism is largely an artifact of broadening the diagnosis and increased surveillance.

It should be noted that the data cannot rule out a small true increase in autism prevalence. Some hypothesize that increasing maternal and paternal age are contributing to the incidence of autism, but I will leave that question for another post.

A new study now adds significant support to the surveillance hypothesis – Ka‐Yuet Liu, Marissa King, and Peter S. Bearman from Columbia University, publishing in the American Journal of Sociology, report that the risk of being diagnosed with an autism spectrum disorder (ASD) correlates with social proximity to another family with a child with an ASD diagnosis. For those interested in this topic, the full paper is worth a read. While it gets technical at times, the authors do an excellent job of reviewing this topic in detail.

To summarize their key points – they begin by reviewing the history of the autism diagnosis. They point out that historically, in the 1950s and 60s, the diagnosis of autism was stigmatized by psychogenic theories of causation (the infamous “refrigerator mothers”). But then:

With hindsight, we can recognize that autism was increasingly destigmatized through the mobilization efforts of Bernard Rimland and the National Society for Autistic Children (NSAC), whose work refuted psychogenic theories of autism and set the stage for the research program that would identify autism as a neurological disorder (Dolnick 1998).

Therefore, prior to the 1970s parents actually mobilized their resources to avoid an autism diagnosis, and instead to seek a diagnosis of mental retardation (MR). This was not only to avoid stigma – the diagnosis of MR was attached to more public resources, while autism was left out in the cold. Then, this situation began to flip. Increasingly after 1990 ASD lost its stigma and became increasingly attached to access to public resources. Parents then began marshaling their resources to obtain an ASD diagnosis, rather than an MR diagnosis.

Liu et al discuss that this interpretation of the history of ASD makes a number of predictions, and they sought to test those predictions. They are not the first to do so, and others have demonstrated (which they review) that as ASD diagnoses increased there was diagnostic substitution – other related diagnoses, such as MR, decreased. Autism prevalence has increased uniformly in all age groups. If an environmental cause was at work, younger age groups should have disproportionately increased. When the same diagnostic criteria are applied autism incidence is largely stable over time. Increasing caseloads of ASD correlate with times when the diagnostic criteria are expanded.

The study authors now go further – looking at the pattern of ASD diagnoses sociologically. They hypothesized that if the increasing rate of ASD is due to sociological, rather than biological, factors then proximity should play a role. That is exactly what they found – a child who lives 250 meters from another child who has been diagnosed with autism is 42% more likely to be diagnosed, and if they live between 250 and 500 meters of another child, they have a 22% increased chance of being diagnosed. Basically, being close to a family with an ASD child provides access to information that allows other parents to more efficiently mobilize their resources to seek an ASD diagnosis.

This matters because obtaining an ASD diagnosis is not always easy or automatic. Many ASD children, on first presentation to a pediatrician, are not diagnosed. Some doctors are more likely to make the diagnosis than others. And school systems may delay diagnosis as well. Parents who therefore pursue a diagnosis more aggressively are more likely to get it.

Of course, an infectious agent would also spread through proximity, so they tested this possibility by looking at specific features that should differ. For example, the sociological spread hypothesis predicts that the proximity effect should be greatest at the milder end of the spectrum. More clear cut cases are likely to be diagnosed regardless of parental resources or effort, while milder cases will be highly dependent on these factors. They found that the proximity effect existed only among the milder end of the spectrum, and was absent for more severe children. They also found that the effect obeyed school districts – so proximity across a school district demarcation did not confer increased chance of having an ASD diagnosis. Further, they found that children with an ASD diagnosis had similar referral sources to other ASD children with proximity.

All of this points to the fact that parents obtain information from other parents in their neighborhood about which doctors to see, what questions to ask, and how to interface with the school system most effectively to obtain an ASD diagnosis and the increased services that come with it.


This study adds powerful evidence to the conclusion that the increasing incidence of autism is largely due to sociological factors rather than a true increase in the incidence of autism. This does not rule out a true increase also hiding in the data, but rather demonstrates that sociological factors are a significant contributor. This study must also be put into the context of the many other studies supporting the conclusion that the dramatic increase in ASD is due largely to increased surveillance and expanded diagnosis.

I was also not previously aware of the extent of the effect that social stigma has played. Previously parents avoided an autism diagnosis and sought an MR diagnosis, while today the situation is reversed, and the evidence shows the diagnostic substitution we would expect to be the result.

Another prediction of the sociological hypothesis is that eventually autism incidence should level off – once diagnoses are saturated. We are probably getting close to that point now.

All of this also means that scientists are justified in focusing their research efforts on characterizing the genetic risks and causes of ASD. Further, calls for a shift in emphasis to environmental causes based upon the premise that there is an autism epidemic are not valid in light of this research. This does not mean that environmental contributors should not be explored – it is often practical to cover all the bases in medical research – but a significant shift in resources is likely not justified.

Posted by Steven Novella

Founder and currently Executive Editor of Science-Based Medicine Steven Novella, MD is an academic clinical neurologist at the Yale University School of Medicine. He is also the president and co-founder of the New England Skeptical Society, the host and producer of the popular weekly science podcast, The Skeptics’ Guide to the Universe, and the author of the NeuroLogicaBlog, a daily blog that covers news and issues in neuroscience, but also general science, scientific skepticism, philosophy of science, critical thinking, and the intersection of science with the media and society. Dr. Novella also contributes every Sunday to The Rogues Gallery, the official blog of the SGU.