Evaluating Change in Autism Prevalence: Change We Can Believe In?
This post is by Alycia Halladay and Michael Rosanoff, members of the Autism Speaks Science Team.
Twenty years ago, autism affected somewhere around 1 in 2000 children. Today, it affects 1 in 110 children. In other words, autism prevalence has increased dramatically over the past 20 years. In fact, it has increased an estimated 600%. Seems straightforward enough, yet why do we keep hearing the same question…
“But is it a true change in autism prevalence?” That is, how much of the change in autism prevalence is truly due to an increase in autism risk?
In an effort to uncover the answers to this question, Autism Speaks and the CDC co-sponsored a workshop entitled “U.S. Data to Evaluate Changes in the Prevalence of Autism Spectrum Disorders,” held on February 1, 2011 at the CDC Headquarters in Atlanta, GA.
The “Evaluating Change” workshop was designed to bring together stakeholders from the autism community and thought leaders in the fields of autism epidemiology and public health to consider the factors driving the change in autism prevalence over the past two decades. A growing body of epidemiologic research has begun to piece together the prevalence puzzle. Factors such as younger age of diagnosis, broadening of diagnostic criteria, improvements in the availability of services, and better awareness of the disorder have all been attributed to the change in autism prevalence. However, recent epidemiological studies indicated that, while these factors do account for a portion of the change, they cannot account for all of the increase alone (see Figure which was adapted from a presentation by Dr. Peter Bearman at the meeting).
So, where is the rest of the increase coming from? What is causing it? And most importantly, what can be done to identify and quantify the influence of multiple genetic and environmental risk factors, and their interaction, on the change in autism prevalence over time? (For more background, please see “What is Causing the Increase in Autism Prevalence”)
Scientists, advocates, family members, and individuals with autism alike served as expert panelists charged to identify promising directions, scientific priorities, and possible approaches for better understanding ASD prevalence trends in the U.S. The beginning of the workshop focused on existing datasets that have been used to estimate autism prevalence and identify potential environmental risk factors associated with autism diagnosis. These include: the CDC’s Autism Developmental Disability Monitoring Network (ADDM), the California Department of Developmental Services (DDS) system, and Special Education data, as well as data from international health registries such as those from the Scandinavian countries. These surveillance systems, while not designed to study the causes of autism, nonetheless have been helpful in identifying possible risk factors. Not only have these resources provided information about the rising prevalence of autism, but they have also identified and replicated findings relating to parental age, obstetrical complications, certain chemical exposures and even genetic influences.
In addition to presentations on what has been already been done to understand reasons for ASD prevalence changes, examples of approaches to understanding prevalence changes from other health conditions were presented. Scientists from the fields of cancer, Parkinson’s, asthma, and schizophrenia research presented methodologies and models for deciphering prevalence trends in their respective disorders and how these previous efforts may inform autism prevalence research. In all cases it was found that changes in prevalence over time are often the result of a complex relationship between genetic and environmental risk factors, as well as factors related to diagnosis and awareness.
In the afternoon, participants broke out into discussion groups focused on the following topics: 1) How can prevalence information be used by stakeholders including parents, professionals, and policy makers? 2) What more can be done with ADDM data? 3) Where are there other sources of data that can be used for prevalence estimates and 4) What else can be done to understand ASD trends?
As the discussions progressed, a number of overarching themes started to become clear. Risk factors studied in isolation are going to underestimate their true influence. They need to be studied in combination. The influence of these risk factors is unlikely to be simply additive, and sometimes they will overlap. Therefore, the field needs new analytic models that can study a combination of multiple genetic and environmental factors. As Dr. Ezra Susser, from Columbia University said “It is helpful now not to look through the lenses of one side or the other, but rather both, and to use the information for better advocacy and awareness.“
Another participant, Dr. William McMahon from the University of Utah put it this way, “think of this black box of prevalence as a pinball machine. You can’t predict the trajectory of the ball based on just one lever. There are dozens of other factors that influence where that ball is going to go”.
Other recurring themes included the need for data on well-defined autism phenotypes, as well as the importance of data quality and accuracy.
Throughout the meeting public attendees contributed their ideas through thoughtful, careful comments. Stakeholders expressed their ideas and their hopes for the way prevalence data would be used – these included lingering and existing concerns about different environmental exposures, need for facilities for adults, societal treatment and acceptance, and reasons for differences in developmental patterns.
As understanding autism prevalence can help us identify priorities in research, services, outreach, and advocacy, the outcomes of this meeting will be published in a meeting report to inform future strategic planning. This workshop was only a first step in better understanding the changes in autism prevalence over time and characterizing the potentially responsible risk factors. Ultimately, as risk factors related to increased autism prevalence are uncovered and if those risk factors are modifiable, then researchers will have new targets for better treatments and prevention measures. The bottom line is that the dramatic increase in the prevalence of autism calls for urgent action to understand why the increase in occurring. It also underscores the crucial need for more and improved services to address the needs of people with autism spectrum disorders and their families.