Bob Wright co-founded Autism Speaks and was their representative to the congressional hearing held yesterday. As part of that hearing, he called autism an “epidemic”, stating:
More than seven years have passed since my wife, Suzanne, and I founded Autism Speaks. During that time, we have seen the prevalence of autism in America nearly double – from 1 in 166 children in 2005 to 1 in 88 today, including 1 of every 54 boys. The prevalence of autism has increased by 1,000 percent over the last 40 years.
Mr. Wright’s testimony also included the statement: “The annual cost of autism in the United States is now estimated at $137 billion – a
figure that exceeds the gross domestic product of 139 countries.”
I tried to work out how Mr. Wright arrived at that figure and couldn’t easily get there. Why, you might ask? It’s nice to cross check: do figures someone assumes match the statements they are making. Luckily a recent interview makes it clearer how he came to these figures:
According to Wright, autism creates costs of $137 billion a year for the U.S., breaking down to $60,000 per year for family. Most of these costs are borne by the parents of the autistic child. Wright told Mitchell he’s “optimistic” after Thursday’s testimony but Congress needs “to have a plan” and “a will to execute it.”
Aside, more burden language. In my family, most of the difficulties with autism are lived by my kid, but I’ll move on. Partly because most of the “costs” in these estimates are incurred by adults.
A bit of quick math to see what autism prevalence Mr. Wright is assuming for autism in his calculation.
$137,000,000,000 cost total divided by 60,000 cost per family is 2,283,333 families. The US Population is currently estimated at 311,591,917. Combining these: 2,283,333 is 0.73% of 311,519,917. That would be an average of 0.73%, across all age groups. 1 in 136. Not so far from the 1 in 88 of the current autism prevalence estimate in the US for eight year olds.
But, wait, he didn’t say it that way. He didn’t say that it was $60,0000 per individual. He said “families”. Let’s take “households” as the estimate of how many families there are in the US: 114,235,996. 2,283,333 is 2% of 114,235,996. Average, across age groups. 2% of families/households have an autistic in them? It’s really the same figure as above (about 0.73%), but in another format. (in the original version of this article, I didn’t catch that fact).
So, Mr. Wright (or whoever in his staff produced these figures) is assuming an autism prevalence somewhere around 1 in 136 (0.73%). Which is pretty close to the current estimate of 1 in 88. Except that Mr. Wright’s figures appear to assume a flat prevalence over age. I.e. no epidemic. So, on the one hand we are told that autism rates are rising to make one scary point, on the other we are told autism costs a lot of money to make another scary point. But both statements are based on polar opposite assumptions. These were political and public relations statements, so it is almost expected that they won’t be self-consistent.
OK, let’s leave behind the “gotcha” phase of the article and re-analyze the statement more closely. He uses the figure of $3.2M as the lifetime cost of autism. That figure comes from this study: The lifetime distribution of the incremental societal costs of autism.
Taking just the results and conclusions of the abstract from that study:
The lifetime per capita incremental societal cost of autism is $3.2 million. Lost productivity and adult care are the largest components of costs. The distribution of costs over the life span varies by cost category.
Although autism is typically thought of as a disorder of childhood, its costs can be felt well into adulthood. The substantial costs resulting from adult care and lost productivity of both individuals with autism and their parents have important implications for those aging members of the baby boom generation approaching retirement, including large financial burdens affecting not only those families but also potentially society in general. These results may imply that physicians and other care professionals should consider recommending that parents of children with autism seek financial counseling to help plan for the transition into adulthood.
If one is going to discuss autism as a “societal cost” issue, one has to focus on where those costs are. The “low hanging fruit” of reducing societal costs are in “Lost productivity and adult care”. Productivity costs were calculated including:
Productivity losses for people with autism were estimated by combining standard average work-life expectancies for all men and women taken from the economics literature (ages 23-57 years for men and 23-53 years for women), 34 with average income and benefits (from Tables 696 and 628 of the Statistical Abstract of the United States36) and estimates of age- and sex-specific labor force participation rates
And a similar estimate assuming some amount of un and underemployment for the parents. But, even with the parental lost income assumed, the largest “costs” to society are for adults. Not really surprising as people spend much more of their lives as adults than as children. This begs many questions. I’ll start with: how much of this “cost” to society, right now, is being incurred because our adult autistics are un- and under-employed? Or to put it in a way to entice a member of Congress, how much money could the U.S. be saving, right now, if we did a better job supporting some fraction of the autism population into employment?
About 2/3 of the U.S. population is in the employable age range (18-65). That’s about 208 million Americans. Assume an autism prevalence of 1%. That’s 2.08 million Americans. Assume 1/10 of those are employable but unemployed. That would give about 208,000 Americans. Let’s take $30,000 per year as salary+benefits for these workers. That’s $6,240,000,000 ($6B) that could be realized if we could get this assumed fraction of autistics from unemployed to employed. Not including whatever is being paid out in unemployment or social security to this unemployed population.
One can quibble with the assumptions here, but we are talking big numbers here. The sort that should catch a legislator’s eye. For those who want to quibble with the idea that the autism prevalence is flat: hey, Bob Wright did it. More to the point, it’s probably correct to assume a relatively flat prevalence. And if you have real data to the contrary, you have data that is either unpublished (and I’d love to see it), incorrect or misinterpreted.
One reason to make this sort of calculation, i.e. focusing on autistics who can be employed, is that it is easy and direct. The math is simple. What about autistics who are not so close to employment, or not close at all? Autistics adults who are similar to my kid. How much do we save by investing in them? I would say a great deal. Each step helping a person move from a more restrictive adult support system to a more independent system will save money. Potentially lots of money. If that’s what congress needs to hear to be interested, fine. These are ways to make the financial impacts that should be attractive in Congress.
Rather than focus on the “costs”, I’d rather focus on what can make a person’s life better. Is the reward of a job merely the salary? I don’t think so. For those who experience even greater challenges, giving a person the ability to self-advocate to the point of not being an even bigger target is invaluable. It could stop problems like those discussed here recently.
By Matt Carey
Note, I made edits to this piece within the first 30 minutes of it being published.