Deep Dive
Shownotes Transcript
This BBC podcast is supported by ads outside the UK.
Earn four times points when you shop for items like Pantene Shampoo, Gillette Fusion 5 Razors, Secret Body Spray, Always Pads, Loves Diapers, Pepto-Bismol, and Nervive Nerve Relief Cream. Then use your rewards for discounts on groceries or gas. Offer ends May 20th. Restrictions apply. Promotions may vary. Visit Safeway.com for more details.
Lowe's helps refresh your garden in time for Mother's Day. Right now, get five bags of 1 1⁄2 cubic foot Scott's NatureScapes mulch for just $10. Plus, select 1 1⁄2 gallon annuals hanging baskets make the perfect gift. Now two for only $15. The best garden starts with great deals. Lowe's. We help. You save. Valid through 5-7. Selection varies by location. While supplies last, discount taking the time of purchase.
BBC Sounds. Music, radio, podcasts. Hello, Charlotte MacDonald here. We're interrupting your regular more or less podcast feed with an episode from another podcast that we thought you might like. It's all about a number that seems to keep on going up and up. And because of that, it's been getting a lot of attention with arguments and conspiracy theories about what might be behind it. The figure in question, the number of diagnoses for autism.
Radio 4 has a new five-part series out today that digs into the data to find out what's really going on. This series is called The Autism Curve, and we're bringing you the first episode. It's presented by Michael Blasland, who, more or less superfans will know, was the founding producer of More or Less some 23 years ago. Here's Michael. Whatever you think of it, however you explain it, something curious is happening.
There has been an increase of 270% in the number of children presenting with autism. Parents tell us of the soul-destroying wait for an autism assessment for their children. There has been a 400% rise in people waiting to see an autism specialist in England.
A dramatic rise in diagnoses of autism in the UK and elsewhere. Unless the long-term exponential increase in autism is stopped by 20... Reports show there's been a massive jump in autism diagnosis in the last two decades. Did you have your numbers like that? Something's wrong. Something's wrong.
New ideas about what autism means. To be able to see things, understand things, process things and remember things in a way that most people can't do has to be seen as a gift. I'm Michael Blastland and from BBC Radio 4, this is The Autism Curve.
A huge rise in numbers, arguments and conspiracy theories about why. Changes in what autism is, who gets to define it, whose experience counts. Affecting both those who are autistic and those who aren't in just a few decades. About 25 years ago, my son, Joe, was described as a prototypical case. Very little speech, profound intellectual disability, sometimes violent.
But is that still the image of autism? In his short life, it has become a different word, changing how we think about how we think. In the course of this series, we'll try to make sense of the autism curve, what's happening, what it means, where it might lead. And we begin with that astonishing trend in the numbers. This is episode one, The Data. My name's Ginny Russell and I'm an associate professor in psychiatry at University College London.
And my book is called The Rise of Autism. So Ginny, we're making a programme which is about the rise in autism diagnosis, but to count something you have to say what it is you're counting. What is autism? Well, the definition that's given in the DSM, that's the Statistical Manual of Mental Disorders, which is the kind of official psychiatric classification system that
is that autism is diagnosed when a person demonstrates two types of behaviours characteristic across settings. So the first group of traits is that an autistic person, according to DSM, has persistent difficulties in social life and in social communication. So like a lack of turn-taking in conversation, or an autistic person might go off on a long monologue, for example, right through to being completely non-verbal.
Difficulty maintaining relationships or initiating relationships. Also, lack of eye contact or two-way interaction. So that's one group, if you like, of traits. And the second type of behaviour that must be present for the diagnosis to be made are restricted interests and repetitive behaviours.
And these can be things like motor movements, like hand flapping or rocking or skin picking, but also sticking to repeated routines, which might cause difficulties when people are trying to shift tasks or if there are unexpected events, for example. So it's things like eating the same food every day. Also, in
In that second group of traits is hyposensitivity or hypersensitivity to sensory stimuli, like when a person's oversensitive to sounds or a particular fabric is attractive or repellent. Some or all of this might already be familiar to you from your own experience perhaps, or if you've read the curious incident of the dog in the night time, or seen the naturalist Chris Packham talk about his autism on TV.
But there's one more detail. It'll be relevant later. What is also important to note is that for many decades we've thought of autism as a spectrum and the DSM has severity levels that range on each type of trait, from needing some support to needing very substantial support. So within the spectrum...
The expression of autism is highly heterogeneous. Highly heterogeneous. Keep that in mind as we try to count who's captured by the definition, who isn't, and how that's changed. So then, the data. I think you did a pretty big study looking at diagnosis in the UK.
What did you find? Yeah, we conducted a study a couple of years ago where we looked at changes in rates of autism diagnosis. And we used data from about 9 million patients in England who were registered at GP surgeries.
So quite a big sample. We looked at shifts in diagnostic practice between 1998 and 2018. So it was a 20-year study. And we found that overall it was about an 800% increase in autism diagnosis between those years. So it was an enormous amount, an enormous increase, best described as exponential. But we've actually seen a quite similar picture in all European countries. Is it global?
And if it's not global, are we seeing it only in rich countries, for example? Well, the honest truth is we lack data from most of the world. In the Anglophone and European countries where we do have data, there is compelling evidence to suggest that other countries have seen a similar sort of rise in diagnosis as in the UK. And in the United States? In the US, the data are gathered by the Centre for Disease Control. And in 2000, in the turn of the century...
One child in every 150 children, I think it is, with autism. So that was less than 1% of children. But that's now gone up, according to their data, which uses the same methods to collect it, to one in every 36 children. So that's about 3%. So again, that's around a sort of 300% to 400% increase in diagnosis in the 20-year period. Maybe 300%, maybe much more.
Depending who's counting, compared with 20 years ago, everywhere we've been looking. We didn't start with a big number, but it's still been a huge change. And what's perhaps more striking? There are more big numbers still to come.
I'm Joshua Stott and I'm a professor of ageing and clinical psychology based in the ADAPT lab at University College London. I'm also an honorary consultant clinical psychologist, which means I work clinically as well as being an academic. Joshua trained as a clinician, then moved into academia. In his clinical years, he worked with older people thinking about dementia.
where he found something puzzling. In the context of kind of trying to understand if someone has dementia, what you very occasionally came up against was that someone had experienced various forms of sort of cognitive difference throughout their life, and it became apparent that for some of these people they were autistic, and people had assumed they had dementia because what was happening to them, I think in many cases, was that they were losing their social supports as they got older...
becoming more isolated, and suddenly some of their difficulties became more apparent in interacting with environments. And then they were referred to our service as if they had dementia, and it became apparent they didn't. So we were seeing a few people, but then I realised, actually, this should be a much bigger proportion of the population, at which point we thought, hmm, this is interesting. Let's try and understand this. So you went away and you did a pretty big study. Yeah. Some GPs submit their data to a central database
computer system and
Those GPs are broadly representative of the GPs in the UK in general. So what we did was we looked at who in those records was diagnosed autistic and split it by age, socioeconomic status and gender and whether people had an intellectual disability or not. What they were looking for was how the rates varied between these different groups. So you were principally, because of this experience of looking at older people, looking for disparities. Exactly. So...
How did the proportion of diagnoses vary across different groups? So it varied a huge amount. He's not kidding. In his published research, he found diagnostic rates about 150 times higher in teenagers than in the elderly.
That can't be plausible as a description of the real prevalences. I would absolutely agree that that isn't plausible of the true prevalence in the population. I think that basically people, as they get older, it's likely the prevalence remains very similar because autism is a lifelong condition and there may be some drop-off with early mortality, but it would be nothing like the amount of drop-off that we're talking about here. Which is to say that if the numbers diagnosed in younger ages are right...
At older ages, they can't be. And a vast number of people with autism were and still are missed. How many? To know that, we have to make assumptions about how many we'd expect to find in the population in general.
Did you have a kind of baseline assumption for what proportion of the population was autistic? When we approached the data in the first place, we were going with a kind of 1% figure, which kind of aligns with a study by someone called Terry Brugger, which did an active case-finding study in England and found about 1%. I'm curious about actively seek.
What does that mean? Do you go around and knock on people's doors? Kind of, metaphorically. So what you do is you get a population sample that's kind of randomly sampled from the population and you call them up and ask them to do a survey. There's a questionnaire called the Autistic Quotient
And they use that and it has a cut off for people that you would expect to be autistic. And they also then assess some of them in more detail later. So they took another sample of those and assess them in more detail in order to kind of verify that their results were roughly right and to adjust their kind of results if they felt there was some error in it when you looked at the other tests. But that 1% is itself contested.
There's reason to think the baseline for how many autistic people we should expect to find at every age could be higher. So we did two things. We thought that that prevalence may underestimate. And a key reason for that was that only one in nine, I think, from that study of people within that age
would have been autistic women, and that would be a kind of a ratio of like nine to one men to women. And we know that that ratio is probably wrong. So we wondered whether that study actually underestimated the prevalence. So what we did was we took that as a lower bound. So if you look at our estimates in the study, what we did was we had a lower bound prevalence, and then we had an upper bound prevalence. And the upper bound prevalence was simply the highest rate that we found in our sample.
Two possibilities, then, for how much undercounting there might be in people at older ages, based on two different baselines for the population as a whole. But the bottom line is that whichever baseline you use, you still finish up with a huge undercount of the elderly. So if you had to stick an overall number on the amount of underdiagnosis in the UK...
What would that number be? We estimated from our paper that depending on which of the prevalence estimates you took, so if you took the 1% it would be a lower bound, if you took the 3% it would be a higher bound, that it was between around 440,000 and 1.2 million people undiagnosed. And that was as of 2018. Obviously the situation may well have changed since then, although I still suspect there will be a very large number, particularly in older age groups, that are undiagnosed.
Either way, a big number. Though at this point I can hear an objection. If people with autism are that easy to miss, do they really have anything worth seeing? That reflects a concern that the problem isn't under diagnosis of the elderly, it's over diagnosis everywhere else. Hold that thought, we'll come back to it. Meanwhile, Joshua Stott thinks it unlikely. Did you find any over diagnosis anywhere?
So that's where I feel slightly trickier because I think this is where my point about the underlying prevalence goes back to. If you were to take the CDC estimate, the American estimate, then no, not really. It was in line with that. If you were to take the Bruja estimate, then perhaps it would be more in children than you'd expect according to that. But I don't really think overdiagnosis, certainly from our study, is the main thing because I think it's in line with estimates of prevalence. I think by far the main thing
is the underdiagnosis that we see in later life, because we know that's happening. We've heard two things. First, a dramatic rise in diagnoses. Second, clear evidence of how a whole category of people can be missed. In this case, the elderly, whose rates of diagnosis are too inconsistent with younger ages for both to be credible. Once we begin to see how counting autism goes wrong like that, and it's hard to dispute that something here has been very wrong indeed...
we become sensitised to how counting can conceal as much as it reveals. Is that the true story of the autism curve? Not a history of rising autism, but a history of miscounting? Who else did we miss? Who else is now called autistic who wasn't before? And what did that do to the numbers? Join us next time on the Autism Curve.
The Autism Curve was presented by me, Michael Blastland, and produced by Simon Mabin. It was mixed by Neil Churchill. The production coordinator was Gemma Ashman, and the editor was Claire Fordham. And if you want to hear the rest of the series, look up The Autism Curve on BBC Sounds. We'll be back with our regular More or Less episode later in the week.
Hey there, it's Ryan Seacrest for Safeway. This spring, refresh your spring personal care items and earn four times points on all your favorites when you shop in-store or online.
Earn four times points when you shop for items like Pantene shampoo, Gillette Fusion 5 razors, Secret Body Spray, Always Pads, Love's Diapers, Pepto-Bismol, and Nervive Nerve Relief Cream. Then use your rewards for discounts on groceries or gas. Offer ends May 20th. Restrictions apply. Promotions may vary. Visit Safeway.com for more details. So you want to start a business?
You might think you need a team of people and fancy tech skills, but listen to me when I say you don't. You just need GoDaddy Arrow. I'm Walton Goggins, an actor, and I like the sound of starting my own business, Walton Goggins Goggle Glasses. But I couldn't do this on my own. GoDaddy Arrow uses AI to create everything you need to grow a business. It'll make you a unique logo. It'll create a custom website. It'll write social posts for you and even set you up with a social media calendar. How cool is that?
Well, listen to this. For a limited time, you can get Aero All Access for just a dollar a week for 12 weeks. We're talking all the AI power of GoDaddy Aero plus a domain, e-commerce store, payments, professional email, a unified inbox, all for less money than I spend on deep tanning lotion while sunbathing off the Amalfi Coast. You know what that sounds like? A plan. Get started at GoDaddy.com. Terms apply.