the trouble with “overdiagnosis” rhetoric

There’s a common sentiment — one I shared before I knew anything about autism — that kids today get too many labels. How can such a huge percentage of kids have autism, ADHD, and so forth? Here’s today’s example article at Salon.

I’ve come to believe that there’s a real concern behind this sentiment, but it accepts too much of the status quo. The problem is not that we’re putting people on the wrong side of the line; the problem is that our bureaucracies insist on drawing a bright line where there isn’t one… and that people are so afraid of autism when they need not be.

If we look at real humans in all their dimensional richness, then we can estimate that they are at 10th or 90th or 99th percentile on “autistic traits,” as observed on some measure such as the ADOS. These tests (not 100% successfully) try to to pick apart autism as its own separate trait, distinct from IQ, personality, or other factors.

I haven’t seen convincing research that supports a bimodal distribution. There isn’t some kind of bright line where you have autism on one side and not-autism on the other side. (If you have a good counterpoint reference here, send it along!) There seems to be evidence to the contrary, in fact (search term: “broad autism phenotype”).

So imagine the current autism line is at the 99th percentile of autistic traits (consistent with around 1 in 100 prevalence). If we measure someone at 98th percentile instead of 99th, or 99th instead of 98th, is that some sort of horrible travesty? Many experts say that these “just outside the line” people would benefit from the same supports that “just inside the line” people will.

If we get worked into a lather about overdiagnosis, our objection is something like “this person has many of the same traits and needs the same help as a person with autism, but isn’t quite autistic!” What does that even mean? Who cares?

If a kid doesn’t interact with peers, they’ll probably benefit from some prodding and help to do so. If an adult can’t manage friendships and it makes them unhappy, they’ll probably benefit from help addressing that need.

When we brought concerns about our son to a doctor, we had a speech evaluation after just a week or two of waiting, and it established that he had a language delay. But insurance didn’t cover “needs help with language”; it only covered “autism.” We had to wait 4 months to get the full autism workup, at which point we could get speech therapy covered. How does this make sense?

If someone “needs help with language” they should get help with language. If they “need help with peer interaction” they should get help with peer interaction. If they “need help with reading” they should get help with reading. What does a binary, yes-or-no diagnostic label have to do with it?

Sure, it’s useful to do a full assessment. Look at IQ; look at personality traits such as anxiety; look at autistic traits such as ADOS score. All these things are valuable in understanding a person. But if we’re trying to draw some kind of bright line between 98th percentile ADOS gets nothing and 99th percentile ADOS gets 20 hours per week of intensive intervention, we’re out of our minds.

I’ve noticed that none of our son’s therapists or educators really use the autism diagnosis at all because it simply does not have significance to their practice. In practice, we want to know “can the child answer wh-questions? if not then let’s work on it” or “does the child make requests to peers? if not then let’s work on it.” In fact, using the diagnosis usually leads them to misjudge what our son needs; they might decide he needs visual rather than written supports even though he’s been reading since age 2, for example, because kids with autism are “visual thinkers.”

People also object to overdiagnosis because they feel that labels are harmful. “Why lower the child’s self-esteem and stress out the family?”

But wait. Is the harmful side of labeling somehow OK if a person has “enough” autism? Why should anyone have to suffer here? How about working to remove the harm of these labels, or how about simply sticking to dimensional measures? The problem is medicalization, pathologization, and stigma; the problem is not our recognition of someone’s individual needs.

Dr. Gnaulati’s argument in Salon boils down to: we mistreat autistics so badly that we should avoid saying that anyone is autistic, unless we absolutely have to. He says if we label someone autistic we’ll “devalue their mental gifts” and treat their strengths as “symptom[s] to be stamped out.” Maybe the problem is mistreating autistics, rather than exactly which people get this mistreatment. I certainly do not treat my son in the way he describes, and neither do any educators or therapists if we’re going to keep working with them.

What I’d like to see, emphasized much more than a diagnostic label, would be something like: “according to our tests, this person has XX percentile IQ, XX percentile creativity, XX percentile autism, XX percentile language, XX percentile whatever else… all of these have a huge margin of error… and overall we think doing A, B, and C would be beneficial for this person, but let’s adapt our approach over time as we learn more. And let’s focus on possibility not probability.” That’s the honest approach.

Labels come about because our institutions, medical and educational, are incapable of dealing with individuality unless they’re forced to. Insurance companies and schools need a label. I say: if that’s what bureaucracy requires, then that’s what bureaucracy gets. Let’s label everyone who needs any kind of individualization — pretty much EVERYONE.  Not my problem and not going to affect my self-image or my image of my son. This is a bureaucratic hoop.

Adopting labels to get help isn’t “overdiagnosis”; it’s an understandable and rational response to inflexible institutions and their refusal to deal in subtlety or individuality. The word “overdiagnosis” implies that there’s something more at stake than nonsensical hoop-jumping. It blames the wrong people for the wrong problem.

Labels are equally problematic when they are “positive,” such as “gifted,” or “high-functioning.” Is the just-barely-gifted child so different from the almost-gifted child? Get real.

Reducing a dimensional, varying-over-time trait to a yes-or-no question throws out a hell of a lot of information.

Maybe a problem is that we tell parents their children are “broken” or “not broken” or “gifted” or some other bucket, instead of giving them useful insights. Maybe the problem is that not all autistics are the same, and individualized education and support therefore has to be exactly that. Add up ADHD and dyslexia and gifted and everything else, and the reality is that people are diverse. Our institutions should be built assuming diversity — but they are not. The solution to this is not to make our diagnostic buckets more accurate; the solution is to think about people instead of buckets.

One more point about autism specifically. Researchers, educators, and clinicians seem to disagree on whether the autism cutoff has to do with “degree of autistic traits” or “level of impairment in practice.” Say you have someone with a high IQ and very resilient personality, with the same degree of “autistic neurology” as someone with lower IQ and a fragile, anxious personality. The outcomes may be very different. Is an autism diagnosis determined by the practical outcome, or by autistic traits?

If your goal is to research the relationship of autism to specific neurological differences, you may want to look at autistic traits, not outcome. If your goal is to decide whether someone qualifies for disability benefits, you want to look at outcome. Much casual discussion, research, and clinical practice is completely unclear on whether autism is a trait or an outcome. Even the DSM is unclear on this.

Overall life outcome will be a function of many factors; many traits, many life experiences. Plenty of people who display autistic traits as toddlers will have a good outcome (not that a good outcome has to involve dropping autistic traits). We don’t have enough research to reliably predict who will have the best outcomes. If we say the ones with optimal outcomes were “misdiagnosed,” we’re implying things like:

  • it was a mistake to offer these kids help
  • the clinician could have or should have predicted the outcome
  • the kids have no trace of “autistic neurology” when older

As far as I know, we can’t say those things yet.

In young children, we don’t know their eventual level of impairment, only their age-normed degree of autistic traits. That’s why childhood diagnosis is based on traits more so than impairment. (Incidentally: scores on the same ADOS module drop as children get older, because kids often learn the behaviors ADOS looks at over time … autism is identified by the delay in some of the behaviors.)

Making this more concrete, let’s take my son. He didn’t spontaneously take up pretend play. But since he’s very bright, as soon as someone shows him a game such as “let’s pretend to go camping” he can learn it and even vary it. Later, someone comes along and they might say “well this kid can play camping, so it isn’t autism.” Has his autism been cured? I don’t think it has. He’s an autistic kid who can play camping.

But I do think his practical difficulty faced is less than that of a child who might be unable to learn to play camping. That’s because my son has an amazing memory which he uses to adapt. The autistic trait is still there (that playing camping didn’t independently occur to him as an interesting thing to do), but his intellectual strengths allowed him to play camping with some friends the other day, and he loved it.

Or take me — I had a hard time as a child, but as an adult, I’m completely capable of handling most social situations, especially professional ones that I’ve practiced thousands of times. Am I “less autistic” than the younger version of myself that couldn’t handle those? I don’t think it makes sense to say that. It does make sense to say that I need no supports now, but I used to need them then. The “it’s misdiagnosed unless the person is greatly impaired” camp might say that I’ve “cured” myself over time, but I’d prefer to say that I’ve learned and adapted.

The DSM, of course, contributes to the mess. It defines autism as a list of autism-like traits (though some disagree and attempt to distinguish “autism-like” from “autism”). If autism-like can be split from autism with the DSM criteria, it’s only by relying on weasel adverbs such as “significantly,” or something like “experts know it when they see it.”

Let’s face it that autism-like is autism for now, because we don’t know the etiology. We don’t know the bright line. All we know is that if a person has certain traits, they likely (though not certainly) would benefit from a certain kind of help. In adults we can ask “are they unhappy about something and do they want help with it?”, in kids we have to guess at the future.

In fact, it’s a minor miracle that the ADOS and other clinical tools work as well as they do; predicting a lifelong pattern based on a few behavior details at age two or three is amazing, even if not perfect.

Sorry to rant at so much length, but as you can see I find this very frustrating.

Getting outraged about “misdiagnosis” misses the point; it blames clinicians and parents and autistics. But our current knowledge does not support the idea of drawing a bright line. It supports the idea of measuring individuals along many dimensions and providing them the individual supports they likely need based on what we know. The need for a bright line comes from politicians and bureaucrats, and those are the people who should catch flak.  Why are we complaining about clinicians and parents for “drawing the line in the wrong place” when many people aren’t interested in this stupid line-drawing exercise in the first place?

My son’s therapists, and his parents, would be more than happy to help him with peer interaction and language purely on the basis that he needs help with those things. “Defining autism” is somebody else’s made-up problem, not mine.

The diagnostic line shouldn’t even matter that much, unless we’re mistreating people based on it. Let’s stop that.

P.S. According to the local school team, an independent autism-specialist diagnostic team, CARS, ADOS, SRS, and more, my son’s autism was “clear.” He’s as tested and certified as it gets. According to me before hearing all that, it was borderline. It turns out that my frame of reference was from my own childhood and my relatives’ behavior — everything I grew up with was toward the autism end of the human spectrum.  So my perception of “normal” is someone else’s perception of “autism.” I’m not the parent claiming my kid has autism while the experts disagree; very much the opposite.

P.P.S. What does the “overdiagnosis” crowd propose we do about it? You don’t just go talk to a psychologist for 15 minutes and get an “autism” sticker. Our son was evaluated by two independent teams, one at the local school (potentially biased against providing services) and one at the local autism center. All had extensive autism experience. Between them they spent many hours, used 5 different well-validated autism-specific diagnostic instruments, interviewed both parents and teachers, observed our son in preschool, did non-autism-specific assessments such as intelligence and language… on the whole it was extremely thorough and evidence-based. If they still turn out to have “misdiagnosed” him because he turns out to be a perfectly happy person — which I expect! — then what were they supposed to do? Ignore what they were looking at? Spend even more time? What is the suggestion?

Dr. Gnaulati uses the case of William as an example, and says “within minutes of playing with William, I knew, unequivocally, that he was not ‘on the spectrum.'” Is his suggestion that his 5-minute expert judgment (gut feel?) is the right way to diagnose autism? Did previous evaluators of William see something different, or did they disagree with Dr. Gnaulati on where the “autism cutoff” lies? What are parents and autistics supposed to make of this clinical infighting?

It’s easy to say in retrospect that autism interventions were an overreaction to a particular person’s autistic traits (though we can’t know what would have happened without intervention). But how do we know what to do looking forward?

8 thoughts on “the trouble with “overdiagnosis” rhetoric

  1. Pingback: an annotated, selected hyperlexia bibliography | Intellectualizing

  2. Pingback: Book: “Rethinking Autism: Variation and Complexity” | Intellectualizing

  3. Pingback: On normalization and social skills: my reaction to “The Kids Who Beat Autism” | Intellectualizing

  4. Pingback: The Parents Who Beat Up “The Kids Who Beat Autism”

  5. Pingback: Schools are disordered, and they blame our kids | Intellectualizing

  6. Pingback: Not explaining autism | Intellectualizing

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s