Ontology of Psychiatric Conditions: Taxometrics
astralcodexten.substack.com
astralcodexten.substack.com
The required concepts are as follows 1. condition or underlying biological etiology 2. diagnostic codes 3. symptoms 4. the patient 5. the person assigning the code to a patient 6. the impact of the symptoms on patient quality of life (this one is nearly always left out, I left it out before I noted as I was wrapping up the post)
There are a number of relationships between these that I won't go into in detail, but to give one example, there is "some underlying condition may give rise to some symptoms."
The problem the author complains about is an inevitable artifact cause by discretization of underlying continuous variables into discrete categories. This is a direct consequence of signal detection theory and is why you see things described as spectrum disorders, because diagnostic codes are applied to sets of symptoms that are produced by multiple different underlying etiologies, including underlying conditions which would fall into the category of "normal" in the population. This is exacerbated by the fact that there is significant variability in how the person assigning the codes assesses the symptoms. An example example of this is that medical professionals consistently underestimate the amount of pain experienced by black Americans [1], and this is pain which is about is unambiguous as you can get.
The question then becomes, can we measure the various biases in these systems to try to recover what is really going on. Usually the answer is no. In the case of Alzheimer's disease post mortem pathology is absolutely required in order for a diagnosis to be confirmed, and we don't do enough autopsies to improve our models for mapping behavioral diagnosis to the pathology (which is definitional). This situation may improve over time, but as long as we cannot directly measure underlying etiology, those who may or may not suffer from an underlying condition will continue to be subjected to the current, clearly inadequate, system.
You also write "The problem the author complains about is an inevitable artifact cause by discretization of underlying continuous variables into discrete categories" - but this is mistaken. There are underlying variables that in any practical interpretation are discrete - you can have flu or not - there is no spectrum. And the problem the author complains about is that we don't know which underlying variables are like having flu and which are like height.
The conceptual data model for vaccination in the US is a table with ~370 million rows and a column for each vaccine with the type being the date the individual last received that vaccine or a null. It is nearly impossible to construct this table with any reasonable level of data quality.
Even measurements of height are subject to error. The question is whether the error is larger than the level required for some particular use case. Thus it is not that height is somehow more discrete (height varies over the course of the day due to gravity), it is that our uses for it aren't affected by the very low uncertainty that its measurements entail. However, if you are building a spacesuit for someone and use a measurement of height taken at the end of the day in a gravitation field, they will be extremely uncomfortable when they get taller in space.
Thus even when there are measurements which are nominally discrete, if the uncertainty is greater than that of the use case, then the fact that it is nominally discrete doesn't matter.
Also - what is the author conflating - can you explain what you mean by that?
But also agree (I think?) that it is pretty good otherwise. I didn't know that taxometrics existed as a field :-o
The problem with this model is that it doesn't permit itself to look backwards in time too long.
If we were to also consider the organisms that have existed down to the common ancestor of rabbits and humans, it would prove quite a bit more difficult.
It reminds me of a pet peeve of mine, the term “on the spectrum” for autistic people. The key part of a spectrum is that everybody is on it, that’s what makes it a spectrum! By implicitly allowing some people to be “off the spectrum”, you’re having your cake and eating it.
Wavelengths are all on a spectrum. Any particular wavelength may or may not be on the visible light spectrum, for example.
"The autism spectrum" does not and has never meant a whole spectrum of all people, some of whom are just "autistic enough" to be considered "on the spectrum."
The vast majority of people simply aren't on the autism spectrum at all.
Well, that's the bit I'm questioning, although I'm definitely no expert. This Astral Codex Ten post would seem to back me up, although on re-reading it's more ambiguous:
Autism is such an annoying thing to have on this list – just when everyone finally agreed there was an “autism spectrum” and no clear difference between neurotypical-people-with-some-autistic-traits and high-functioning-autistics, along comes this meta-analysis to say nope, it’s a taxon, neurotypicals and autistics are two clear objectively separate categories. I did hunt down the studies here and they actually seemed pretty good. But none of these results are strong enough to be more than suggestive, so I am going to keep believing in an autism spectrum until more research comes in.
I don't know enough to rate the research here. But it does seem to me a bit strange to work so hard on communicating to laypeople that there isn't just a single condition called "autism", it's a whole spectrum with degrees of severity, but at the same time to still insist that only a small discrete set of people "have autism" and the rest of us are completely free and easy. As other commenters have noted, all the symptoms of low-level autism (social awkwardness, sensitivity to overstimulus, difficulty in communication, etc) are things that almost everybody experiences now and then. So it seems much more helpful to me to think of autistic symptoms as things that everybody has to a greater or lesser degree; so the autistic spectrum has room for everybody.
For example - here are some possible causes of a person exhibiting social awkwardness:
1. Slow/Deep processing of language
2. Distraction by other stimuli such as bright lights, visual noise.
3. Introversion
4. Lack of experience in the particular social setting.
5. Religious structures agains certain prevailing social behaviors.
6. A headache caused by forgetting to hydrate adequately during exercise.
7. A pressing need to empty one’s bladder.
1 & 2 are often associated with Autism, but 4 could result from 1 & 2 making social situations unrewarding. 6 & 7 could just be rare flukes or they could happen frequently due to ADHD.
The point I’m getting at is that Autism (or ADHD) can be conditions with varying degrees of severity and that can be orthogonal to the fact that the symptoms present in a spectrum across the entire population, because each ‘symptom’ has many causes and is fundamentally stochastic.
What about the broad autism phenotype (BAP)? BAP is very common in close relatives of people diagnosed with ASD. And there is no clearcut boundary between ASD and BAP – the boundary varies from clinician to clinician. (On that topic, see [0] which argues that a >30 percentage point difference in ASD diagnosis rates between two specialist clinics is best explained by differences in the culture of the clinicians.)
If ASD is 1-2% of the population, BAP could be several times that.
Autism doesn't just mean, "Damn this person is really awkward." If you look into the screening tools, you'll realize many of the questions are actually about sensory processing, not just human interaction.
If there were a spectrum of human awkwardness, sure, everybody could be on it somewhere, but that's not what the autism spectrum is.
A person doesn't have to have abnormal "sensory processing" to have ASD, even though a lot of people have convinced themselves that's what ASD is all about. Abnormal sensory responses is DSM-5 criteria B(4). You need at least two of B(1)-(4) to be diagnosed with ASD, so a person can meet the diagnostic criteria without abnormal sensory responses, if they have at least two of criteria B(1)-B(3) (along with criteria A,C,D,E).
But even assuming that's true (which I don't think is actually known) how would you distinguish them?
There isn't yet a genetic test or an antibody test or anything like that to distinguish autistic from non-autistic people, right? So it's based on judging the symptoms; and given there's a wide spectrum of symptoms, the categorisation is always going to be somewhat arbitrary.
[0] is a PhD thesis which (in its first part) examines three specialist ASD diagnostic clinics in the US, with a >30% percentage point difference in diagnosis rates. It argues that the best explanation for that observed difference is differences in clinical culture between the clinics; and, that despite this substantial variation in diagnosis rates, the clinics took an unusually rigorous approach to diagnosis, suggesting that in a broader set of clinics (most of whom would approach diagnosis with less rigour) the variation is likely to be even higher.
Although we have diagnostic instruments, such as the ADOS-2 (which along with the ADI-R is considered the "gold standard"), they have problems. One problem is that the standards for using these instruments clinically are a lot weaker than the standards for research use. [1] is a study which attempts to measure the accuracy of ADOS(-2) coding by clinicians, by comparing it to the consensus coding of experienced researchers – it found a high variability in the accuracy of the clinical codings – some clinicians were quite close to the research standard, others were a fair way off.
Agreed, but it’s not clear why that is relevant.
Whether it is hard to diagnose or not doesn’t tell us whether the condition is discrete or distortion on a spectrum.
I think the answer to "whether the condition is discrete or distortion on a spectrum" is likely to be "both". I think the most severe cases – such as non-verbal classic Kannerian autism with intellectual impairment – are likely discrete conditions (albeit not a single discrete condition, more likely several). By contrast, the lower severity cases are likely to just be the extreme end of the normal variation of traits in the general population.
I think, one of the many mistakes made by the authors of the DSM-5, was creating a single diagnosis encompassing everything from the most severe to the mildest cases, when it seems likely that rather different things are going on at those two extremes. There's a bunch of research that supports that view, see for example: https://link.springer.com/article/10.1007%2Fs10803-017-3451-...
That's a pet peeve of mine too, but for a different reason – there is more than one spectrum in the DSM-5. There is a whole big section entitled "Schizophrenia Spectrum Disorders and Other Psychotic Disorders". The concept of "obsessive compulsive spectrum" was also considered for inclusion in DSM-5, but ultimately didn't make it, but if you look at the work group list inside the front matter, one of the work groups was called "Anxiety, Obsessive-Compulsive Spectrum, Posttraumatic, and Dissociative Disorders". And there are other spectrums discussed in the literature of psychology and psychiatry, such as the "bipolar spectrum", although that isn't mentioned in the DSM-5. People who say "the spectrum" instead of "the autism spectrum" appear to be ignorant of those other spectrums.
> So we can just graph GAD-7 scores and then we…
>…in fact, this doesn’t work. Let’s look at the flu graph above again.
Seems unfinished, in need of reviewing or an editor, and makes steps towards an explanation that remains incomplete.
> So we can just graph GAD-7 scores and then we, uh...
>...in fact, this approach doesn’t work. Let’s look at the flu graph above again.