Star Ford

Essays on lots of things since 1989.

Autism diagnostic criteria in DSM-V

on 2012 January 25

There’s been a lot of buzz about the DSM-V update to the autism diagnostic criteria. (NY Times article)

The autblogs seem to oppose it, the reasons for which were not readily apparent to me, so I thought I would pull together some thoughts and opinions about the change. The first thing I noticed about the web-buzz is the rush to judgment about what will happen based on this or that study, such as the idea that they are trying to deny people services (or alternately, that they are trying to sell people more drugs).

In my analysis, as I will elaborate on below, my main conclusions are (1) It is important to get the definition right, but the new DSM-V scattered language is not a help. As currently proposed, it could cause a more random, less helpful rate of diagnosis. (2) The DSM should be less judgmental, and not define thresholds or attempt to allocate services. The allocation of services is a separate issue entirely, and should be based on the observed needs of people in their environment. And (3), the definition of autism should be based on a more structurally sound understanding, using the categories of focus, sensitivity, acculturation, and naivete. I’ve drafted completely new language for the diagnostic critiera.

1. What will the changed definition actually do?

Overall, the proposed definition is wordy and fuzzy. It is the kind of writing that you get when people don’t really know what they are trying to say. It hedges more than it defines.

First, the new criteria are more vague and subjective. For example, the old criterion – “a lack of spontaneous seeking to share enjoyment, interests, or achievements with other people (e.g., by a lack of showing, bringing, or pointing out objects of interest)” – lists specific things that typify the behavior. The new criterion – “deficits in social-emotional reciprocity; ranging from abnormal social approach and failure of normal back and forth conversation through reduced sharing of interests, emotions, and affect and response to total lack of initiation of social interaction” – is scattered and much harder to read. With the old wording, you could make a case that a particular person does that or not. But with the new wording, there is no center to it; it is so indefinite, that it has become meaningless. Almost anyone in the world fits somewhere in that vagueness.

Second, the new criteria are more judgmental. For example, the old – “stereotyped and repetitive use of language or idiosyncratic language” – is subjective, but not value-laden. The new wording – “deficits in nonverbal communicative behaviors … poorly integrated … or deficits in understanding..” – reads like a list of things we hate about you. Why would they go out of their way to change non-judgmental language into purposefully judgmental language? Much more than the old wording, the new version forces the diagnostician to make value judgments based on how things seem to them, such as whether behaviors are “excessive” or “appropriate”.

Because of these changes (from specific to vague, and from objective to judgmental), my prediction is that the new manual will result in a more random categorization of people. Some diagnosticians will include more people, and others will include fewer, based on their personal values. More people who are not truly autistic will be misclassified as such, and more who truly are will be excluded. It is just overall less meaningful.

2. Why define autism in the first place?

The DSM appears to fulfill two main purposes. As a glossary of terms, it allows people to talk about neurological and psychological conditions using shared vocabulary. Researchers can, for example, study a data set where subjects have been labeled with standard terms, and others can use those results meaningfully in other research.

The second purpose is as a taxonomy, it allows non-specialists to place someone in a large sea of fine distinctions, much like you might use a leaf identification book to find out the species of an unfamiliar tree. Someone needing help can use this taxonomic placement to connect with specialists and teachers who can help.

I feel it is important to define autism and other conditions accurately because understanding who we are is vital to our development. Being mislabeled and then being blindly mistreated for the wrong diagnosis (or being unlabeled and being punished for being who we are) is a bad outcome. Having a good definition is a first step in avoiding that trap, and is a step towards finding others like us, finding others who understand us, and finding self awareness.

Some of the web buzz appears to assume that the DSM has further purposes, which I don’t think it has. First, it does not provide diagnostic tests, so it cannot be used by itself to establish a diagnosis. Second, it does not define the need for services or allocate services.

3. Why make a change in the definition of autism?

The DSM authors’ reasons for changing the definition appear to boil down to (1) simplifying, and (2) recalibrating. This understanding is only from reading secondhand sources.

Simplifying: They did simplify in terms of coalescing several related categories into the umbrella of autism, but they created more complexity at the same time by making it scattered and vague. Simplification makes sense, as the old categories did not appear to be based on valid science. But I don’t think they have achieved it.

Recalibrating: This is the more controversial reason. It is as if the DSM-IV accidentally made it too easy to get a diagnosis, so the DSM-V is supposed to fix that mistake by un-diagnosing people who were never really broken in the first place. This is where the mental fog thickens into mental mashed potatoes. It was never the job of the book to widen or narrow definitions in order to capture more or less of the population. That’s like saying we are going to reclassify 25% of birds into “neo-reptilions” because the budget for studying birds went down and so we need the bird category to be smaller. If that happened, then you (the bird watcher) would be obligated to declare that a particular thing perched on your bird feeder is not a bird, because it does not fit the newly restricted criteria for being a bird.

I don’t think they achieved either of what appear to be their reasons, and I also don’t think recalibrating even makes sense as a reason. There are good reasons to make a change as I’ll get into below.

But first, here is another reason sometimes cited for never making a change: “If we change definitions, then you cannot compare data sets over time, and we won’t know if autism is really a growing phenomenon or just an artifact of the measurement.” This is true, but not important. When you study anything, you frequently have to take into account changing thresholds and valuations. For example, economists express things in “constant dollars” because prices cannot be compared directly over time. The EPA adjusted auto fuel efficiency ratings at one point, so the ratings before and after that change are not comparable. That is what you deal with when you analyze things. Moreover, culture itself changes, and the incidence of expressed autism is a function of that changing culture, because people who might appear outside the norm in one environment could appear within the norm in a different environment or at a different time.

4. What are the problems with the DSM approach?

The DSM approach to defining autism has several fundamental flaws.

First, the DSM is a taxonomy of brokenness. It only includes disorders, but the idea of a threshold of disorder is nonscientific and doesn’t follow from the taxonomy itself. It should be a taxonomy of all mental states, and should not judge them as good or bad. The NIMH says that about half of the population has a mental disorder in their life (higher if you include addictions). A professor told me that in one experiment, a random pool of adults was sent for screening, and two thirds of them got some disorder label. That should not be shocking: it just means that psychologists are getting better at describing more mental states. There is no scientific basis to exclude a certain chunk of the population on moral grounds that they are deemed “healthy”.

Second, the DSM creates artificial lines between health and disorder. If you don’t artificially exclude a chunk of the population from the taxonomy and you realize that most conditions are part of a larger continuum, then there is no need for threshold lines at all. However, some conditions are the result of a single gene, and those conditions are not a continuum at the causal level. The point is to describe nature the way it is, based on observable measures, not just project manufactured categories onto nature.

Third, the DSM relies on surface behavioral markers for autism, as if to deny that there are any roots. It doesn’t distinguish neurological from psychiatric from behavioral. The reality is that there are physical brain differences, and real cognitive and perceptual differences, which are expressed as common patterns of behavior. But the part you can see is not the thing. You can use the behavior as a screening, or as the first level of understanding, but it should not be written as if the behavior is itself autism.

Fourth, the DSM confuses social with cultural. When we say humans are social animals, we mean we rely on each other, learn from each other, love each other, and live in tribes. When we say humans are cultural animals, we mean that we communicate in symbols and have a shared vocabulary of words and other cultural symbols (architecture, music and so on). The notable thing about autism is the difference in acculturation, not in socialization.

Finally, and this isn’t pointed at the DSM in particular: autism is not a health issue. Autism doesn’t make people sick or give then less chance of happiness than anyone else.

5. On taxonomy

The question that drives how you design a filing system is: “what information will I have when I need to get something out of the filing system.” A diagnostic manual is a taxonomy (or nosology) of mental or physical states – a filing system where the known variable is an expressed behavior and the desired knowledge is the underlying condition. To even understand a diagnostic manual, you cannot be a pure behaviorist; you have to at least accept the notion of underlying conditions. (The swing to behaviorism may explain the cloudiness of the new wording: the authors may no longer accept the basic purpose of the manual.)

The manual categorizes, but people aren’t really in categories – the taxonomy is just a tool, and we should not confuse the tool with reality. People are not strictly categorizable because we change, and because we are in relationships and environments that largely affect how our mental state is expressed and how functional it is. It is still useful as long as we remember the reason for having the tool (connecting people with solutions, and having common vocabulary), and we don’t worship it (we understand that people are not really in categories), and we understand that expressed behavior is not the same thing as a mental state.

The ideal manual would allow a diagnostician to look up notable behaviors in an index and find matching conditions, then read the conditions to determine if they fit. A condition fits if it puts those behaviors in a structural context that explains the underlying condition, and this sheds light on the whole. The ideal manual has differentiators (factors that distinguish conditions) that are close to the root of the condition, and therefore unlikely to relate to many different conditions.

The ideal manual is not concerned with the extent of behaviors, unless the extent is a true differentiator. The amount of behavior is not the amount of a condition, and the amount of a condition is not the amount of need.

6. My proposed definition

I propose the following for the DSM-V.

Short description: Autism is a condition whose name implies being “of the self,” or independent. Sensory experiences are less filtered, and are relatively uninformed by the culture of which the person is a part, often leading to isolation. In everyday language, the person is said to be “in their own world”, not readily adapting to the fabric of the family and larger society.

A. At least one aspect of all four of the following principal factors, unless masked by compensations:

(1) Focus – a single sensory focus at a time (for example, not processing visual and auditory information at the same time); having long-lasting highly specific focus of interest (such as wheels or other parts of objects, a certain toy, naming the presidents, or an academic subject); a tendency to focus attention in an undivided way, rather than the divide attention among many competing stimuli

(2) Sensitivity – abnormal sensory sensitivity to high stimuli (such as bright or flashing lights, or sudden noises) causing a strong reaction; abnormal sensitivity to very low stimuli (such as hearing things that others do not hear); abnormally low sensitivity to some stimuli (such as not noticing cold or pain)

(3) Non-acculturation – lack of adoption of language (such as not speaking, or delayed start of speech); lack of acquisition of language in context (such as the meaning of gestures, sarcasm, or non-literal statements; or lack of use of gesture or eye contact); lack of interest in or conformity to culturally significant topics (such as manners, popular songs, or sports) unless one of those things happens to be a focus of interest; not imitating the behavior of others

(4) Naivete – lack of development of strategic social game playing in teen years (such as taking sides, manipulation, awareness of social conflict); maintaining a naive, guileless, or childlike manner into maturity

B. Any if the following associated factors which may be used to identify the condition, but are not always present:

(1) Behavior – self-stimulating behaviors which are either communicative (such as flapping hands when happy) or as a reaction (such as rocking or other repetitive movements which are more pronounced under high stress)

(2) Inconsistency – a high variation of intellectual capacity (unusually strong in one area and unusually weak in another); inconsistency of language use over time (such as becoming non-verbal or “shutting down” periodically)

(3) Compensations – a repertoire of appearances and behaviors which mask any factor listed in part A

(4) Isolation – low ability to develop peer relationships, or lack of interest

(5) Routine – adherence to specific routines or rituals, often as a technique to retain autonomous control of the environment

In the case of B(3), compensations may render the condition difficult to detect. In this case, more intense examination of focus, acculturation, sensitivity, and naivete are required.

7. How do we allocate services?

If the DSM would include a non-judgmental and more explanatory definition as proposed above, and would not attempt to exclude so-called normal people from its taxonomy, then we will have come a lot closer to accepting all people as we are, while also having a way for people to systematically identify treatments and other opportunities for growth. However, we have not touched on the issue of allocating scarce resources, which is a separate issue.

Allocation is the question of who gets services such as special education services and disability income support. It is important to get this right, because getting enough of the right kind of support early can put someone on a path that is more successful and ultimately costs less or costs nothing in the long term. Also, there are at least a few people who will try to get as much as possible at public expense, regardless of their level of need, so there must be reasonable limits.

A person’s level of need is almost completely unrelated to the kind or “amount” of autism that they may have. Therefore it is nonsensical to use diagnostic criteria as a basis for allocation. Instead, agencies that fund disability-related services should have their own metric that they use to gauge the level of need based on observation of the person functioning in their environment. These metrics should allow for change over time, and should not have artificial limits that are tied to disability categories. For example, one autistic person adult who is blind and has not completed high school may need income support and job training, and 24 hour supervision, while another autistic person may need only help with transportation. The first person costs a lot more, but she is not necessarily more “severe” according to the diagnostic criteria. The second person may require less money, but could have more pronounced autism characteristics.

This article is another systematic way to assess level of need


3 responses to “Autism diagnostic criteria in DSM-V

  1. Lucy says:

    Very interesting, thank you.

    I agree that the new criteria are subjective and judgmental. As you point out, they don’t account for the ways that friendship (for example) is blocked or facilitated by culture and environment. There’s neurotypical bias in judging the quality of such friendships anyway.

    “A person’s level of need is almost completely unrelated to the kind or “amount” of autism that they may have.” — I agree. But although the dx shouldn’t determine services, the reality is that it is very influential in this regard and a vital gateway for many people.

    At the AANE (where I’m on the Board) we are very concerned about the impact of the new criteria. I don’t see this as a rush to judgment: it’s about caution — anticipating the likely consequences for individuals and families based on the evidence so far, and working to ensure the committee is fully aware of those potential consequences and guarding against the negative ones. The DSM committee is responding to various competing priorities and pressures. What works for researchers (greater heterogeneity within categories) doesn’t necessarily work for individuals who need ways to determine and access the supports that can help them.

    Recalibration is a major goal of the exercise — some committee members have acknowledged that this is about screening out the more nuanced cases, people who might not be truly autistic — but of course these mechanisms are clunky and some people who need services and protections will now not meet the criteria for dx. A study from LSU is indicating that the people who’ll be disqualified under the new criteria have the same symptom severity as the people who won’t be. Will some new diagnostic label be created for them? In any case, we need more research on the impact.

    We should also remember what a struggle it was, over many years, to get Asperger’s into the DSM in the first place — to have it formally recognized as part of the autism spectrum. It seems there’s still a huge misunderstanding about what Asperger’s is — even the committee members talk about it as “mild”. In fact some of the committee members seem amazingly tin-eared and prone to blurting out comments that do anything but reassure those with concerns. Hence the lack of trust.

  2. Luke Allen says:

    I would love to see the DSM-VI include Culturally Aggressive Survivalist Disorder (neurotypicalism) among its catalog. Just because their numbers are statistically greater doesn’t mean they’re less disordered.

    I would also like to see Nonverbal Learning Disorder and Hyperlexia rolled into the Autism Spectrum Disorders.

  3. Jessie says:

    Your system makes a lot more sense, but while we have a type of capitalism without guaranteed health care and other basic needs, individual people will need stigmatizing “broken” labels to get any help. No amount of refining and redefining is going to change that.

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 )

Google+ photo

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

Connecting to %s

%d bloggers like this: