Levels in Autism Assessments

Assigning a level is one of the hardest parts of an assessment. This blog outlines how I try to approach it.

The levels in the DSM-5-TR describe how much support a person needs. They are not a ranking of intelligence, verbal fluency or capability. Confusion can come partly from the layout: the DSM offers a table of illustrative traits under each level rather than a definition of the levels themselves. Read in isolation, that table suggests someone who speaks fluently could never require substantial support, even if they cannot sustain work or manage daily living without assistance. Read alongside the introductory chapters, which state clearly that clinical judgement takes precedence over a recipe approach, that reading does not hold.

Cognitive ability and autism are independent of one another. Strong cognitive abilities often allow a person to compensate and mask for decades at considerable cost, and those same abilities are then used as evidence that they cannot really be autistic. (On a side note, please note that I do not administer IQ or cognitive testing and I do not formally assess intellectual giftedness).

Speech is not the same thing as communication, either. A person can be articulate in a quiet room with a familiar clinician and be unable to say what they need when they are overloaded, or lose speech altogether in a shutdown period. Situational mutism sits alongside fluent expressive language far more often than the level descriptors would suggest.

Support needs can also change. Capacity shifts across a day, a week and a setting, which no single number can capture. One way to look at levels is this- if we were to strip all existing supports away, what are we left with? How much support does the person need?

What I look at, then, is the sustainable picture rather than the peak one. Assessment appointments tend to catch people at their most prepared and most masked, and a few hours of capacity says very little about a year of it. I look at how much work or study can be held long term without tipping into burnout, shutdown or overload, how long recovery takes afterwards, how the tasks of daily living are managed.

I also look at where capacity sits across different domains, because it is rarely even. Someone may be flourishing  in their job, while meals, showering and the state of the house have quietly fallen away. Trajectory matters too. Capacity that has dropped sharply following burnout tells a different story from capacity that has always sat where it sits. So does reserve, meaning what happens when something goes wrong, when there is an illness, a move or a change of manager, which tends to be far more revealing than how things run when nothing does.

Self-report in high-masking adults tends to run low, not because people are unreliable but because they have spent decades treating an extraordinary level of effort as ordinary. Informant accounts help here, as does what has actually happened: jobs that ended, courses that were dropped, bills that went unpaid, relationships that could not be maintained alongside everything else.

Careful attribution matters as well. Where other diagnoses are in the picture, I (try!) to think about which difficulties belong to what. Someone with a significant trauma history or a mood disorder may need a great deal of support without that support need arising from their autistic neurotype, and the reverse is equally true. Defaulting to autism is a poor habit, and so is defaulting away from it.

Much of this difficulty is an artefact of where the levels sit. Functional capacity would be far better described as its own dimension, applied consistently across every diagnosis, rather than embedded inside the autism criteria as though it were a feature of the neurotype itself. That would let us say what a person actually needs without forcing the answer through a single label.

So, when we meet, I briefly explain the levels, I ask the person what they think theirs is, and we work towards a shared position. They have been living inside their own capacity their whole life, which makes their view rather good data. The level as reviewable, because support needs change and a level assigned at one point in someone's life should never be treated as a permanent fact about them. Clients often arrive having been assigned a level with no idea what it means or where it came from, which leaves them holding a diagnosis they cannot use.