I think it’s fair to say that there are a lot of problems with the way that we currently do diagnostic assessments for autism and other divergent neurotypes.  This is by no means an exhaustive list of all the issues, but I thought I would make an effort to identify some of the ones that I think are most important.

• One important problem is that our diagnostic systems are inaccessible.  Parents are likely to notice some evidence of atypical neurodevelopment long before doctors and other professionals actually refer an autistic child to get a diagnosis.  Doctors are trained to have a certain amount of confidence in their own judgement, which is often fine and even necessary, but this becomes more problematic when doctors are trying to rule on questions of neurodevelopment, since doctors don’t tend to know much about autism.  This can create delays that not only prevent people from accessing supports, but it may leave parents distrustful of established scientific and medical authority. This distrust may make people more vulnerable to the expensive, and often dangerous, cures and treatments peddled by alternative practitioners and quack doctors.
• Another problem with accessibility is a financial one.  In various jurisdictions, diagnostic assessments for children can be covered as part of public healthcare systems, at least once people have a referral.  Some jurisdictions even cover assessments for adults, which is vitally necessary!  However, in other places, access to these assessments is privatized – requiring either private medical insurance or out-of-pocket spending.  I think such privatization is not acceptable.[1]
• After people have been referred to a diagnostic assessment, we tend to focus on evaluating them for some given neurotype or condition.  For example, we might try to assess whether they are autistic or not.  The ADI and the ADOS, the main diagnostic instruments for autism that exist today, were designed to answer precisely this question.[2]  I don’t think this question is particularly practical, because neurotypes often occur together, as do neurotypes and various mental conditions.  If a given autistic person also has ADHD and anxiety, we need to make sure that they are given all three diagnoses.[3]  Thus, whatever protocol we follow in a neurodevelopmental or psychiatric assessment should include a mandatory component of at least screening to rule out other neurotypes and conditions.  Ideally, clinicians should have the training and expertise[4] to actually assess whether the individual meets criteria for any neurotype or condition that cannot be easily ruled out.
• A related issue is a lack of genetic testing.  To be sure, there are a lot of controversies surrounding genetics, particularly as they relate to gene editing, selective abortions, and other domains of eugenics. These ethical questions are crucial and we mustn’t ignore them.  However, if we’re just speaking in terms of diagnosing somebody who already has a genetically-defined neurotype, I think that having the information offered by the diagnosis could be very helpful: it could clarify the nature of the neurotype.  Unfortunately, we don’t do enough genetic testing as part of diagnostic assessments out in the community, even when people have low scores on cognitive assessments (IQ tests) or if there are other signs that a genetic diagnosis might be possible.
• The diagnostic assessment may include a lot of interaction with the autistic person in a clinician’s office.  This is not a typical social setting, so it is unclear how results might map onto someone’s ability to function in the real world.  This is a particularly big issue for autistic teenagers and adults, as we have often spent many years learning to camouflage our autistic features.  I suspect that I would now appear superficially typical if I was interacting with someone in a doctor’s office – you might have to take me somewhere else, like a party, before my social challenges became obvious.  In short, we need more naturalistic methods of assessment.
• We’re also increasingly aware of the fact that autism presents in different ways in different people.  Most notably, a lot of recent research has revealed differences in the way autism appears in men and women – the research suggests that there is a “female autistic phenotype.”  Furthermore, these differences seem to result in many women being diagnosed substantially later than men.  We not only need to make doctors and clinicians in communities aware of these realities (because many remain staggeringly ignorant of autism in women and girls), but we also need to continue exploring whether there might be other autistic phenotypes.  When we do so, we should resist the issue to view these phenotypes as categorical in nature – they are almost certainly continuously distributed.  I have no doubt that many autistic men have an autistic phenotype more similar to the “female” type than to the stereotypical “male” type, and that is even without considering the fact that many autistic people have diverse gender identities beyond this simple male-female binary. I don’t think we should define autism differently in these populations, but we need to be sensitive to how autism presents differently in them.
• It is probably a good idea to have multidisciplinary assessment teams, instead of individual clinicians, providing diagnoses.  This could be particularly useful if we are broadening the diagnostic evaluation beyond one thing (like autism) and trying to gather information relevant to different neurotypes and conditions.
• Moreover, whether multidisciplinary or not, diagnostic assessments are always done by professionals with clinical/medical/etc. training and experience. This sort of expertise is hardly useless, but the fascinating critique of the ADOS by Timini et al. (2019) highlights how such expertise is far more subjective and far less objective than often supposed. The power and authority of the examiner are not irrelevant, nor are the biases of the examiner irrelevant. While autistic people might struggle to understand others’ thoughts and perspectives, neurotypical people also struggle to understand the autistic perspective (see, e.g., Edey et al., 2016; Heasman & Gillespie, 2017), which might partly reflect the unfamiliarity of the autistic perspective and might partly reflect the general tendency for those with power to invest less effort in understanding the perspectives of those less powerful than themselves. If the examiner and those who designed an assessment struggle to see and understand the perspective of the population being assessed, the quality of the assessment will suffer. Furthermore, an assessment like the ADOS inherently involves an interaction between the examiner and the examinee, and the examiner’s behaviour in this interaction is far from irrelevant. Thus, it is essential for autistic people to be involved in the design of assessment tools, and while expert professionals (most of whom will not be autistic) must necessarily deliver assessments in community settings, these expert professionals must be sensitive to the perspectives of the populations they assess. This is no different from the need for cultural sensitivity in diagnostic evaluations with minority and marginalized populations – which is, of course, another important area where diagnostic practices have room to improve.
• Many diagnostic tools place a great deal of emphasis on parent reports. This is fine for children who have lived with the same caregivers for their entire lives: parents may have biases, just like anyone else (including autistic people themselves), but parents know more about their own children’s early development than anyone else. However, not everyone is a child living with their parents. For example, some autistic adults I know are pretty much estranged from their families, and then there are adults with intellectual disabilities in community living facilities or institutions. I was delighted to see the ADOS has now been adapted for minimally verbal adults, as this does help with this problem, but we still need more work on self-report measures and interviews.[5]
• Assessment tools like the ADI and the ADOS are often proprietary, owned by private companies that charge high fees to distribute test materials.  This, of course, imposes additional cost burdens on individuals, families, and public healthcare systems.  Furthermore, to me, there is something grossly obscene, something outrageous, about a private company controlling access to a diagnostic evaluation that represents an identity for a designated group of people!  An autism diagnosis is not merely a passport to access services and supports, but also something that defines who we are as people.  Is that truly something that can legitimately be owned?
• Once a diagnosis has been issued, this is often seen as the end of the diagnostic process.  However, there should definitely be follow-up consultation in the short term, because adjusting to the diagnosis may take some time, and people may need more than one conversation to understand everything they need to know about their neurotype or that of their child.
• There needs to be longer-term follow-up to monitor whether the constellation of diagnoses an individual is eligible for shifts over the years.  I’m not so much thinking here of autism itself,[6] but instead of changes in things like mental health conditions.  For example, if we diagnose autism in a toddler or a preschooler, we might want to check in again when the individual is attending school to see if mental health challenges are emerging.

What do you think of this list?  Can you think of any other problems I missed?  If so, please add a comment!

---

[1] It’s also incredibly inefficient.  Just look at healthcare costs in the United States – in that country, huge amounts of money are wasted on unnecessary healthcare costs like advertising and marketing, the creation of an atmosphere of luxury in elite institutions, and on administrative efforts to keep track of all the money flowing around.

[2] Everybody calls these tools the “gold standard,” which I frankly find a little disturbing.  If we accept the idea that the ADOS and the ADI represent some ideal of what a diagnostic instrument should be, it would follow that it is impossible to develop better instruments, so why should one bother trying? They’re good measures, and they were a massive advance over the old CARS when first introduced, but that doesn’t mean we should give up trying to do better. We can at least develop a more diverse and flexible suite of tools.

[3] Not only are the additional diagnoses information that helps us better understand what is going on, but they may be of vital importance when we are trying to figure out how to best support people.  For example, it is probably easier to treat co-occurring anxiety than to work on teaching social skills to autistic person.  Indeed, treating social anxieties may actually have a larger effect on social comfort and competence than social skills training!

[4] Unfortunately, we do have a shortage of clinicians with autism expertise, and because of the high prevalence of mental health challenges in autism, these clinicians should all be comfortable dealing with mental health.

[5] There are also other important populations, like people with motor/catatonic problems and/or severe attention problems who cannot communicate through speech but who (unlike the minimally-verbal population that is the focus of the new ADOS module) can understand language and communicate through typing. How is diagnosis to be approached in such cases?

[6] Research does show that autistic people can sometimes become ineligible for an autism diagnosis – this is called an “optimal outcome.”  However, the term is misleading: autistic people with “optimal outcomes” may still have serious challenges that simply fall short of a diagnostic threshold.  I am not sure that it is useful to punish autistic people for effective camouflaging by taking away their diagnosis.