Module 10 — A Therapist Who Gets It

 

How to find one, what “neuro-affirming” should mean in practice rather than on a profile page, and what a bad fit looks like early.

 

Self-Identification · Part Four — Other people

 

The big idea

 

Short on capacity today? The big idea: The thing that usually goes wrong is not that the therapist is bad at therapy.

 

Most are competent. But competence with the general population does not transfer to this automatically, and when it fails to transfer the failure gets written up as something about the client — not engaging, intellectualising, resistant, poor insight. Plenty of people arrive here having been told four or five times that the mismatch was theirs.

 

It mostly was not. This is a fit problem with a research literature behind it, and the useful skill is not persistence. It is knowing what to look for, and how early you are allowed to conclude you have not found it.

 

Step 1 — The lesson

 

A. The fit problem is real, and it has been measured

 

Start with the people receiving the care.

 

Two hundred autistic adults aged 18 to 67 were surveyed about treatment and support for mental health difficulties, self-injury and suicidality. Three themes came out: difficulties accessing treatment; “lack of understanding and knowledge of autistic people with co-occurring mental health difficulties” — the professionals’ knowledge, not the participants’; and the fact that appropriate support, or its absence, moved people’s wellbeing and their sense of whether suicide was in their future.

 

Now the people delivering it. A study interviewed 22 autistic adults with mental healthcare experience, 44 community mental health clinicians and 11 agency leaders. Across all three groups, “most of the reported barriers involved clinicians’ limited knowledge, lack of experience, poor competence, and low confidence working with autistic adults.” It also found autistic adults could be “turned away from services when they contact the mental health division and disclose their autism diagnosis during the intake process.”

 

Then the gap measured directly. One hundred community mental health clinicians were asked about starting cognitive behavioural therapy with autistic and non-autistic adults, and reported weaker intentions, less favourable attitudes and lower self-efficacy for the autistic clients. Same clinicians, same therapy, different confidence.

 

And at scale: 498 psychotherapists rated their training and competence across a list of diagnoses. They rated autism lowest of everything surveyed — for education received in training and for competence in both assessment and treatment — including against conditions of comparable prevalence. Many held “misconceptions and outdated beliefs.” Few had done additional training; most wanted to.

 

The conclusion is not that therapists are hostile. It is that most were never taught this, know they were never taught it, and would like to be. You are not being fussy. You are shopping in a market with a published, self-reported shortage of the exact thing you need.

 

Diagram — A · One tool the kit does not carry. Start with the people receiving the care. Two hundred autistic adults aged 18 to 67 were surveyed about treatment and support for mental health difficulties, self-injury and suicidality, and three themes came out: difficulties accessing treatment; lack of understanding and knowledge of autistic people with co-occurring mental health difficulties — the professionals' knowledge, not the participants'; and the fact that appropriate support, or its absence, moved people's wellbeing and their sense of whether suicide was in their future. Now the people delivering it. A study interviewed 22 autistic adults with mental healthcare experience, 44 community mental health clinicians and 11 agency leaders, and across all three groups most of the reported barriers involved clinicians' limited knowledge, lack of experience, poor competence and low confidence working with autistic adults; it also found autistic adults could be turned away from services when they contact the mental health division and disclose their autism diagnosis during the intake process. Then the gap measured directly: one hundred community mental health clinicians were asked about starting cognitive behavioural therapy with autistic and non-autistic adults, and reported weaker intentions, less favourable attitudes and lower self-efficacy for the autistic clients — same clinicians, same therapy, different confidence. And at scale, 498 psychotherapists rated their training and competence across a list of diagnoses and rated autism lowest of everything surveyed, for education received in training and for competence in both assessment and treatment, including against conditions of comparable prevalence; many held misconceptions and outdated beliefs, few had done additional training, and most wanted to. That is the empty pocket in the drawing, and the note in it is in the profession's own handwriting. The conclusion is not that therapists are hostile. It is that most were never taught this, know they were never taught it, and would like to be. You are not being fussy. You are shopping in a market with a published, self-reported shortage of the exact thing you need.

 

B. What “neuro-affirming” should mean, in behaviours

 

The word is now on thousands of profiles and it costs nothing to type. Treat it as a claim to be checked, not a credential. Here is what it looks like when it is real — all of it observable inside two sessions.

 

They adapt the communication, rather than expecting you to. Open questions with no edges — and how did that make you feel? — are hard for a lot of people to answer, and the difficulty is not avoidance. A therapist who gets it notices you stalling and offers a structured version, or a scale, or a choice of three. The adjustment happens on their side.

 

They do not read your face as data about your engagement. Flat delivery, little eye contact, describing something devastating in a level voice, looking away all session because looking away is how you think — none of these mean what the textbook says. A therapist who gets it says so early and takes it off the table.

 

Masking is not the goal. This is what separates the profiles from the practitioners. A great deal of therapy aimed at autistic adults is, functionally, coaching in better performance: more eye contact, warmer small talk, fewer visible differences. That may make other people more comfortable. It is not a mental health outcome, and for many adults it is the mechanism of the exhaustion they came in with. The goal is a life that costs you less, not a version of you that reads as more typical.

 

They will change the conditions of the session. Camera off, or audio only. No small talk at the start. Forty minutes instead of sixty, or ninety every fortnight instead of fifty every week. Notes sent afterwards. Permission to fidget or to look somewhere else while talking. A therapist who gets it treats these as ordinary configuration rather than concessions, and offers some before you ask.

 

They accept writing. A short written message between sessions — the thing you could not say out loud, the thing you only worked out on Thursday — is part of the work, not a boundary violation to be managed. A clear limit is fine: “send me anything, I’ll read it before we next speak” and “I can’t do between-session contact, so let’s build in ten minutes at the start for what you’ve written down” are both good answers. Silence, or being treated as a nuisance, is not.

 

They will say what they do not know. “I’ve worked with three autistic adults and I’d be learning alongside you” is a better sign than an unbroken confidence, and it is the answer the 498-therapist study predicts you should be hearing more often than you do.

 

Diagram — B · A stamp, not a credential. The word neuro-affirming is now on thousands of profiles and it costs nothing to type, so treat it as a claim to be checked rather than a credential. Here is what it looks like when it is real, and all of it is observable inside two sessions. They adapt the communication rather than expecting you to: open questions with no edges — and how did that make you feel — are hard for a lot of people to answer, and the difficulty is not avoidance; a therapist who gets it notices you stalling and offers a structured version, or a scale, or a choice of three, and the adjustment happens on their side. They do not read your face as data about your engagement: flat delivery, little eye contact, describing something devastating in a level voice, looking away all session because looking away is how you think — none of these mean what the textbook says, and a therapist who gets it says so early and takes it off the table. Masking is not the goal, and this is what separates the profiles from the practitioners: a great deal of therapy aimed at autistic adults is functionally coaching in better performance — more eye contact, warmer small talk, fewer visible differences — which may make other people more comfortable but is not a mental health outcome, and for many adults it is the mechanism of the exhaustion they came in with. They will change the conditions of the session: camera off or audio only, no small talk at the start, forty minutes instead of sixty, or ninety every fortnight instead of fifty every week, notes sent afterwards, permission to fidget or to look somewhere else while talking — treated as ordinary configuration rather than concessions, and some of it offered before you ask. They accept writing: a short message between sessions is part of the work, not a boundary violation, and a clear limit is a fine answer where silence or being treated as a nuisance is not. And they will say what they do not know — I have worked with three autistic adults and I would be learning alongside you is a better sign than an unbroken confidence.

 

C. What the evidence says about adapting therapy

 

Two things are true at once here, and you should have both.

 

The first: the adaptations are real, specific and documented. A three-round Delphi study put 221 candidate statements about delivering cognitive behavioural therapy to autistic people through expert rating by clinicians and clinical researchers in England; 155 were rated essential or important. Adaptations to the structure and process of therapy were consistently endorsed — the mechanics, not the model. Among the endorsed items: “allowing clients to do things that make them feel more comfortable such as fiddling with objects or not making eye contact,” using “a range of methods to facilitate communication,” and checking the client’s understanding of the social rules of the appointment. Rather than a single autism protocol, the panel favoured an individualised, formulation-driven choice of technique.

 

A smaller consultation with eight specialist therapists, led by a researcher with lived experience, named the list from the other direction: increase collaboration, support emotional literacy, use visual prompts, be consistent, accommodate sensory needs, and avoid metaphors. If you have ever been asked what colour your anger is, that last one is a finding, not a preference.

 

The second thing, which the honest version of this module has to include: the outcome evidence is thin. A 2026 systematic review found only 10 randomised controlled trials, 537 participants in total, most with small samples and high risk of bias, and concluded that CBT “showed very uncertain effects across most outcomes in autistic adults.”

 

That is a reason to hold the right expectation, not to skip therapy. Nobody can promise you an evidence-backed protocol here, because it does not yet exist at that strength. What exists is a documented set of adjustments to how therapy is run. So judge a therapist on whether they do the adjusting — not on whether they claim an evidence base nobody has.

 

Diagram — C · Agreed on paper, thin on the plate. Two things are true at once here, and you should have both. The first: the adaptations are real, specific and documented. A three-round Delphi study put 221 candidate statements about delivering cognitive behavioural therapy to autistic people through expert rating by clinicians and clinical researchers in England, and 155 were rated essential or important. Adaptations to the structure and process of therapy were consistently endorsed — the mechanics, not the model. Among the endorsed items: allowing clients to do things that make them feel more comfortable such as fiddling with objects or not making eye contact, using a range of methods to facilitate communication, and checking the client's understanding of the social rules of the appointment. Rather than a single autism protocol, the panel favoured an individualised, formulation-driven choice of technique. A smaller consultation with eight specialist therapists, led by a researcher with lived experience, named the list from the other direction: increase collaboration, support emotional literacy, use visual prompts, be consistent, accommodate sensory needs, and avoid metaphors. If you have ever been asked what colour your anger is, that last one is a finding, not a preference. The second thing, which the honest version of this module has to include: the outcome evidence is thin. A 2026 systematic review found only 10 randomised controlled trials, 537 participants in total, most with small samples and high risk of bias, and concluded that cognitive behavioural therapy showed very uncertain effects across most outcomes in autistic adults. That is the signed card and the nearly empty plate in the drawing, and it is a reason to hold the right expectation, not to skip therapy. What exists is a documented set of adjustments to how therapy is run, so judge a therapist on whether they do the adjusting rather than on whether they claim an evidence base nobody has.

 

D. What a bad fit looks like early

 

You can usually tell inside three sessions. Four signals, in rough order of how often they turn up.

 

The identification becomes the presenting problem. You mention you have worked out you are autistic, and the work quietly reorients onto that: why you need the label, what you are avoiding by having it, whether this is really about your mother. There is a legitimate version of that conversation, once, early. What is not legitimate is the topic never handing back. If four sessions in you are still defending the description rather than using it, that is the fit failing.

 

Everything is anxiety. Sensory overload becomes anxiety. Needing plans in advance becomes anxiety. Exhaustion after socialising becomes anxiety, then avoidance, then something to tackle with graded exposure. Anxiety is genuinely common alongside all of this and often needs treating — Module 1 says plainly that the possibility deserves a proper look. But a clinician with one bucket puts everything in it, and the treatment that follows will ask you to keep doing the thing that is depleting you until it stops feeling bad. It will not stop feeling bad.

 

Visible discomfort when you do not perform distress the expected way. You describe something severe without crying, or laugh in the wrong place, or give the timeline before the feeling. Watch what happens next. A therapist who gets it takes the content at face value. One who does not keeps fishing for the affect — that sounds like it must have been very painful? — and, if it does not arrive, may start recording you as detached or as not ready for this work.

 

The outcome measure is normality. Listen to how success gets described. If the picture is more eye contact, more socialising, fewer accommodations, a version of you that draws less attention — that is someone else’s target. Ask what improvement would look like, and check whether the answer is about your suffering or about your visibility.

 

None of these means the person is unkind. Most are one training course away from being useful. That is not your project to run.

 

Diagram — D · The coat does not fit, you got the ticket. The thing that usually goes wrong is not that the therapist is bad at therapy. Most are competent. But competence with the general population does not transfer to this automatically, and when it fails to transfer the failure gets written up as something about the client — not engaging, intellectualising, resistant, poor insight, which are the four tickets pinned to the coat in the drawing. Plenty of people arrive here having been told four or five times that the mismatch was theirs. It mostly was not. This is a fit problem with a research literature behind it, and the useful skill is not persistence; it is knowing what to look for, and how early you are allowed to conclude you have not found it. You can usually tell inside three sessions, and there are four signals. The identification becomes the presenting problem: you mention you have worked out you are autistic and the work quietly reorients onto that — why you need the label, what you are avoiding by having it — and there is a legitimate version of that conversation, once, early, but if four sessions in you are still defending the description rather than using it, that is the fit failing. Everything is anxiety: sensory overload becomes anxiety, needing plans in advance becomes anxiety, exhaustion after socialising becomes anxiety and then avoidance and then something to tackle with graded exposure — and a clinician with one bucket puts everything in it, then asks you to keep doing the thing that is depleting you until it stops feeling bad. It will not stop feeling bad. Visible discomfort when you do not perform distress the expected way: a therapist who gets it takes the content at face value, and one who does not keeps fishing for the affect and may start recording you as detached or as not ready for this work. And the outcome measure is normality: if the picture of success is more eye contact, more socialising, fewer accommodations, a version of you that draws less attention, that is someone else's target, so ask what improvement would look like and check whether the answer is about your suffering or about your visibility. None of these means the person is unkind. Most are one training course away from being useful. That is not your project to run.

 

E. Before you book, and if you have already been through several

 

Six questions, for a first message, an introductory call or a form. Copy them.

 

1. What experience do you have of working with autistic adults? Not do you work with — what experience. You want a number and a story, not a tick box.
2. What do you actually change about how you work with an autistic client? The single most diagnostic question in the list. Good answers are boring and specific: pacing, structure, written summaries, camera off, less open-ended questioning. Bad answers are warm and general.
3. Would you work with someone who identifies as autistic and does not have a diagnosis? Ask it plainly. Many say yes without hesitating; some will not; a few need a diagnosis for funding or service-rule reasons that have nothing to do with believing you. All three are answers you want before you have paid anything.
4. How do you handle contact between sessions, and do you accept things in writing?
5. If I am hard to read — flat voice, not much eye contact — how do you work with that? Listen for whether they treat it as information or as a symptom.
6. How would we know this was not working, and what happens then? A therapist who has a real answer to this is a therapist who has ended things well before.

 

And if you have already been through three, or five. This is common here, and demoralising in a particular way: each ending leaves a bit more evidence for the theory that the problem is you, and that evidence travels with you. Two things.

 

The first is arithmetic. If a large share of practitioners rate their training in this as the weakest thing they have, a run of poor fits is what sampling at random from that population produces. Three failures is not an unlikely event. It is the base rate.

 

The second is that the search improves once you stop starting from scratch. You now know more than most first-time clients about what breaks. Write down what went wrong each time — not the person, the mechanic. Too many open questions. Kept pushing the eye contact. Wanted me to be more sociable. Went quiet when I emailed. That list is your screening criteria, and it turns question 2 from a general enquiry into a test you can score.

 

Where self-identification sits, plainly: therapy does not require a diagnosis. You are describing your own presentation, which is what you are meant to do in session, and Module 8 makes the same point about doctors. Most therapists will work on that basis; a minority will not, and asking costs one sentence and saves a month.

 

Diagram — E · Three strikes is the base rate. Six questions, for a first message, an introductory call or a form. One: what experience do you have of working with autistic adults? Not do you work with — what experience; you want a number and a story, not a tick box. Two: what do you actually change about how you work with an autistic client? The single most diagnostic question in the list; good answers are boring and specific — pacing, structure, written summaries, camera off, less open-ended questioning — and bad answers are warm and general. Three: would you work with someone who identifies as autistic and does not have a diagnosis? Ask it plainly; many say yes without hesitating, some will not, and a few need a diagnosis for funding or service-rule reasons that have nothing to do with believing you, and all three are answers you want before you have paid anything. Four: how do you handle contact between sessions, and do you accept things in writing? Five: if I am hard to read — flat voice, not much eye contact — how do you work with that? Listen for whether they treat it as information or as a symptom. Six: how would we know this was not working, and what happens then? A therapist who has a real answer to this is a therapist who has ended things well before. And if you have already been through three, or five: each ending leaves a bit more evidence for the theory that the problem is you, and that evidence travels with you. Two things. The first is arithmetic — if a large share of practitioners rate their training in this as the weakest thing they have, a run of poor fits is what sampling at random from that population produces, so three failures is not an unlikely event, it is the base rate, which is what the spent matches in the drawing are and what they are not. The second is that the search improves once you stop starting from scratch: write down what went wrong each time, not the person but the mechanic — too many open questions, kept pushing the eye contact, wanted me to be more sociable, went quiet when I emailed — and that list is your screening criteria, which turns question two from a general enquiry into a test you can score. Where self-identification sits, plainly: therapy does not require a diagnosis. You are describing your own presentation, which is what you are meant to do in session. Most therapists will work on that basis; a minority will not, and asking costs one sentence and saves a month.

 

F. What helps

 

1. Treat the first two sessions as an assessment you are conducting.

 

Both of you are deciding. Saying so — “I’d like the first couple of sessions to be us working out whether this is a fit” — is normal practice, and makes stopping at three much easier.

 

2. Ask question 2 before anything else.

 

What do you change about how you work? A therapist who cannot name three concrete adjustments has probably never made any. Faster and more reliable than reading profiles.

 

3. Send the brief in advance, in writing.

 

One page: how you communicate best, what you need a session to be like, what went wrong before, what you want. It saves four sessions of ramp-up, and how they respond to receiving it tells you most of what you need to know.

 

4. Name the affect thing early, once.

 

“I don’t show much on my face and my voice stays flat. It doesn’t mean I’m not affected — please don’t read it as distance.” Said in session one, it heads off the most common misreading in this whole area. How it lands is itself information.

 

5. Leave earlier than feels polite.

 

The best predictor of another wasted six months is staying because you do not want to offend someone who has been kind to you. A bad fit is not a moral event, it is a procurement outcome. You may say the fit is not right and go — no justification, and no reference to the last four.

 

Up next

 

Module 11 — Do I Want an In-Depth Assessment?

 

All modules in Self-Identification