Talking therapies adapted for autistic adults

On average, psychological and social interventions help autistic adults a little. A 2026 meta-analysis pooled 35 studies of interventions for autistic adults, 1,631 people in all, covering mental health, social cognition and education or work outcomes.¹ The 16 studies that measured mental health found a small to moderate benefit (an effect size of 0.35).¹ The pooled effect was larger for quality of life, from eight studies, and for social cognition, from sixteen, but both were imprecise, and neither is a measure of talking therapy.¹ The effects on behavior, thinking skills, daily-living skills and employment were not statistically clear.¹ The studies compared the interventions with treatment as usual or with a waiting list, and the review took in randomized trials and studies that did not randomize people, while leaving out single-case and before-and-after designs with no comparison group.¹ The reviewers' conclusion, shared by the network meta-analysis below, is that autistic people should be offered the treatments given to everyone else, following person-centered care.^1,4

In real services the picture is a little less bright. In England, national health records for 2012 to 2019 held 8,761 autistic adults who had completed a course of psychological therapy. Matched with adults without an identified autism diagnosis, the autistic group was less likely to improve reliably (56.1% against 61.7%) and more likely to get reliably worse (9.2% against 7.2%).² Most autistic adults in the study still improved.

A 2025 network meta-analysis pooled 67 randomized trials and 3,604 people to compare non-drug interventions for anxiety, depression and quality of life.³ The participants were children, adolescents and adults together (average age about 18), about 7% were female, and most studies recruited people described as high-functioning, with a normal or near-normal IQ and some spoken language.³ Within that network, mindfulness-based programs ranked highest for anxiety, cognitive behavioral therapy (CBT) for depression, and physical activity for quality of life.³ The rankings estimate the probability that a treatment is among the best, not proof that it beats the others. An age subgroup analysis gave larger anxiety effects in adults than in children and adolescents; for the other outcomes the age pattern was mixed.³ The certainty the authors gave was moderate for CBT against control for anxiety and depression, and low or very low for the other comparisons.³

An earlier network meta-analysis of 71 trials took a more cautious view. Its network comparisons found no reliable differences between treatments, at very low certainty, though some forms of CBT "may decrease" anxiety and depression and mindfulness may help adults who already had a mental-health condition.⁴ Hints from direct comparisons in adults went to some forms of CBT for anxiety and to self-directed adapted CBT or skills training for depression, both at low certainty.⁴ Its authors recommend that autistic people are given access to the mental-health interventions available to non-autistic people, following person-centered care.⁴ A 2016 meta-analysis of CBT, across all ages, found only a small effect on anxiety and depression when people rated their own symptoms, which was not statistically clear, and larger effects when relatives or clinicians did the rating.⁵

A 2026 review of how these interventions have been studied found that most reuse existing CBT or mindfulness protocols, so the range on offer is narrow.⁶ The trials behind these numbers mostly recruited people without intellectual disability and without a spoken-language difference; one review could not tell which interventions worked for people with intellectual disability.⁴

One therapy has been tested specifically in suicidal autistic adults. In a Dutch trial at six mental-health centers, 123 autistic outpatients aged 18 to 65 with suicidal behavior were assigned to dialectical behavior therapy (DBT) or to usual care.⁷ By the end of treatment DBT had reduced suicidal thoughts and suicide attempts more than usual care, but both differences were no longer statistically significant at the 12-month follow-up, which the authors attribute to the usual-care group improving rather than to the therapy wearing off; they describe the benefit as short-term. Depression fell and stayed lower, and there was no clear effect on social anxiety. Two people in the usual-care group died by suicide.⁷ Judged from the health service's side, the same trial found DBT cost slightly less than usual care while adding quality-adjusted life years; judged from a wider societal side it cost a little more for each year of good health gained.⁸

A 2026 meta-analysis pooled four randomized trials of DBT-based interventions, 375 people in all, from school-age children to adults (average ages 10.2 to 37.9 years).⁹ It reported large improvements in emotion regulation, suicidal thoughts and depression, but the estimates for suicidal thoughts and for depression each pool only two trials, one in adults and one in adolescents, with an I-squared of 93 to 98%, and it found no clear effect on anxiety. The adult share of that evidence is one trial.⁹

Trauma is common in autistic people's lives, and treating it has barely been studied. A 2026 systematic review, searching the literature to 2022, found only four studies of PTSD treatment in autistic people: three described a single person each, and one was a non-randomized group study.¹⁰ In that group study, eight sessions of eye movement desensitization and reprocessing (EMDR) were added for 21 autistic adults after six to eight weeks of usual care; their PTSD symptoms had not improved with usual care but did improve after the EMDR sessions, and held at a six- to eight-week follow-up.¹¹ The reviewers note that trauma-focused CBT, the first-line treatment for PTSD in the general population, had not been systematically tested in autistic people, and that autism-specific adaptations of it had only been proposed, not tested.¹⁰ No randomized trial of trauma-focused therapy in autistic adults was found as of October 2026.

Two newer studies address the gap without closing it. A 2026 pilot gave 17 autistic young people aged 10 to 17 an adapted, telehealth form of trauma-focused CBT, and their PTSD symptoms fell across self, caregiver and clinician report; it was a single-arm pilot, not randomized, and it was in children and adolescents, not adults.¹² A 2026 study of 175 people in intensive trauma-focused treatment found that those with high autistic traits improved as much as those with low traits, but the participants were not a diagnosed autistic sample and there was no control group.¹³

The prevalence picture is itself argued over. A 2025 meta-analysis of diagnosed PTSD put lifetime rates in autistic adults at about 2.7%, close to general-population figures.¹⁴ A 2026 systematic review, collating cases across studies that used diagnostic interviews, put the lifetime diagnosis rate higher, at about 16% with a range of 0 to 20%, and reported that autistic adults and children have more severe PTSD symptoms than non-autistic peers, at at least comparable rates.¹⁰ Both note that screening tools find much higher rates still, which the 2025 authors read as the diagnostic criteria missing how trauma shows in autistic people.¹⁴

What a study actually changes to suit autistic people is rarely spelled out and almost never tested. In a 2026 review of 69 studies of psychological interventions for autistic adults, the adaptations most often described, across the 42 adult-only studies (a further 27 mixed adults and children), were the use of concrete supports and language, and less text with more visual supports (18 studies each), clear and concise content (13) and therapists experienced in working with autistic clients (12).⁶ The authors note that studies gave examples rather than full lists, so this undercounts what is done, and that only four studies discussed whether the adaptations helped.⁶ They found no agreed best set of adaptations, and only about 14% of the outcome measures used had been validated with autistic people.⁶

What to ask for, then, is the list the studies support: therapy tailored to the person, changes to the sensory environment or remote sessions, longer or more sessions, written or visual information, practical as well as emotional help, a clinician who knows about autism and understands how varied it is, and one who supports a positive autistic identity and connection to the autistic community.^15,16 These findings come mostly from samples that were without intellectual disability, and the first-person studies were mostly White and in Western countries.¹⁵

Talking therapy is not risk-free for autistic adults, and how often it goes wrong is largely unmeasured. In the 2023 network meta-analysis of 71 trials, only 9 reported how many people had an adverse event.⁴ In the 2025 network meta-analysis the average drop-out was 13.6% of people in the therapy groups and 11.2% of those in the control groups, and it was higher for CBT (15.9%) and for behavioral and functional training (17.2%).³ In the English records study, autistic adults were more likely than others to get reliably worse during treatment.²

The commonest barrier reported is the therapist. The most commonly reported barrier across studies of autistic people of all ages, reported by them or by their parents, is therapists' lack of knowledge about autism, or an inability or unwillingness to tailor their approach.¹⁷ Reviews of autistic adults' own accounts put clinicians' lack of understanding and their inflexibility among the central barriers, and one concluded that services "do not adequately support autistic adults and can even cause additional harm".¹⁸

Autistic writers describe the same mismatch between what therapy assumes and how autistic minds work. Daniel Burford, writing about England's national talking-therapy service and the records that show autistic adults improving a little less often, puts it plainly: "Most people conclude they failed at therapy. That isn't what happened."^c1 He argues that cognitive therapy asks a person to challenge a distorted thought, when an autistic person's prediction that a busy office will be unbearable, or that a party will cost them three days, is usually accurate, so being asked to argue against a true belief does not take.^c1 His advice is to say the hard thing out loud: "A therapist worth having will change how they ask."^c1

Megan Griffith, an AuDHD therapist, writes that she does not want "a therapist or coach who is only understanding of my particular neurodivergence while holding stigmatizing beliefs about other types of neurodivergence", and treats the working relationship as something to repair: she suggests practicing "rupture and repair" when sessions go wrong.^c2