Personality-disorder diagnoses and misdiagnosis

Personality-disorder diagnoses appear in autistic people's records more often than in other adults' records. The largest study used UK primary-care records: 22,112 autistic adults and 221,120 comparison adults matched by age, sex and general practice, followed from 2000 to 2019. Autistic men without an intellectual disability were given a new personality-disorder diagnosis 4.8 times as often as comparison men, and autistic women 4.6 times as often as comparison women. Among autistic people who also had an intellectual disability the rates were 2.0 times as high for men and 8.3 times for women. For borderline (emotionally unstable) personality disorder specifically, autistic women with an intellectual disability were diagnosed 16.6 times as often.¹

Other record studies point the same way at different magnitudes. In US Medicare records for 2016 and 2017, 3.1% of 4,685 autistic adults aged 65 and over had a personality-disorder diagnosis, against 0.1% of 46,850 matched comparison adults, an adjusted odds ratio of 24.1.¹¹ In 2020 US Medicaid records covering 993,965 people with autism of all ages, personality-disorder diagnoses were recorded in 1.75%, against 0.81% of beneficiaries without autism, an adjusted prevalence ratio of 2.97.²¹ The UK cohort's authors also report a Danish population cohort, covering people born from 1980 to 2010, in which 3.4% of autistic people had a personality-disorder diagnosis against 1.2% of the comparison group.¹

Over the period studied, the rate of new personality-disorder diagnoses rose in women, and among autistic people it fell to 2009 and then rose again; the authors say the early years are imprecise.¹

ICD-11 removed several categorical personality-disorder types, including schizoid, avoidant and obsessive-compulsive, for lack of evidence of their validity; the record and clinical studies above were coded with the older categorical types.¹

Borderline personality disorder is the one that overlaps most with autism. A meta-analysis of 19 studies pooled a borderline-personality prevalence of 4% in autistic samples (95% confidence interval 0 to 9) and an autism prevalence of 3% in borderline samples (1 to 8), both within the usual population prevalence of each condition. Eleven of the 19 studies were at high risk of bias, and the authors said the studies could not show whether one diagnosis is being made in place of the other.⁷

A review of 22 studies of autistic adults without an intellectual disability found cluster A and C personality disorders the most frequent co-occurring types, and that roughly 40% to 62% of autistic samples assessed with structured interviews met criteria for at least one personality disorder.²² Two older Swedish samples found the same cluster pattern: among 54 young adults with Asperger syndrome, about half met criteria for a personality disorder, all in clusters A and C, more often in men (65%) than in women (32%), though that sample was small,¹³ and among 240 adults referred for a neuropsychiatric assessment, cluster A and C disorders were more common in those with autism and cluster B disorders in those with ADHD.¹²

Where the two diagnoses are compared directly, they differ more than expected. In one study, 51 autistic women and people assigned female at birth and 51 with a borderline-personality diagnosis, each group with few or subclinical traits of the other diagnosis, filled in questionnaires. Emotional reactivity, emotional dysregulation, identity disruption, difficulty being alone and rejection sensitivity were more characteristic of the borderline-personality group, with identity disruption the largest difference. Sensory processing, social cognition and behavior, a preference for sameness and repetitive movement were more characteristic of the autistic group, with sensory processing the largest difference. A combination of measures sorted 95.1% of participants correctly, though the model was built and tested on the same 102 people, all chosen to have few signs of the other condition, so it does not show how the measures perform on the ambiguous cases the two diagnoses usually present.³ An earlier study found that a small borderline-personality group scored above controls on autistic traits, and the gap between the borderline and autistic groups was marginal in the full sample and not statistically significant in a smaller subsample.⁹ A systematic review of the differential diagnosis concluded that the overlap, especially with borderline personality disorder, reduces diagnostic accuracy.⁴

In a Dutch register of 1,211 autistic adults, 24.6% said they had been given at least one psychiatric diagnosis before their autism diagnosis that they now considered a misdiagnosis. Personality disorders were the most frequent, ahead of anxiety disorders and mood disorders. Women reported this more often than men (31.7% against 16.7%), and specifically more personality-disorder, anxiety and mood misdiagnoses; women also reported more prior diagnoses in general, and among people who had any prior diagnosis, men and women were about equally likely to call it a misdiagnosis.⁵

An earlier survey from the same register found that half of 1,019 autistic adults (50.2%) had at least one psychiatric diagnosis before autism, that mood and personality disorders were the most frequent, and that 37.7% reported at least one prior diagnosis that was no longer present after the autism diagnosis. Personality disorders were the most frequently dropped.¹⁸

A survey of 420 adults who identified as autistic, possibly autistic or non-autistic found that the autistic and possibly autistic participants were more likely to report mental-health diagnoses but less likely to agree with them. The reasons they gave were that their autistic traits had been confused with a mental-health condition, and that their difficulties resulted from being autistic.⁶

Other studies show the same sequence. In a Swedish national cohort of 72,331 autistic people, 54.2% of the women and 40.9% of the men had at least one psychiatric diagnosis before autism.²⁰ In 161 adults given a first autism diagnosis at two Italian centers, the median age at diagnosis was 23 and the first mental-health appointment had come a median of 11 years earlier; personality disorder was among the most common earlier diagnoses, and women were diagnosed with autism significantly later than men.¹⁰ In a Dutch register of 2,722 mostly late-diagnosed adults, gender-diverse participants reported the most prior perceived misdiagnoses, followed by cisgender women and then cisgender men.¹⁵

Thirteen adults who had been diagnosed with borderline personality disorder before autism described in interviews how the earlier diagnosis had not explained their difficulties, how stigma and stereotyping had followed them, and how the autism diagnosis let them reframe their past.²⁴

A chart review of 1,027 psychiatric emergency-department visits by 345 autistic adults aged 16 to 63 at one Canadian hospital found that a borderline-personality diagnosis was recorded in 33.5% of visits and in 11.0% of the people, more often in people assigned female at birth and in people with a documented intellectual disability. An autism diagnosis was recorded in 60.2% of the visits, and a borderline-personality diagnosis was recorded less often when autism was documented. The authors read this as possible diagnostic overshadowing and state that it points only indirectly and cannot show which diagnosis came first.²

In the UK cohort, the relative difference for a personality-disorder diagnosis already recorded at the start was larger than for a new one, which its authors read as the diagnosis being made at a younger age in autistic people.¹ They attribute the pattern partly to autism being harder to identify in women and girls who mask, and to emotional dysregulation from trauma, social exclusion and the demands of masking being attributed to a personality disorder.¹

Undiagnosed autism is common in adult psychiatric settings. In one Swedish adult psychiatric outpatient clinic, screening every new referral estimated autism in at least 18.9% of the population, though that figure extrapolates from a screening questionnaire and a partly assessed subsample.¹⁹ A systematic review of autism among people with a personality disorder recommends a careful developmental assessment when a personality disorder is possible or confirmed.⁸ One case report describes a young man with severe self-injury who had been treated for borderline personality disorder for years and was later found, on reassessment by clinicians with autism expertise, to meet criteria for autism, ADHD and depression but not personality disorder; the authors report that the adapted treatment seemed to improve his self-injury and his relationships with family and caregivers.¹⁴

In the emergency-department chart review, suicidality as the main reason for the visit was associated with a recorded borderline-personality diagnosis, and admission, discharge and restraint use did not differ significantly between visits with and without one.²

The stigma attached to the diagnosis is documented in its own right. A review of the personality-disorder stigma literature found public knowledge low, people with personality disorders seen as deliberately misbehaving, and provider stigma "particularly pernicious for those with borderline personality disorder".²⁵ In-depth interviews with 22 US mental-health clinicians found patients with borderline personality disorder routinely labeled "difficult" and routed out of care.²⁶

The UK cohort's authors argue that personality-disorder stigma may increase the likelihood of inappropriate treatment, exclusion from services and not being considered deserving of care, and that diagnosing co-occurring conditions such as depression, psychosis or post-traumatic stress disorder, rather than attributing symptoms to personality disorder, may improve access to treatment.¹ A 2024 editorial in the British Journal of Psychiatry argues instead that personality difficulties are inherent to the autistic spectrum, that they are best assessed dimensionally rather than as categorical "personality disorders", and that therapies developed for personality disorder may help autistic people.¹⁶ A 2015 review in The Lancet Psychiatry described a "lost generation" of adults diagnosed with autism only in adulthood and said that possible misdiagnosis, especially in women, should be explored, with personality disorders among the overlapping conditions in the differential.¹⁷

Systematic reviews recommend a careful developmental history and a structured, standardized differential assessment when autism and a personality disorder may both be present, and advise considering autism when assessing someone with a possible or confirmed personality disorder.^8,4 An editorial in BJPsych Open reports a recommendation from a Swedish screening study that autism be "ruled out" in adult psychiatric patients by taking a developmental history, and points to NICE adult-autism guidance.²³ A Delphi study of 20 expert clinicians from 7 countries reached consensus that trauma and borderline personality disorder "could be difficult to differentiate from autism" in verbal women of typical intelligence, and that questionnaires must be paired with clinical judgment. The same study cautioned that autism criteria applied as a checklist can lead to over-diagnosis as well as under-diagnosis, and that people who are assessed but not diagnosed should still be offered support.²⁷ The case report above shows what the reassessment can change for one person.¹⁴

No study found as of October 2026 has tested whether ruling out autism before a personality-disorder diagnosis reduces misdiagnosis or diagnostic delay.

A clinician's guide by Donna Henderson, Sarah Wayland and Jamell White writes that many people diagnosed with personality disorders may in fact be undiagnosed autistic, and that personality-disorder-like traits can result from living with autism that is unsupported. When autism better explains the pattern, the guide says, a personality-disorder diagnosis "would not be appropriate (or helpful)".^c1 In the same guide, autistic adults describe the label arriving before anyone considered autism. Andi F. writes, "Almost all the traits that got me diagnosed with BPD are gone now that I know to avoid my sensory triggers."^c2

Devon Price writes that "It's quite common for adult Autistic women to be incorrectly labeled with Borderline Personality Disorder" and calls it "a really disastrous diagnosis", describing a stigma that makes affirming care harder to find.^c3 Amy Pearson and Kieran Rose use the term diagnostic overshadowing for the way autistic difficulties get attributed to another condition, note that borderline personality disorder is one of the most prevalent diagnoses autistic women receive before being identified as autistic, and write that the experience of misdiagnosis can be traumatizing and lead to medication and therapy that do not fit.^c4

Daniel Burford writes that personality disorders are "the single most common category of diagnosis autistic adults report having been handed before anyone said autism", with borderline the one that comes up most, and that the label changes both the language written in a person's notes and the treatment offered.^c5 In an edited collection of autistic adults' accounts, Annie writes that for 20 years professionals told her she had a mental illness, from bipolar disorder to dissociative identity disorder to borderline personality disorder, that she was given 13 kinds of medication and admitted to a psychiatric ward, and that she had to plead for a referral to an autism specialist.^c6