The core insight is structural rather than about individual bad actors. When a patient reports harm, the same institution that may be implicated in that harm defines what counts as legitimate knowledge, classifies the person making the report, decides what is recorded in the notes and determines whether reform is needed. Philosophers call the resulting wrongs epistemic injustice: harms done to people specifically in their capacity as knowers. In mental healthcare, stereotypes of irrationality, unreliability or manipulation can diminish a patient’s credibility while professional interpretations receive presumptive authority. The authors stress that psychiatry is not monolithic and has genuine traditions of self-critique, supervision and professional disagreement, but they insist that internal reflection alone is not learning. Learning becomes visible only when concerns are detected, reforms implemented, outcomes measured and recurrence reduced.

A central technical contribution of the Essay is a careful dissection of what different kinds of evidence can and cannot show. Qualitative research can rigorously investigate experiences, meanings and processes, but cannot estimate how common a phenomenon is. Epidemiological studies can estimate frequency with appropriate sampling. Randomised trials can estimate causal treatment effects under defined conditions, but applying those results elsewhere requires attention to participants, interventions, outcomes and settings. The authors warn against collapsing these distinct tasks into a single number. An adverse event occurs during treatment without necessarily being caused by it; iatrogenic harm requires an attribution to care. Negative experiences, reliable deterioration and serious adverse events differ in severity and how they are ascertained, and combining them into one prevalence figure would obscure clinically important differences.

The evidence on harm in psychological treatment is striking precisely because harm is so rarely sought. In one review of 132 randomised trials of psychological interventions, 79 per cent gave no indication that harms were monitored at all, and only 3 per cent described both adverse events and the methods used to collect them. Later reviews identify inconsistent definitions and terminology across the field. Reported estimates of clients experiencing at least one negative episode in psychotherapy range from 7 to 96 per cent across studies, with reliable deterioration estimated at 2 to 14 per cent, figures that vary depending on what is counted and how. Limited safety evidence, the authors argue, ultimately constrains clinical guidelines and the informed consent patients can genuinely give. Tools such as a patient-rated scale for positive and negative experiences of psychotherapy offer a starting point for systematic monitoring.

The Essay also challenges the assumption that professional knowledge is a neutral comparator against which patient accounts should be judged. In an experimental diagnostic task, psychiatrists who searched for confirmatory information were more likely to retain an incorrect preliminary diagnosis. A meta-meta-analysis found an association between researcher allegiance and psychotherapy outcomes, meaning the preferred therapy of a research team tends to perform better in that team’s trials. The authors are careful to note that such findings do not establish deliberate distortion, but they support what they call reciprocal reflexivity: lived experience and professional expertise should each be scrutinised, matched to the question at hand and neither presumed infallible nor excluded from evaluating the other’s claims.

Representativeness problems compound the picture. Among articles in a leading American psychiatry journal, only 43 per cent reported the race or ethnicity of participants, and just a quarter acknowledged limited representation as a limitation. In perinatal mental health studies conducted during the COVID-19 pandemic, 76.5 per cent of participants were White; in an eating disorder review, roughly 70 per cent of participants in studies reporting ethnicity were White. A sample containing few or no members of a population cannot provide precise safety estimates for that population, and extrapolation requires explicit justification. The published record itself may misrepresent the research conducted: comparing 74 antidepressant trials with regulatory reviews, 94 per cent appeared positive in the published literature versus 51 per cent in the assessments, and the published effect size was 32 per cent larger overall. Among federally funded psychological treatment trials for depression, nearly a quarter went unpublished, and including available unpublished results reduced the estimated effect by 25 per cent.

From these threads the authors build a vocabulary of institutional dysfunction. Ignorance culture describes the handling of alarms through silence, minimisation, deflection or blame. Exclusion culture describes practices that make denying or withdrawing care appear clinically appropriate. Defensiveness names the redirection of scrutiny away from a practice and toward the credibility, pathology or emotionality of the person raising the concern. Most provocatively, diagnostic weaponisation denotes the unwarranted extension of diagnostic authority to discredit, exclude, coerce or control, using a label beyond its evidentiary scope or as a substitute for an independently justified decision. Empirical studies lend this concern weight: higher rates of borderline personality disorder diagnosis have been found among lesbian, gay and bisexual patients, and among transgender and gender diverse patients, even after adjustment for relevant clinical correlates, while an experiment showed that a suggested borderline personality disorder label adversely influenced clinicians’ judgements of an unrelated panic disorder beyond equivalent behavioural information.

The consequences of classification ripple outward. A clinician’s account enters the medical record and can shape diagnosis, risk assessment, coercion and access to care; a researcher’s choices determine which outcomes are foregrounded; institutions decide whether complaints become part of safety systems. Accounts from people harmed in care describe withdrawal from services after coercion or invalidation, which removes their perspectives from clinical learning precisely because feedback depends on continued engagement. Qualitative studies of anorexia nervosa care describe treatment experienced as simultaneously life-saving and unsupportive, coercive or enduringly harmful, and interviews with people reporting long-term adverse effects after electroconvulsive therapy describe memory changes, losses of identity and dismissive professional responses. Antidepressant withdrawal tells a similar story: one survey found withdrawal symptoms in 79 per cent of respondents who had attempted discontinuation, though the low response rate limits population inference, and patient accounts document dismissal, misdiagnosis and inadequate tapering support. Uncertainty, the authors argue, calls for better ascertainment, not dismissal of reported experiences.

The proposed remedy is a responsibility culture built on four linked standards. Serious or recurrent concerns require scrutiny sufficiently independent to challenge the institution, including epistemic independence from the shared assumptions and allegiances under examination. Safety surveillance must actively solicit clearly defined harms rather than passively waiting for complaints, since fewer reports may reflect diminished trust rather than safer care. Relevant lived experience expertise needs remunerated decision-making authority over harm definitions, outcomes, priorities and evaluation, not tokenistic participation. And institutions need mechanisms for remedy, redress and public reporting of implementation, because policies, training and commitments are inputs, not evidence that safety has improved. Research transparency, including trial registration, accessible protocols and complete outcome reporting regardless of results, forms part of the surveillance apparatus.

The Essay’s conclusion is a challenge to the field’s self-image. Psychiatric expertise has genuine value, but professional status does not confer epistemic neutrality, and the authority that accompanies it makes reciprocal scrutiny especially important. Trust, the authors argue, should rest on transparent self-correction rather than assurances of good intent. The practical test they set is disarmingly simple: can psychiatric institutions recognise when their own practices are implicated in harm, respond to those affected, and demonstrate publicly what changed and whether safety improved? Until learning is measured in those terms, warnings that recur across misconduct, coercion, diagnosis and iatrogenic harm will continue to be attenuated by the very systems that should be heeding them.

Subject of Research: Institutional non-learning, epistemic injustice and diagnostic weaponisation in psychiatry

Article Title: When psychiatry does not learn: Epistemic humility, ignorance culture, exclusion culture and diagnostic weaponisation

Article References: Cobbaert, L., Elwyn, R., Downs, J., James, S., & Jackman, M. (2026). When psychiatry does not learn: Epistemic humility, ignorance culture, exclusion culture and diagnostic weaponisation. PLOS Mental Health, 3(10), e0000741. https://doi.org/10.1371/journal.pmen.0000741

Image Credits: AI Generated

DOI: 10.1371/journal.pmen.0000741

Keywords: psychiatry, epistemic injustice, institutional non-learning, iatrogenic harm, diagnostic weaponisation, psychotherapy safety, publication bias, lived experience, coercive care, diagnostic bias, epistemic humility, accountability

The research team, led by Jeremy J. Cottle of Nationwide Children’s Hospital together with colleagues at the University of Alabama at Birmingham, analyzed records from 458 children and adolescents who had been referred for an autism evaluation at a tertiary care clinic between 2013 and 2020. Crucially, the sample included every child referred for assessment, not only those who ultimately received a diagnosis. Of the participants, 118, or 25.8 percent, were female, and 61.6 percent were diagnosed with autism spectrum disorder. The average age was just under seven years. By including children who did not receive an autism diagnosis, the researchers deliberately avoided a common pitfall: studies restricted to diagnosed individuals can systematically exclude the very girls whose presentations current tools fail to capture.

The study focused on the two instruments widely regarded as the gold standard for autism assessment. The Autism Diagnostic Observation Schedule, Second Edition, or ADOS-2, is a semi-structured, play-based observation administered directly to the child by a trained clinician. The Autism Diagnostic Interview, Revised, or ADI-R, is a structured interview conducted with caregivers, drawing on their long-term knowledge of the child’s behavior across settings and years. Both produce scores across domains of social communication and restricted, repetitive behaviors, but they capture autism traits through very different lenses: one through a clinician’s real-time observation, the other through a caregiver’s retrospective report.

To detect bias at the level of individual test items, the team used a statistical technique called multiple indicator multiple cause modeling, or MIMIC. The approach works by first estimating latent variables, in this case underlying traits such as social affect and repetitive behaviors, from the pattern of item scores. It then asks a precise question: does a child’s sex predict how they score on a particular item even after holding their overall level of the underlying trait constant? If so, that item exhibits differential item functioning, meaning it behaves differently for boys and girls beyond what overall trait levels can explain. The researchers converted these effects into odds ratios, flagging differences as meaningful when the ratio exceeded 2.0 or fell below 0.5, and used delta transformations to compare the magnitude of bias across items, domains, and instruments.

The results split cleanly along the two measures. On the ADOS-2, girls did show lower overall levels of restricted and repetitive behaviors than boys, consistent with a large body of prior research. Yet when the researchers probed individual items, they found no differential item functioning at all. No single ADOS-2 item, whether measuring social affect or repetitive behavior, functioned differently for boys and girls once overall trait levels were accounted for. In other words, the observed domain-level sex difference on the ADOS-2 appears to reflect genuine differences in the traits children display during observation, not biased scoring of the instrument itself.

The ADI-R told a different story. Four items showed differential item functioning, spread across all three of the diagnostic content areas the researchers examined. In the social interaction domain, girls were scored higher on inappropriate facial expressions even at the same overall level of social interaction difficulties, an odds ratio of exactly 2.0. In the communication domain, girls were scored lower on spontaneous imitation of actions and imaginative play, with odds ratios of 0.225 and 0.256 respectively, indicating substantially higher endorsement of autistic traits for boys on these items. These two communication items also carried the largest delta values in the study, marking them as the sites of the greatest sex bias. In the repetitive behavior domain, girls were scored lower on circumscribed interests, though this effect narrowly missed the formal threshold for meaningfulness.

Each of these biased items can be interpreted through the leading theories of why autism is missed in girls. The circumscribed interests finding suggests that girls’ intense interests may take forms, such as art, animals, or celebrities, that clinicians do not immediately recognize as unusual, whereas interests in trains or schedules are more readily flagged. Intense interests in dolls or babies may even be mistaken for typical pretend play unless their intensity and peculiarity are explicitly probed. The imaginative play result aligns with the idea that play centered on dolls is more easily labeled as imaginative by caregivers than play with cars or trucks. The imitation finding fits camouflage theory: girls may more often observe and copy peers’ behavior, skills that caregivers notice and that can mask underlying difficulties. The inappropriate facial expressions result, meanwhile, may reflect the perception hypothesis, with caregivers judging girls’ expressions against stricter social expectations.

The contrast between the two instruments carries practical weight. On the ADI-R, males were significantly more likely to receive an autism classification than females, a pattern that did not appear on the ADOS-2. The authors caution that the ADI-R contains more diagnostic items than the subset of ADOS-2 items analyzed, which raises the statistical chance of detecting bias, and that the cross-module comparison excluded some ADOS-2 items that could not be validly compared across versions. Still, the finding that boys were more likely to clear the ADI-R’s diagnostic threshold, combined with item-level bias on that same instrument, suggests that caregiver-report measures may be a specific point where female presentations slip through. The authors note that if the ADI-R is used to determine eligibility for research studies, girls with characteristically female presentations may be disproportionately excluded, reinforcing the male skew of autism research samples.

Notably, the study found no overall sex difference in the likelihood of receiving a clinical autism diagnosis, a result that diverges from earlier research but may reflect growing clinician awareness of female presentations in recent years. The authors also acknowledge limitations: the retrospective clinical data could not examine gender as distinct from sex assigned at birth, could not account for cognitive ability, and spanned a wide age range from two to eighteen years. Some girls who are never referred for evaluation in the first place remain invisible to any clinic-based study.

The implications reach beyond statistics. The authors argue that the findings point toward concrete adaptations, such as guidance for ADI-R interviewers to use gender-balanced prompts and examples when probing interests, play, and imitation, and greater education about the female autism phenotype for everyone involved in referral and diagnosis, from pediatricians and educators to psychologists and caregivers. Even when boys and girls show similar overall levels of autistic traits, the specific ways those traits appear can differ, and diagnostic conversations need to be structured to notice that. As the authors conclude, a well-informed understanding of sex differences in how autism traits are presented and measured is critical for correcting the sex bias that persists in clinics and research alike, and for ensuring that children of both sexes receive accurate identification and the support that follows from it.

Subject of Research: Item-level sex differences in autism diagnostic measurement using ADOS-2 and ADI-R in a referred pediatric sample

Article Title: Item-Level Sex Differences in Autism Spectrum Disorder Measures: A MIMIC Model Analysis of ADOS-2 and ADI-R in a Referred Pediatric Sample

Article References: Cottle, J. J., Arnold, Z. E., Guest, K. C., Mrug, S., Brisendine, A. E., Khatri, S., & O’Kelley, S. E. (2026). Item-Level Sex Differences in Autism Spectrum Disorder Measures: A MIMIC Model Analysis of ADOS-2 and ADI-R in a Referred Pediatric Sample. Journal of Autism and Developmental Disorders. https://doi.org/10.1007/s10803-026-07527-6

Image Credits: AI Generated

DOI: 10.1007/s10803-026-07527-6

Keywords: autism spectrum disorder, sex differences, ADOS-2, ADI-R, differential item functioning, MIMIC model, female autism phenotype, diagnostic bias, restricted repetitive behaviors, camouflaging, pediatric assessment, underdiagnosis