Skip to content

Hero image: A diverse group of women sharing a relaxed conversation in a sensory-considerate community space

All insights

Mental & Emotional Wellbeing

Autism in Girls and Women: Seeing the Needs Behind Camouflage

9 min readEvidence synthesis
Read the evidence

The question in focus

An evidence-based guide to autism recognition in girls and women, diagnostic assessment, camouflaging, support and the limits of current tools.

Evidence at a glance

1 in 31

US eight-year-olds identified with autism in 2022 surveillance

CDC surveillance provides population context, not an individual diagnostic test. Boys were identified more often than girls, and identification varies across communities.

Prevalence and Early Identification of Autism Spectrum Disorder, 2022

Why familiar stereotypes miss people

Autism is a lifelong neurodevelopmental difference involving social communication and interaction alongside restricted or repetitive patterns, interests, routines or sensory experiences. The way these features appear is diverse. Systematic reviews describe barriers for girls and women when clinicians, families or schools expect a narrow, male-typical presentation [1]. Strong vocabulary, apparent eye contact, friendships or imaginative interests do not by themselves rule out autism.

Some girls and women consciously or unconsciously camouflage by rehearsing conversations, copying peers, forcing eye contact, suppressing movement or analysing social rules. Camouflage may reduce visible signs during a brief appointment while increasing fatigue, anxiety and loss of a stable sense of self. It is not unique to women and not every autistic person camouflages. Evidence is also shaped by samples that have often been White, educated and recruited from specialist communities [1][2].

  • Look for the effort and recovery time behind social performance.
  • Ask about sensory needs, routines and focused interests using age-relevant examples.
  • Do not infer support needs from eye contact or verbal fluency alone.

Assessment should reconstruct development, not reward performance

A comprehensive assessment integrates current experience, early development, direct observation and information from someone who knew the person earlier when available. It should evaluate education, work, relationships, daily living, sensory processing, mental health and other neurodevelopmental conditions. NICE recommends a structured comprehensive assessment and warns against using a screening score as the sole basis for diagnosis [3].

Tools such as the AQ-10, ADOS or developmental interviews can contribute evidence, but every instrument has a validation context and none is a biological confirmation. An apparently typical performance in a structured session may not represent unstructured daily life. Clinicians should ask what preparation preceded the appointment and what recovery follows it. They should also consider language, culture, intellectual disability, trauma, ADHD, anxiety, eating disorders and medical causes of behavioural change [3][4].

  • Use multiple sources while respecting that family informants may not have seen internal distress.
  • Record strengths and preferences as well as impairments.
  • Explain uncertainty instead of forcing a binary answer from limited evidence.

Diagnosis is a doorway, not the intervention

A diagnosis can provide language, access and relief, but support should respond to specific needs. Useful adjustments can include predictable agendas, written information, reduced sensory load, direct language, time to process questions and permission to communicate in writing. NICE recommends adapting psychological treatments for coexisting mental health conditions with greater structure, visual or written information, explicit rules and regular breaks [3].

There is no medicine that removes autism, and attempts to make a person appear non-autistic should not be the goal of care. Medicines may be used for particular coexisting conditions after ordinary benefit-risk review. Support can address sleep, anxiety, depression, pain, gastrointestinal symptoms, executive functioning, communication, employment and relationships. The person should decide which outcomes matter, rather than having compliance or eye contact treated as success.

  • Ask which environments create overload and which adjustments reduce it.
  • Treat pain and mental health symptoms rather than attributing everything to autism.
  • Include partners or family only with consent and clear purpose.

Women's health services need sensory and communication competence

Menstrual care, contraception, fertility treatment, pregnancy, pelvic examination and menopause can combine sensory, communication and uncertainty demands. NICE advises adapting the setting and communication to an autistic adult's needs [3]. In women's health, that means explaining each step before touch, offering choices about positioning and support people, avoiding figurative language and allowing extra processing time. It also means asking about the patient's own pain signals rather than assuming conventional facial expression or tone.

Digital portals can help people prepare questions and receive written summaries, but inaccessible forms and rapid messaging can create new barriers. A profile should record communication and sensory preferences with permission and carry them across settings. It should not become a stereotype. Preferences can change by procedure, stress level and life stage, so clinicians need to confirm them rather than rely solely on a permanent flag.

  • Offer a written sequence of what will happen and why.
  • Ask before touch and make stopping signals explicit.
  • Provide the plan and safety-net instructions in the patient's preferred format.

Better data should not become automated diagnosis

CDC estimated that about 1 in 31 eight-year-old children in its 2022 surveillance network had been identified with autism [5]. This is an estimate from selected US communities and cannot be applied as a universal global rate. Identification was more common in boys, but surveillance cannot determine how much of the difference reflects prevalence, recognition, referral or diagnostic practice.

Researchers are studying digital phenotyping, voice, movement and machine-learning models. These tools remain investigational for broad diagnostic use. Performance may degrade across sex, race, language, intellectual ability or clinic context, and a model can reproduce the bias in its training data. Technology is most defensible when it improves access, communication or follow-up while leaving diagnosis to a transparent, accountable clinical process [1][4].

  • Require subgroup validation before deploying an assessment algorithm.
  • Do not use digital behaviour as covert surveillance.
  • Give people a route to correct records and challenge automated outputs.

What the evidence cannot yet answer

  • Female autism research remains affected by referral bias and historically male-weighted diagnostic samples.
  • Camouflaging evidence is heterogeneous and often drawn from White, educated or self-selected populations.
  • Surveillance identification rates cannot separate underlying prevalence from access and diagnostic practice.

Questions worth taking into care

  1. What social or sensory demands require preparation, scripting or recovery?
  2. Which early developmental information is available, and what may have been overlooked?
  3. How will the assessment distinguish autism from and alongside ADHD, anxiety, trauma or other conditions?
  4. Which communication and sensory adjustments should be recorded for healthcare visits?
  5. What support is available whether or not a final diagnosis is reached?

Source record

Evidence used in this review

Sources were selected for clinical authority, methodological relevance and traceability. Links open the original guidance, public-health record or research publication.

  1. [1]
  2. [2]
  3. [3]
    Autism Spectrum Disorder in Adults: Diagnosis and Management

    National Institute for Health and Care Excellence · 2021

  4. [4]
  5. [5]
Editorial standard

This evidence synthesis is for general information. It does not diagnose a condition or replace care from a qualified health professional. Treatment choices depend on individual history, examination, local guidance and informed preference. Emergency or rapidly worsening symptoms need urgent local medical assessment.