What autism is

You do not need to become a specialist. You need enough to understand what you are seeing, ask better questions, and put down blame that was never yours.

What autism actually is

The CDC's own words: autism spectrum disorder is a developmental disability caused by differences in the brain. CDC says it "begins before the age of 3 years and can last throughout a person's life, although symptoms may change over time."

Clinicians diagnose it on two areas, and only two. This is the DSM-5-TR, the manual they work from.

• Social communication and interaction. Lasting differences in back-and-forth conversation, in body language, and in building relationships. CDC's examples: avoiding eye contact, not responding to a name by 9 months, not pointing to show you something by 18 months.
• Restricted and repetitive patterns. At least two of: repeated movements or speech (lining things up, echoing phrases, hand flapping, rocking), needing sameness, intense focused interests, and sensory differences, meaning over- or under-reactivity to sensory input. Sensory is the part families notice first and hear about last. A hand dryer that makes a shop unbearable is not fussiness. It is in the diagnosis.

The signs must be present in early childhood, though the manual says they may be "masked by learned strategies in later life." That is why a person can be diagnosed at 40 and still have been autistic at 4. Autism is not an illness, and there is no cure to hunt for. A different brain is not a broken one.

How common it is

The CDC's ADDM Network published its latest count in April 2025, from data collected in 2022 at 16 US sites. About 1 in 31 children aged 8 (3.2%) was identified as autistic. Autism was 3.4 times as common in boys (49.2 per 1,000) as in girls (14.3 per 1,000).

The number keeps climbing. In 2000 it was 1 in 150. Here is what CDC itself says about that. Sites in the same country came out wildly different, from 1 in 103 in Laredo, Texas to 1 in 19 in California. CDC's explanation: "differences in ASD prevalence are likely due to differences in evaluation and testing practices." Research, it adds, "does not show that living in certain communities puts children at greater risk for developing ASD."

So the count largely tracks who gets looked at. The definition broadened when separate diagnoses were folded into one spectrum, recognition came earlier, and more children are evaluated at all. Your child is not part of an epidemic, but of a group that used to be missed.

The varying degrees of need

The DSM-5-TR asks clinicians to record how much support a person needs. Three levels, with their real names:

• Level 1, "requiring support." Without supports, the social differences "cause noticeable impairments." In life: a teenager who talks happily about trains, cannot start a conversation with a classmate, and falls apart in the car after a day of holding it together.
• Level 2, "requiring substantial support." Marked differences that stay obvious even with supports in place. In life: a child who speaks in short phrases, needs the same route home, and cannot recover from a changed plan without an hour of help.
• Level 3, "requiring very substantial support." Very little social initiation, and behaviours that "markedly interfere with functioning in all spheres." In life: an adult who communicates with a device or a few signs, and needs someone within reach all day.

Now set the level down. It is a snapshot of the assessment day, not a verdict, and it moves with the year, the school, the sleep and the noise. It is written for services and funding, so it tells you what to build, not what to expect.

It is also two numbers, not one. The manual says severity "should be separately rated" for social communication and for restricted, repetitive behaviours. The same person can be Level 1 in one and Level 3 in the other. If your report gives a single number, ask for both.

This site does not say "high-functioning" or "low-functioning". Those words hide the person twice: a Level 1 person's real struggles go invisible, and less gets expected of a Level 3 person. In the survey of the UK autism community described below, both terms were endorsed by only a minority of autistic adults.

The words "profound autism"

In 2021 the Lancet Commission on the future of care and clinical research in autism proposed the term profound autism. It is for people who are "minimally verbal or non-verbal, are not able to advocate for themselves, and require 24-hour access to an adult who can care for them." Its purpose was to make clinicians and researchers prioritise a group it called vulnerable and underserved.

CDC researchers then measured it across 15 ADDM sites, counting 8-year-olds who were nonverbal, minimally verbal, or had an IQ under 50. 26.7% of autistic 8-year-olds met that definition. In 2016 it affected 4.6 per 1,000 8-year-olds, and those children were more likely to have seizure disorders and self-injurious behaviour.

Autistic self-advocates object, and the objection is specific. The Autistic Self Advocacy Network says the definition "conflates needing 24/7 support with having a measured IQ of 50 or below or being nonspeaking," which it calls "a harmful oversimplification." Some people who speak fluently need round-the-clock support. Some device users live independently. ASAN adds that "IQ tests are notoriously unreliable for autistic people, particularly nonspeaking autistics."

Both are worth hearing. If the term gets your family a placement, use it. If a clinician uses it as a reason to stop teaching your child to read, do not accept it.

Across a lifetime

Autism spans a life, and the caring changes shape every few years.

What often comes with it

Autism rarely travels alone, and the companions are usually more treatable than the autism itself.

• Intellectual disability. 39.6% of autistic 8-year-olds with cognitive testing in CDC's 2022 data (IQ of 70 or below).
• Epilepsy. Children's Hospital of Philadelphia states 25% to 40% of autistic patients, against 2% to 3% of the general population.
• ADHD. Pooled prevalence of 28% in The Lancet Psychiatry meta-analysis.
• Anxiety. 20% in the same analysis, with obsessive-compulsive disorder at 10%.
• Gut problems. Estimates vary widely by study, from 9% to 70%, with chronic constipation at a median of 22%.
• Sleep. Chronic sleep problems in 50% to 80% of autistic children, against 9% to 50% of non-autistic children.
• Feeding differences. Around 27% have problems with food selectivity, often linked to the gut problems above.

Why the list matters. Every condition on it hurts, and an autistic person in pain may not be able to tell you where. Constipation, reflux, an ear infection, a bad tooth and a seizure can all arrive as screaming, hitting, refusing food or not sleeping.

Behaviour is communication

The site runs this principle across every condition it covers. Behaviour is a message from someone who cannot deliver it another way. The job is to read it, not to stop it.

A meltdown is not a tantrum. The National Autistic Society defines it as "an intense response to an overwhelming situation" in which someone "temporarily loses control of their behaviour", and says directly: "A meltdown is not the same as a temper tantrum. It is not bad or naughty behaviour." A tantrum has an audience and a goal. A meltdown has neither, and it does not end when the person is given what they wanted, because there was nothing they wanted.

A shutdown is not defiance. The same overload can go inward: going still, going silent, refusing to move, appearing to ignore you. Pushing harder makes it longer. NAS describes an earlier "rumble stage" of pacing, repeated questions, rocking or becoming very still. That stage is your window.

Check these five before you correct anything:

1. Pain or illness. Teeth, ears, gut, head, seizures.
2. Sensory load. Noise, lights, smell, crowding, a label in a shirt, a hand dryer two rooms away.
3. Change. A different route, a substitute teacher, a plan that shifted without warning.
4. Fear. Something that happened here before, or something coming that nobody explained.
5. An unmet need with no words attached. Hunger, thirst, the toilet, needing to leave.

NAS's first advice for the moment itself is to give time and make space. More on Help now.

The words, and why they matter

This site says "autistic person" rather than "person with autism", because autistic adults asked for it. Kenny and colleagues' 2016 study in the journal Autism surveyed the UK autism community: more autistic adults endorsed "autistic" and "autistic person" than "person with autism". The same study found no single term everyone accepted, and said so plainly.

• "Autistic" is not an insult. It is a description, the way "left-handed" is. Lowering your voice to say it teaches a child that it is shameful.
• "Nonspeaking" beats "nonverbal". Many people who do not speak do use language, by typing, writing, signing or a device. "Nonverbal" gets heard as "understands nothing", and that leads to being talked over.
• Never "suffers from autism". A person may suffer from pain, from anxiety, or from a world built for someone else. That is a different sentence.
• Say "non-autistic", not "normal". The opposite of a normal child is a wounded one.

Loving families use both kinds of language, and some autistic adults prefer person-first for their own reasons. This page will not police anyone. If the person can tell you what they want to be called, that outranks every survey.

What causes it, and what does not

The weight of the evidence sits on genetics. NICHD states that "a great deal of evidence supports the idea that genes are one of the main causes of or a major contributor to ASD," and that "more than 100 genes on different chromosomes may be involved." There is no single autism gene, which is why there is no simple test.

CDC says "scientists believe there are multiple causes of ASD that act together", and lists what makes it more likely: an autistic sibling, genetic conditions such as fragile X syndrome or tuberous sclerosis, complications at birth, older parents. Other environmental factors are under study, and honest sources say study rather than conclusion.

Vaccines do not cause autism. The American Academy of Pediatrics puts it in one sentence: "Decades of rigorous research have shown vaccines do not cause autism." On the 1998 study that started the claim, AAP says "the results of the study were later proven false". On the mercury preservative: "no causal association was found between ASD and thimerosal (mercury)."

It is nobody's fault. Not how you held your baby, not screen time, not working, not divorce, not diet, not anything you have replayed at 3am. Cold parenting was blamed for decades, and that theory was wrong. Autism starts before birth, in the wiring.

The hard choices, both sides

Loving, faithful people land on opposite sides of these. Here is each one fairly, so you decide rather than get sold.

Intensive behavioural therapy (ABA)

The parents' case. It is the most available therapy in the US, the one insurers cover, often the only funded option in a county. For a child who bites through skin or runs into traffic, a programme that reduces that is not a philosophy question. Cochrane's 2018 review found children in early intensive behavioural intervention did better than comparison groups on adaptive behaviour, IQ and language after about two years.

The autistic adults' case. Many autistic adults describe these programmes as training in compliance: suppressing the stimming that calms them, accepting touch they hate, learning that no is not accepted. ASAN holds that autistic people "deserve access to high-quality, evidence-based, ethical services that work towards our goals, as defined by autistic people ourselves." The evidence is also thinner than the marketing. Cochrane found "no evidence at post-treatment that EIBI improves autism symptom severity", called its findings "weak evidence", and drew them from five studies of 219 children, only one randomised.

The questions that tell a respectful programme from a harmful one. Ask them on the tour, and watch the faces.

• Show me a goal you set last year. Whose goal was it? Goals should be things the person would want: asking for help, making a friend, staying safe. Not quiet hands.
• What happens when a child says no, or walks away? A no should be respected. Anyone who calls it "escape behaviour" to be extinguished has told you what you needed to know.
• Do you use any punishment, response cost, or planned ignoring? There is no acceptable amount of punishment. Ask twice.
• How much of a session is play, and how much is drills at a table? Play-based beats compliance drills, and 40 hours a week of anything is a lot for a small child.
• Can I watch a whole session, unannounced, any time? A yes without hesitation is the best single sign you will get.

Declining a programme is a legitimate choice. So is taking one and setting your own limits on hours and goals. Both are done in love.

Medication

No medicine treats autism. Medicines do treat what comes with it: seizures, ADHD, anxiety, depression, severe sleep problems. This site names no drugs and no doses. Some families find that treating anxiety hands their child back their week. Others find side effects that cost more than the symptom did. What you can ask: What is this treating, how will we know in six weeks whether it worked, and what would make us stop?

Telling the child their diagnosis

Some parents wait, out of a real fear of handing a child a label to carry. The case for telling early is simple. They already know they are different, and without a word for it they usually decide the difference is a fault in them.

The National Autistic Society advises starting with difference itself, making clear that difference is not bad or scary, introducing it slowly across many conversations, and explaining that they do not have an illness. Told early and plainly, it is a fact about themselves, like being tall. Told at 15, after years of quiet appointments, it lands as a secret that was kept.

Verified September 2026 against the CDC's autism pages and the ADDM Network's 2025 report, the DSM-5-TR criteria, NICHD, the American Academy of Pediatrics, the Lancet Commission, the Autistic Self Advocacy Network, the National Autistic Society, and Cochrane. Sources on the sources page. Background for a caregiver, not medical, legal or financial advice.