Autism in children: signs, diagnosis, and care that supports development.

Autism spectrum disorder (ASD) is a neurodevelopmental difference characterised by persistent challenges in social communication and the presence of restricted, repetitive patterns of behaviour, interests, or activities. About 1 in 100 children in India are on the autism spectrum (US community-based estimates run higher at ~1 in 36, primarily reflecting better identification rather than true incidence), with the earliest reliable signs visible from 12–18 months. The American Academy of Pediatrics recommends universal autism-specific screening of every child at the 18-month and 24-month well-child visits using the M-CHAT-R/F. Both core features must be present from early childhood and must cause meaningful functional difficulty for the diagnosis to apply; diagnosis is clinical, not based on a blood test or brain scan. Early structured intervention — particularly between ages 1 and 5, when neuroplasticity is highest — is the most reliable lever for long-term outcomes.

Severity at diagnosis: what the three levels mean

The DSM-5-TR specifies three severity levels at the time of autism diagnosis, each describing how much support the child currently needs. Levels are not permanent — many autistic children move down a level with intervention; some need more support over time as social and academic expectations shift.

Level	Social-communication needs	Restricted / repetitive patterns	Support implication
Level 1 — Requiring support	Difficulty initiating interaction; responses to others may be atypical or unsuccessful	Inflexibility causes interference in one or more contexts	Visible difficulty without support; everyday function largely preserved
Level 2 — Requiring substantial support	Marked deficits in verbal and non-verbal social communication, even with supports in place	Frequent restricted or repetitive behaviours obvious to a casual observer	Difficulty coping with change of focus or action
Level 3 — Requiring very substantial support	Severe deficits; very limited initiation, minimal response to overtures from others	Extreme difficulty with change; repetitive behaviours markedly interfere with function	High level of structure and support across daily activities

Source: DSM-5-TR, American Psychiatric Association.

What “spectrum” actually means

The shift from older sub-categories (autistic disorder, Asperger’s syndrome, PDD-NOS) to a single “spectrum” was a deliberate clinical decision in the 2013 DSM-5. It reflected the strong research evidence that drawing sharp lines between sub-types was not clinically useful — the underlying mechanisms appear continuous, not categorical. In practice, this means that two children who both have autism may look quite different from each other. One may be highly verbal and academically advanced, with social difficulties most visible in nuanced peer relationships. Another may have minimal verbal communication and need substantial support across daily activities. Both meet the diagnostic criteria; both belong on the same spectrum.

How common is autism

Global prevalence estimates have risen substantially over the past two decades, primarily reflecting better identification rather than a true increase in incidence. The CDC’s ADDM Network places US prevalence at about 1 in 36 children. Indian community-based studies have produced lower estimates — typically in the range of 1 in 100 to 1 in 125 — though these likely under-identify autistic children, particularly girls, children with co-occurring intellectual disability, and children in non-metro settings where screening reach is limited.

What causes autism

The current scientific consensus is that autism is predominantly genetic, with heritability estimates from twin studies typically falling between 60 and 90 percent. Hundreds of genes have been linked to autism risk; in most cases, no single genetic cause is identifiable. A small number of environmental contributors — notably advanced parental age and certain prenatal exposures — appear to add modest risk.

Vaccines do not cause autism. The 1998 paper proposing this link was fraudulent, was retracted in 2010, and has been disproven by every replication attempt at scale.

Early screen exposure: what the recent Indian evidence actually says

A 2026 AIIMS Delhi study led by Prof. Shefali Gulati (Head, Pediatric Neurology), published as an abstract in the Journal of the International Child Neurology Association, examined 250 children aged 3–18 across socio-economic backgrounds. The researchers found that more than 80% of the autistic children in the cohort had been exposed to screens for over 20 minutes per session before 18 months of age, compared with roughly 66% of the non-autistic comparison group. The authors are emphatic about what the finding does not say: screens do not cause autism. What the study suggests is that intensive screen exposure during the critical 0–18 month window may interfere with the social-communication experiences a developing brain needs — the joint attention, face-to-face turn-taking, and gestural exchange that lay the foundation for language. For families with a child already showing developmental concerns, the practical implication is that limiting unsupervised early screen time and protecting face-to-face interaction time is a defensible, low-cost intervention worth starting now, regardless of whether a formal diagnosis is in progress.

How autism is diagnosed

Diagnosis is clinical — not based on a blood test, brain scan, or genetic test. A developmental paediatrician or child psychiatrist combines (1) developmental history from parents and caregivers, (2) direct observation of the child using standardised tools (most commonly ADOS-2), and (3) review of the child’s functional level across communication, social interaction, and behavioural domains. Where appropriate, formal cognitive testing and adaptive-functioning assessment (Vineland-3) are added.

The AAP recommends universal autism-specific screening of all children at the 18 and 24-month well-child visit, using the M-CHAT-R/F. This catches a meaningful proportion of autistic children at an age when intervention is most plastic. Children who screen positive are referred for diagnostic assessment.

What good support looks like

The strongest evidence supports comprehensive, multi-modal early intervention — typically combining speech-and-language therapy, naturalistic behaviour-analytic intervention (ABA done well), occupational therapy where sensory-motor needs are present, and structured parent coaching. The intensity, mix, and duration of intervention should be calibrated to the child’s profile and recalibrated regularly as the child develops.

Support that is not well-supported — and in some cases is actively harmful — includes nutritional or biomedical “cures”, chelation, and intervention models built on the goal of making a child appear non-autistic at the cost of suppressing the child’s own communication or self-regulation strategies. These are widely available in India and elsewhere; we recommend caution.