Autism in young children: what the evidence says on early signs and screening

Autism can be reliably identified in some children as young as 18 months, and US paediatricians recommend standardised screening at both the 18- and 24-month checks alongside ongoing developmental surveillance [s1]. A separate US expert body concluded that the evidence to recommend screening every child was insufficient — a call to gather more evidence, not a verdict that screening is unhelpful [s2]. On the question parents still ask most, very large studies are unambiguous: vaccines do not cause autism [s3][s4].

What autism is, and its early signs

Autism spectrum disorder is a common neurodevelopmental condition defined by differences in two domains: social communication and interaction, and restricted or repetitive patterns of behaviour and interests [s1]. In the American Academy of Pediatrics clinical report, US prevalence was reported as approximately 1 in 59 children, about 1.7% [s1]. It is a spectrum: support needs range widely across behaviour, education, health and family life [s1].

Early signs that prompt a closer look include limited eye contact and shared attention, delayed or unusual language, not responding to a name, few gestures such as pointing, intense focus on particular objects or topics, and strong reactions to changes in routine. No single sign is diagnostic, which is exactly why structured tools exist.

Why screening happens at 18 and 24 months

The AAP report reaffirms standardised screening for autism at 18 and 24 months of age within routine primary care, on the reasoning that autism is common, can be diagnosed as young as 18 months, and has evidence-based interventions that may improve function [s1]. Screening is not diagnosis: a screen that raises concern leads to a fuller assessment, not a label. The report also acknowledges that more accurate and culturally sensitive screening approaches are still needed [s1].

That is where a genuine scientific disagreement sits. The US Preventive Services Task Force reviewed screening tools administered in routine visits to children aged 18 to 30 months for whom no concerns had been raised, and issued an "I statement" — concluding that current evidence is insufficient to weigh the benefits and harms of screening every such child [s2]. This is frequently misreported as advice against screening. It is not: an I statement means the trials to settle the question have not been done, and it explicitly does not apply to children whose parents or clinicians already have concerns, who should be evaluated regardless [s2].

The vaccine question, settled by scale

Because a discredited 1998 paper linked the MMR vaccine to autism, the claim continues to circulate, and the evidence against it is now overwhelming. A meta-analysis pooled five cohort studies covering 1,256,407 children and five case-control studies covering 9,920 children [s3]. It found no relationship between vaccination and autism (odds ratio 0.99, 95% confidence interval 0.92 to 1.06), no link specifically for MMR (OR 0.84, 95% CI 0.70 to 1.01), and none for the vaccine preservative thimerosal (OR 1.00, 95% CI 0.77 to 1.31) [s3].

A later nationwide cohort added decisive statistical power. It followed 657,461 children born in Denmark from 1999 through 2010; during follow-up 6,517 were diagnosed with autism, and comparing vaccinated with unvaccinated children gave a fully adjusted hazard ratio of 0.93 (95% CI 0.85 to 1.02) [s4]. The study found no increased risk even in subgroups thought to be susceptible, and no clustering of diagnoses after vaccination [s4]. The rise in recorded autism prevalence reflects broader diagnostic criteria and better recognition, not an environmental trigger delivered by a needle. Our coverage of the childhood vaccine schedule sets out how that schedule is set and reviewed.

Attention has since shifted to other prenatal exposures, including paracetamol; we assess that claim against the evidence in our explainer on paracetamol use in pregnancy. Our coverage of the overlap between ADHD and autism diagnosis and of the male-to-female autism ratio covers who gets diagnosed, and when.

How to read this without overreaching

The evidence supports a calm, specific message. Learn the early signs; take part in the 18- and 24-month screens where they are offered; and if you have a concern at any age, ask for an assessment rather than waiting for a scheduled screen [s1][s2]. The debate captured by the task force's I statement is about whether to screen children with no signs at all, not about whether autism is real or whether early support helps [s2].

The limits are worth stating. Prevalence figures shift as criteria and ascertainment change, and the 1 in 59 estimate reflects a particular US surveillance period [s1]. Screening tools miss some children and over-refer others, which is why a positive screen must be followed by proper evaluation [s1][s2].

This article is informational and is not medical advice. It does not diagnose any child or recommend a specific screening test, and any concern about a child's development should be discussed with a qualified clinician.