Established care

Early behavioural and developmental intervention

ABA, EIBI, ESDM, PRT, JASPER, DIR/Floortime, parent-mediated programmes

Early structured support for an autistic child is recommended care. Which kind, how many hours, and towards what goal are all still genuinely argued over — and the arguing is the useful part.

Where it stands, condition by condition

The same therapy can be well supported for one problem and completely untested for another. This is the single most common place families are misled.

What it is

Three families of approach share this space, and they are closer to each other than their supporters usually admit.

Applied behaviour analysis (ABA) is an umbrella term, not a single method. Classical early intensive behavioural intervention used structured, adult-led teaching trials in long weekly blocks. Modern practice has largely moved away from that format.

Naturalistic developmental behavioural interventions (NDBI) — the Early Start Denver Model, Pivotal Response Treatment, JASPER — are where behavioural and developmental thinking met. The adult follows the child's lead, teaching inside play and daily routines. This is the group with the strongest evidence.

DIR / Floortime comes from the developmental side: the adult gets down on the floor, joins whatever the child is already doing, and builds back-and-forth exchanges from there. Its logic overlaps substantially with NDBI; its published evidence base is smaller.

In practice, a good programme today looks much the same whichever label is on the door: the child's own interests, natural settings, the family doing most of the hours, and goals the family chose.

How it is meant to work

All three rest on the same foundation — a child learns what they practise, and practises what is rewarding — and differ in where the reward comes from.

Classical ABA arranges the reward externally; NDBI and Floortime try to make the activity itself the reward, on the reasoning that a child who is enjoying an interaction will produce far more repetitions of it than a child who is complying with one.

That difference matters more than it sounds. Motivation determines dose, and dose is what drives learning.

What has actually been tested

• The most careful synthesis is Project AIM (Sandbank and colleagues), which by its 2023 update pooled 252 studies of non-drug interventions. Its findings should be read closely, because they are more sobering than the field's usual summaries.
• When studies at high risk of detection bias were set aside, exactly one significant pooled effect survived across every type of intervention: NDBI on diagnostic measures, around 0.30 — a small effect. Excluding caregiver- and teacher-rated outcomes, NDBI effects on social communication (about 0.36) and diagnostic measures held. Restricted to randomised trials only, NDBI retained a significant effect on adaptive behaviour.
• In other words, the approach that survives the strictest filters is the naturalistic, play-based one — and even there the effects are modest rather than transformative.
• The underlying record is incomplete. A 2024 audit of early-childhood autism trials found results posted for only 7% of completed registered trials, 36% of registered outcomes omitted from publications, and 61% of published outcomes never registered in the first place. That pattern makes published estimates more likely to be optimistic than pessimistic.
• Other reviews are more positive. A 2025 synthesis of eleven meta-analyses reports IQ gains of roughly 9–15 points with intensive behavioural and NDBI programmes, and language improvement — while noting that effects on core autistic features were much less consistent. A 2025 review of sixteen randomised trials found larger language effects for higher-intensity ABA-based programmes than lower-intensity ones, with several estimates graded low certainty.
• Harms are barely studied. Project AIM noted that most papers mentioning adverse events gave no account of how they were monitored. For a field this large, that is a real gap.

What we still do not know

• How many hours are actually needed. The familiar 25–40 hour figure is inherited, not established, and the 2024 dose analyses contradict each other.
• Which children respond to which approach — no reliable predictor exists, so the first programme is always partly a trial.
• Whether early gains persist into adolescence and adulthood, which almost no study has followed.
• Which outcomes should count. Measures built around reducing autistic behaviours answer a different question from measures of the child's skills, independence and wellbeing.

Risks and costs

• Load on the child and the family. A programme filling 30 hours a week leaves little childhood around it, and the parent delivering it is also the parent. Exhaustion is the most common reason good programmes stop.
• The goals, not the techniques, are what the long-standing criticism is about. Many autistic adults describe distress associated with intensive behavioural programmes, particularly where the aim was compliance or appearing typical. The research on this is largely survey-based and its methods are contested — but the underlying point stands on its own: a programme should be judged by whether the child's life got better, not by whether the child looks less autistic.
• Older aversive practices are not part of acceptable contemporary programmes. If anything punishment-based is described to you, that is a reason to leave.
• Cost. Intensive private programmes are among the most expensive things a family will be offered, and the evidence does not support the most intensive versions over moderate, well-targeted ones.

Questions to ask before you agree

1. What are the goals, in plain words — and would my child's life be better if we achieved them?
2. How much of this is led by my child's own interests, and how much is adult-directed?
3. How many hours, and what is that number based on for this child?
4. What is my role, and will you coach me rather than only treating my child?
5. How will we know it is working, who measures it, and does that person know which group my child is in?
6. How do you respond when my child is distressed during a session?

More in this section