The bottom line
  • Some approaches showed small, targeted average benefits in randomized trials.
  • The strength of findings fell when the review removed outcomes at higher risk of bias.
  • Possible adverse effects were too poorly monitored to balance benefits against harms.
Studies
252
Participants
13,304
Age
Under 8
01

Why this review matters

Families are often presented with confident claims about early support. Project AIM pooled a very broad evidence base: 289 reports covering 252 controlled studies and 13,304 participants under eight.

A review this large can reveal patterns that a single trial cannot. It still cannot tell you which option is right for one particular child.

02

What the researchers found

In randomized trials, several intervention categories were associated with improvements in specific outcomes. Average effects were generally small to moderate and depended on what was measured.

When analyses excluded caregiver- or teacher-reported outcomes and then outcomes at high risk of detection bias, far fewer results remained statistically clear. This does not mean those supports never help. It means confidence should match the quality of the measurement.

03

What this means for a family

Ask what goal a support is designed to address: communication, play, participation, distress, daily living or something else. A result in one area should not be sold as a global transformation.

A useful plan should have goals that matter to the child and family, a way to notice both benefit and burden, and permission to adapt or stop.

Limitations to keep in view

  • Interventions and outcomes were highly varied.
  • Many outcomes relied on people who knew which treatment was received.
  • Adverse events were poorly monitored and reported.

A careful next step

Use this evidence to ask sharper questions, not to choose support without an individualized discussion.

Original source

Autism intervention meta-analysis of early childhood studies (Project AIM): updated systematic review and secondary analysis

Sandbank M, Bottema-Beutel K, Crowley LaPoint S, et al.

BMJ · 2023

PMID 37963634DOI 10.1136/bmj-2023-076733
Open the source

This article provides general information and does not replace individualized medical, psychological or educational advice.