Assessment & Research

Non-pharmacological interventions for autistic children: An umbrella review.

Trembath et al. (2023) · Autism : the international journal of research and practice 2023
★ The Verdict

No single non-drug treatment works for every autistic child, so pick targets, probe, and pivot fast.

✓ Read this if BCBAs choosing interventions for autistic clients in clinic, home, or school settings.
✗ Skip if Practitioners who only run pharmacological or medical protocols.

01Research in Context

01

What this study did

Trembath et al. (2023) pulled together every high-quality review of non-drug treatments for autistic children. They looked at meta-analyses and systematic reviews, not single trials. The goal was to see which interventions help which kids and which outcomes.

The team screened thousands of papers and kept only the strongest reviews. They mapped what works, what does not, and where the evidence is still thin.

02

What they found

No single intervention helped every child or every outcome. Parent-training programs, sleep plans, and behavioral packages each showed some wins, but never across the board. The clearest signal: you must match the tool to the child and goal.

Even within well-studied areas like parent-mediated work, the size of benefit stayed small. Moderators such as child language level, parent adherence, or study quality changed results, but not in a steady way.

03

How this fits with other research

Stewart et al. (2018) and Lee et al. (2022) both report small but real gains from parent coaching. Trembath et al. (2023) agree, yet stress the gains are too tiny and shaky to call parent training a universal best pick. The umbrella view widens the lens and cools the enthusiasm.

Bergmann et al. (2019) show behavioral sleep interventions add about 24 minutes of sleep. The umbrella review folds this in as a bright spot, but notes the benefit is narrow: sleep improves, yet daytime behavior or family stress may not budge.

Heyvaert et al. (2014) find behavioral interventions cut problem behavior, especially after a functional analysis. Trembath et al. (2023) keep that on the "likely helpful" list, again with the warning that effects vary and long-term data are scarce.

Lord et al. (2005) warned the field about weak methods. Eighteen years later, Trembath et al. (2023) echo the same worry: too many different measures, too few RCTs, and little replication still block firm choices.

04

Why it matters

For BCBAs, the takeaway is humility. Use the intervention that has partial support for your specific target, but do not expect miracles. Measure the child’s baseline, run brief probes, and track data closely. If progress stalls after a fair trial, pivot early rather than doubling down. Build parent buy-in and monitor adherence; reviews show that factor alone can make or break small effects. Finally, add common outcome tools (like ABC scores or sleep logs) so your data can stack up with future reviews instead of adding to the 327-measure mess Provenzani et al. (2020) describe.

Free CEUs

Want CEUs on This Topic?

The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.

Join Free →
→ Action — try this Monday

Pick one target behavior, run a one-week probe with your chosen evidence-based package, and graph daily so you can pivot quickly if gains are flat.

02At a glance

Intervention
not applicable
Design
systematic review
Population
autism spectrum disorder
Finding
not reported

03Original abstract

What is already known about the topic?The delivery of evidence-based interventions is an important part of the clinical pathway for many autistic children and their families. However, parents, practitioners, and policymakers face challenges making evidence informed decisions, due to the wide variety of interventions available and the large, and often inconsistent, body of evidence regarding their effectiveness.What this paper adds?This is a comprehensive umbrella review, also known as a 'review of reviews', which examined the range of interventions available, the evidence for their effectiveness, and whether effects were influenced by factors relating to individual children (e.g. chronological age, core autism characteristics, and related skills) or the ways interventions were delivered (by whom and in what setting, format, mode, and amount). There was evidence for positive therapeutic effects for some, but not all, interventions. No single intervention had a positive effect for all child and family outcomes of interest. The influence of child and delivery characteristics on effects was unclear.Implications for practice, research, and policyThe findings provide parents, practitioners, and policymakers with a synthesis of the research evidence to inform decision-making and highlight the importance of individualised approaches in the absence of clear and consistent evidence. The findings also highlight the need to improve consistency and completeness in reporting of research studies, so that the same questions may be answered more comprehensively in the future.

Autism : the international journal of research and practice, 2023 · doi:10.1177/13623613221119368