Questions to Ask About Any Autism Intervention’s Evidence

When a service describes an autism intervention as evidence-based, the next question is: evidence for what outcome, for whom and under which conditions? A broad label cannot answer whether the proposed support matches a particular person’s priorities, circumstances and needs.

This guide provides questions for a discussion with a provider. It does not rank all interventions or recommend a programme. The aim is to make the claimed benefit, the research behind it and the practical decision easier to examine together.

Define the claim before reviewing the evidence

Ask the provider to describe what the intervention is intended to change in everyday terms. “Improve communication” could refer to several different outcomes. Expressing a preference, understanding instructions and changing a score on a research measure are related possibilities, but they are not the same claim.

Identify what will actually be offered: the activities, setting, frequency, people delivering them and expected involvement between sessions. A familiar programme name does not establish that the local service is delivered in the same way as the intervention in a study.

Write a short question linking the proposal to the person’s priority. For example: “What evidence suggests this approach could help me communicate a change of plan at work?” A specific question makes it easier to notice when the answer shifts to an unrelated outcome.

Ask for the original research and its scope

The NIH overview of clinical studies explains that different designs answer different questions. Ask whether the evidence comes from a comparison study, a trial, qualitative accounts or another approach, and what that design allows the authors to conclude.

Request a direct source rather than only a brochure, testimonial or list of impressive institutions. If a provider cites a review, ask about its publication date and search period. A recent webpage may be summarising evidence gathered several years earlier.

Check the age group, setting and outcome. Evidence about young children does not automatically establish effectiveness for autistic adults. A result during a structured task does not automatically establish an improvement across school, home and community life.

Discuss the quality and limits of the evidence

The 2023 Project AIM review examined non-drug interventions for young autistic children. It found evidence for some specific outcomes, while conclusions depended on study quality and how outcomes were measured. Inadequate monitoring of adverse events limited assessment of benefits against possible unwanted effects. This is not a verdict about every service or age group.

A 2026 correction added fuller reference information to supplementary tables. It is useful to check corrections as part of reading a source; a correction’s meaning depends on what it actually changes.

Ask the provider to explain the strongest relevant evidence and the main uncertainty in the same conversation. An answer that acknowledges limitations can be more informative than a guarantee. Conversely, uncertainty should not be used to avoid explaining why the proposed approach is reasonable for the stated purpose.

Check fit with the person and the setting

The Canadian Paediatric Society’s post-diagnostic guidance emphasises individualised care, coordination and informing families about options and evidence. Its 2019 statement concerns children; it does not determine whether a particular service is suitable for an adult or guarantee access to a programme.

Ask how the proposed goal reflects the autistic person’s priorities. A change that matters to a researcher, parent or staff member may need further discussion before becoming the person’s own goal. Identify whose view each outcome represents.

Consider communication access, sensory needs, time, travel and the resources required between sessions. If the service differs from the research setting, ask what that difference means for the recommendation. A plan needs a practical explanation as well as a citation.

Do not assume that doing more is necessarily better. Ask why the proposed schedule is appropriate and how fatigue, competing activities and preferences will be considered. The decision should address the actual person rather than rely only on a standard package.

Ask how unwanted effects and concerns are monitored

A useful review considers more than the desired outcome. Ask how the provider will learn about distress, discomfort, loss of valued activities or other concerns that arise during the plan. Agree a way for the person to communicate that something is not working.

Clarify whom to contact, what would prompt a pause and how a concern will be documented. If a new health symptom appears, seek appropriate medical advice rather than assuming it is an expected part of progress. A support provider should explain the limits of its own role.

Keep uncertainty accurate. Limited reporting of unwanted effects does not prove that an intervention is harmless, and it does not prove that every participant will be harmed. It means the decision needs an honest account of what is known and a practical response to problems.

Agree a manageable review before starting

Ask what information will be collected and how it connects to the goal. A brief account of an everyday activity may complement formal measures. Avoid collecting private information simply because it is easy to record; clarify its purpose and handling.

In a fictional example, Omar is considering a programme intended to help with group conversations. He asks whether the cited studies examined adults, what participants actually practised and whether outcomes were measured outside sessions. He also asks how the plan would account for his preference for written preparation. Those answers inform the decision together.

Set a review point and identify the available choices: continue, adapt, obtain more information or discuss ending the service. A review should be a real decision, not a formality that always produces the same recommendation.

Common questions