If you are arranging an autism assessment, the ADI-R and the ADOS-2 are the two instruments you are most likely to encounter. They are the international gold standards for autism diagnosis in research and clinical practice. This article walks through what each one is, what they actually involve, what they do well, and what they do not do on their own.
Why two tools, not one
Autism diagnostic criteria require evidence across two domains: social communication, and restricted and repetitive behaviour [5]. They also require evidence that the picture has been present from early development. No single tool captures both well.
The two gold-standard tools each handle a different part of the diagnostic question:
- ADOS-2 is an observational assessment. It looks at how social communication is structured in the patient now, in a clinical setting.
- ADI-R is a developmental interview, usually with a parent or someone who knew the patient in childhood. It looks at the historical pattern from early life.
Used together, they triangulate the diagnostic picture. Used alongside the wider clinical interview, sensory profile and informant input, they form the core of a NICE-aligned assessment [3, 4].
ADOS-2: the Autism Diagnostic Observation Schedule, second edition
What it is
The ADOS-2 is a structured set of activities and conversations that a trained clinician guides the patient through [1]. Tasks are designed to elicit social communication behaviours, restricted or repetitive behaviour, and play (for younger children). The clinician scores observed behaviours rather than relying on what the patient says they do.
How it works
The ADOS-2 has five modules, chosen based on age and language level:
- Module 1: pre-verbal or single words (typically young children)
- Module 2: phrase speech (typically young children with some verbal language)
- Module 3: verbal children and adolescents with fluent speech
- Module 4: verbally fluent adolescents and adults
- Toddler Module: 12 to 30 months
For most autistic adults, Module 4 is used. The assessment runs around 45 to 60 minutes and includes activities such as describing a picture, telling a story, talking about emotions and relationships, and conversational exchange about everyday topics.
What it scores
Behaviours observed in the session are coded against the diagnostic criteria. The output is a series of standardised scores for social affect and restricted and repetitive behaviour. A cut-off score indicates whether the picture meets the autism threshold within the ADOS-2 framework. The score is not a diagnosis on its own; the clinician integrates it with the wider assessment.
What it does well
- It produces a structured observation that is reproducible across clinicians and providers
- It captures behaviours the patient may not self-report
- It works across language and literacy levels
- It is well established in research and clinical practice internationally
What it does not do
- It does not capture the developmental history; that is the ADI-R's role
- It does not capture how the patient functions outside the clinic; informant input matters
- It does not perfectly capture autistic women who mask effectively; the score can underestimate the picture, particularly for late-diagnosed adults
- It is not a test of intelligence or success; there is no pass or fail
The clinical literature increasingly recognises that ADOS-2 can miss masking autistic adults, particularly women, and that scores below the threshold do not rule out autism where the wider picture fits.
ADI-R: the Autism Diagnostic Interview-Revised
What it is
The ADI-R is a structured interview, usually with a parent or someone who knew the patient in childhood [2]. The interview covers the developmental trajectory across the diagnostic domains, with particular attention to the period between ages four and five (the most diagnostically informative for many autistic children).
How it works
The ADI-R is conducted with the informant, usually a parent. It takes around 90 to 120 minutes. The interviewer asks about specific behaviours and developmental milestones, with detailed follow-up where relevant. The questions cover:
- Communication development, including language milestones and use
- Reciprocal social interaction, including peer relationships and play
- Restricted, repetitive and stereotyped behaviour
- Specific concerns and behaviours observed across childhood
What it scores
The ADI-R produces algorithm scores in three domains: qualitative abnormalities in reciprocal social interaction, communication, and restricted/repetitive behaviours. Scores above the cut-off indicate that the developmental pattern meets the autism threshold within the ADI-R framework. As with ADOS-2, the result is not a diagnosis on its own.
What it does well
- It captures the developmental history in detail
- It anchors the diagnosis in the criterion of early-developmental onset
- It is reproducible across clinicians
- It is informative even when the patient is too anxious or masking to reveal the picture themselves
What it does not do
- It depends on the availability of a reliable informant who knew the patient in childhood
- It can miss autistic women whose presentations were quiet and gendered, where the parent did not notice or remember the picture as unusual at the time
- It is time-intensive; some private and NHS services do not always have time to complete it
- It is not a substitute for the clinical observation of the patient now
When a parent is not available, or memory is too partial to be useful, clinicians use alternative approaches: school reports, sibling memories, childhood photos, and structured developmental history-taking with the patient directly. A diagnosis is still possible without a full ADI-R, but the developmental section will take more clinical time.
How NeuroFX uses both
Standard adult autism assessment at NeuroFX uses both the ADOS-2 and the ADI-R wherever a developmental informant is available. The assessment also includes:
- A full clinical interview with the patient
- A sensory profile
- Information about masking and the home/public split
- A differential diagnosis review for co-occurring conditions
- Where suspected, combined ADHD and autism assessment covers both conditions in the same pathway
The integrated picture, not the individual tool scores, drives the diagnostic decision.
What this means in practice
- The ADI-R and ADOS-2 are gold-standard tools, used together where possible.
- Neither tool diagnoses autism on its own. The clinical synthesis does.
- A below-threshold ADOS-2 in a masking adult does not rule out autism.
- The ADI-R is much easier to complete with a parent or sibling who can remember childhood specifically. Where this is not available, the assessment uses other developmental evidence.
- Tools are part of a NICE-aligned assessment, not the whole of it.
When to speak to a professional
If autism is on your mind for yourself or your child, NHS routes start with GP referral. NeuroFX offers private autism assessment for adults and for children aged 6 and upwards using both the ADOS-2 and the ADI-R from our Bedford clinic. Seek same-day help via 111 (or 999 in an emergency) for any mental health crisis.
Sources
- Lord C, Rutter M, DiLavore PC, et al. Autism Diagnostic Observation Schedule, Second Edition (ADOS-2). Torrance: Western Psychological Services; 2012.
- Rutter M, Le Couteur A, Lord C. Autism Diagnostic Interview-Revised (ADI-R). Torrance: Western Psychological Services; 2003.
- NICE. Autism spectrum disorder in adults: diagnosis and management. CG142. National Institute for Health and Care Excellence. https://www.nice.org.uk/guidance/cg142
- NICE. Autism spectrum disorder in under 19s: recognition, referral and diagnosis. CG128. National Institute for Health and Care Excellence. https://www.nice.org.uk/guidance/cg128
- American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition, Text Revision (DSM-5-TR). Washington, DC: American Psychiatric Association Publishing; 2022.
