Sign Up
Login
Try for Free
Login
The ADI-R (Autism Diagnostic Interview-Revised) is a standardized, semi-structured caregiver interview that gathers developmental history and current behavior when autism is part of the question. Psychologists and other trained clinicians use it as one input to an autism evaluation, and it yields a categorical classification, not a diagnosis. This page covers how to write up ADI-R results, with a fictional sample and a results-section template.
Psychologists, developmental pediatricians, and clinicians trained on the ADI-R; publisher qualification level C
Autism evaluation teams, physicians, IEP and early-intervention teams, BC Autism Funding and NDIS reviewers, parents
300 to 600 words for the ADI-R section · administration 90 to 150 minutes including scoring
Semi-structured caregiver developmental-history interview
Autism diagnostic evaluations across the lifespan, mental age above two years; paired with direct observation for ABA authorization, school, BC Autism Funding, and NDIS support
Published by Western Psychological Services (2003); current edition as of July 2026; described here for write-up purposes, no test content reproduced
The Autism Diagnostic Interview-Revised (ADI-R) is a standardized, semi-structured interview conducted with a parent or caregiver who knows the person's early development well. Western Psychological Services published the revised edition in 2003, and it descends from the original ADI (Le Couteur and colleagues, 1994); its authors are Michael Rutter, Ann Le Couteur, and Catherine Lord. It is composed of 93 items across three functional domains, Language and Communication, Reciprocal Social Interactions, and Restricted, Repetitive, and Stereotyped Behaviors and Interests, and the interviewer codes the caregiver's descriptions across eight content areas covering background, early development, language, social development and play, interests, and behavior. It can be used with children and adults whose mental age is above two years, and administration with scoring runs about 90 to 150 minutes. As of July 2026 the 2003 edition remains current, with no successor edition released.
One point is load-bearing for write-ups: the ADI-R yields a categorical result, not a score. WPS states that because it focuses on behaviors rare in unaffected individuals, it "provides categorical results rather than scales or norms," so there are no standard scores, percentiles, or severity metrics to report, only whether a developmental-history pattern met the instrument's threshold. Its Comprehensive Algorithm Form supports two kinds of algorithm: Diagnostic Algorithms, based on developmental history and used to support a formal diagnosis, and Current Behavior Algorithms, which describe present functioning for treatment and educational planning and are not diagnostic. The original diagnostic algorithm was built around DSM-IV Autistic Disorder, which creates a genuine alignment gap with the DSM-5-TR two-domain model, so the classification supports rather than settles a diagnosis. Convention pairs the caregiver interview with direct observation such as the ADOS-2 and a rating scale such as the SRS-2, and the whole set feeds a psychological evaluation report, not a verdict on its own.
Autism diagnostic clinics, developmental-behavioral pediatricians, child and adult autism services, and psychologists use the ADI-R as the developmental-history component of an evaluation, drawing on a caregiver who can describe early childhood in detail. It rarely stands alone: convention pairs it with direct observation, usually the ADOS-2, a social-communication rating scale such as the SRS-2, and cognitive and adaptive measures such as an adaptive interview, and the whole set anchors a psychological evaluation report. The caregiver-history piece is where the ADI-R does work the others cannot: it carries weight when direct observation understates difficulty, as with an adolescent who has learned to mask socially or an adult whose current presentation is subtle but whose early history is clear. Two boundaries belong in the report itself. First, the diagnostic algorithm depends on early developmental history, so it needs a caregiver who was present and can recall that period; without one, for many adults, the algorithm cannot be validly completed. Second, WPS states that remote administration by phone or videoconference "has not been specifically validated," so a remote interview belongs in a limitations line, not a silent footnote.
No statute, payer, or publisher mandates a format for the ADI-R section of an autism evaluation. The sequence below is the convention experienced evaluators converge on because it survives review, and because it is built around one rule the top search results routinely break: report the classification and describe the developmental history in your own words, and leave the interview's questions, coding conventions, algorithm totals, and cutoff numbers out. That omission is a test-security obligation under APA Ethics Standard 9.11 and WPS policy, not an oversight. Each section below carries the pitfall that most often undermines it.
Instrument, informant, and administration. Name the ADI-R and its 2003 edition, identify the informant and their relationship to the person and how well placed they are to recall early childhood, note who administered and coded the interview and their ADI-R training, and state that it was one component of the evaluation. Pitfall: not identifying the informant or their recall quality, or writing the interview up as though it produced the diagnosis on its own.
Developmental-history findings, in your own words. Describe what the caregiver reported across the three domains, early social development and reciprocity, language acquisition and any loss of skills, and restricted or repetitive behavior, with attention to the early-childhood window the diagnostic algorithm focuses on. Pitfall: reproducing the interview questions, prompts, or coding conventions. Transcribing the instrument exposes protected content and erodes its validity.
Algorithm used and classification, reported not decoded. State which algorithm you applied, the developmental-history Diagnostic Algorithm for diagnostic support or a Current Behavior Algorithm for planning, and whether the diagnostic classification was consistent with autism. Pitfall: printing domain totals or cutoff values, presenting the categorical result as a score or percentile, or mixing the two algorithm families so a current-behavior result reads as diagnostic.
Integration with observation, history, and other measures. Place the caregiver history next to direct observation such as the ADOS-2, prior records, and any rating, cognitive, or adaptive data, and say where the sources agree and where they diverge. Pitfall: an ADI-R section that never connects to the observation, or that leaves an interview-versus-observation disagreement unexplained.
Diagnostic formulation under DSM-5-TR. State the diagnostic determination and the clinical judgment behind it, with the ADI-R named as one input, and note that its algorithm reflects DSM-IV criteria while the diagnosis is made against DSM-5-TR. Pitfall: a diagnosis that rests on the classification alone. An ADI-R classification is not a DSM-5-TR diagnosis, and a result not consistent with autism does not rule autism out.
Limitations and informant recall. Note anything that bears on interpretation: the informant's recall confidence, retrospective bias, especially for adults reporting distant childhood, any masking or camouflaging that shaped the observation, and in-person versus remote administration. Pitfall: silence on informant recall quality, the single most common gap for adults and older adolescents.
Recommendations linkage. Tie each recommendation, whether supports, services, or further assessment, to a specific finding. Pitfall: recommendations that could follow any evaluation and trace back to nothing in the interview.
ADI-R RESULTS SECTION Instrument/edition: ADI-R (2003) Informant: [relationship, recall quality] Administered/coded by: [name, credentials, ADI-R training] Date: [ ] DEVELOPMENTAL-HISTORY FINDINGS (narrative, in your own words) [Early social development and reciprocity, language acquisition and any loss of skills, restricted and repetitive behavior; describe what the caregiver reported, not the interview questions] ALGORITHM USED AND CLASSIFICATION [Diagnostic Algorithm (developmental history) or Current Behavior Algorithm] [Whether the diagnostic classification was consistent with autism] [Categorical result only; no domain totals, cutoffs, or item scores] INTEGRATION WITH OBSERVATION, HISTORY, AND OTHER MEASURES [Direct observation (e.g., ADOS-2), records, rating scales, cognitive and adaptive data; state where they converge or diverge and reconcile it] DSM-5-TR FORMULATION [The diagnostic determination and the clinical judgment behind it; the ADI-R is one input, and its algorithm reflects DSM-IV criteria] LIMITATIONS AND INFORMANT RECALL [Recall confidence and retrospective bias, masking or camouflaging, in-person vs remote administration] RECOMMENDATIONS LINKAGE [Each recommendation tied to a specific finding] Evaluator signature / credentials: Date:
Free to use and share, no signup. The PDF includes a one-page cheat sheet with section-by-section pitfalls and a pre-sign checklist; the DOCX is the blank results-section skeleton, ready to adapt.
Scenario: a 15-year-old is referred by the school counselor and family doctor for social difficulties and long-standing anxiety. Language was on time, and the adolescent is verbally fluent and socially motivated, but the mother describes years of one-sided friendships and rigid routines. The ADI-R was completed with the mother as informant, and an ADOS-2 was administered separately; the direct observation came in nearer the threshold than the developmental history. This is the ADI-R section only, and it deliberately reports the classification and narrative without item content, coding, or cutoff numbers, which are protected. All details are fictional.
Client: M.R., 15 · Informant: mother, lived with client since birth, strong early-childhood recall · Evaluator: L. Okafor, PhD, Licensed Psychologist (ADI-R trained) · Interview date: 07/09/2026 · Report date: 07/15/2026
Instrument, informant, and administration: The Autism Diagnostic Interview-Revised (ADI-R, 2003) was completed with M.R.'s mother, who has lived with him since birth and recalled the early-childhood period in detail. The interview was administered and coded in person by the evaluator, who is trained on the ADI-R, and it served as the developmental-history component of a broader autism evaluation. Informant recall was rated as good, and the account was internally consistent with school records from the early grades.
Developmental-history findings: Across the reciprocal social domain, the mother described early and persistent difficulty forming and keeping friendships, a tendency to relate to adults rather than peers, and limited sharing of interest and enjoyment in early childhood. Language milestones were reported on time, but early communication featured repetitive questioning and a formal, one-sided conversational style. In the restricted and repetitive domain, she reported intense circumscribed interests that dominated play, strong insistence on routine, and marked distress at small changes, together with early sensitivity to certain sounds and textures. The pattern was most pronounced in the early-childhood window the diagnostic algorithm focuses on.
Algorithm used and classification: The developmental-history Diagnostic Algorithm was applied. The pattern the mother reported was consistent with autism on that algorithm. Consistent with test-security convention, domain totals and cutoff values are not reproduced here, and the result is reported as a categorical classification rather than a score.
Integration with observation, history, and other measures: The caregiver history and the direct observation diverged in degree. On the ADOS-2, M.R. presented as verbally fluent and socially motivated, and the current-observation profile fell nearer the threshold than the developmental history did. This pattern is expected for an adolescent who has learned to compensate socially, and the divergence is reconciled rather than averaged: the detailed early history carries weight where present observation understates difficulty because of camouflaging. A completed social-communication rating scale fell in the elevated range, and school records from the early grades corroborated the caregiver account.
DSM-5-TR formulation: Integrating the ADI-R developmental-history classification, the ADOS-2 observation, the rating-scale data, and clinical judgment, M.R. meets DSM-5-TR criteria for autism spectrum disorder, requiring support (Level 1), with co-occurring anxiety documented as an associated condition rather than the primary explanation. The ADI-R is one input to this determination, not the determination itself, and its algorithm reflects DSM-IV criteria while the diagnosis is made against DSM-5-TR.
Limitations and informant recall: The findings rest on a single informant's retrospective account; recall was judged good and was corroborated, but it remains retrospective. The lower current-observation profile reflects likely camouflaging and does not negate the developmental history. Administration was in person; no remote interview was involved.
Recommendations linkage: The reciprocal-communication history supports recommendation 1 (social-communication support at school with structured peer opportunities). The routine and sensory findings support recommendation 2 (predictability and sensory accommodations in the classroom). The co-occurring anxiety supports recommendation 3 (anxiety-focused therapy that accounts for an autistic profile). The diagnosis supports recommendation 4 (eligibility and service discussions by the school and payer teams applying their own criteria).
This sample is fictional and for educational purposes. It does not describe a real person or record, and the classification and details are invented for illustration and correspond to no real child or record.
Writing these after every session? BastionGPT drafts complete notes from bullets, dictation, or a transcript.
Generate a note from bulletsWrite the ADI-R section knowing which decision it will feed, and label the strength of each requirement honestly. No United States or Australian authority legally requires the ADI-R for an autism diagnosis. DSM-5-TR governs the diagnosis, and the American Academy of Pediatrics clinical report (Hyman, Levy, and Myers, 2020, reaffirmed October 2025) affirms that clinicians comfortable with DSM-5 criteria can diagnose without mandating any single instrument. What names the ADI-R is payer or program policy, always as an example rather than a rule: Aetna's clinical policy lists it among diagnostic tools used with clinical assessment, and TRICARE's Autism Care Demonstration has encouraged standardized tools "such as the ADOS-2 and the ADI-R" while recognizing others (PAYER POLICY, not a clinical rule). The clear exception is jurisdictional: in British Columbia, for a private diagnosis to qualify for Autism Funding, both the ADOS-2 and the ADI-R are required instruments, and the province accepts the ADI-R done virtually but not the ADOS (PROGRAM POLICY). The Canadian Paediatric Society notes only that "in some jurisdictions, the ADOS and ADI-R are required," without itself mandating them. In Australia the Autism CRC National Guideline (second edition 2023, NHMRC-approved) calls for comprehensive assessment without naming a required tool, and NDIS access turns on a DSM-5 diagnosis with functional-impact evidence. Interview interpretation and integration time is billed under the psychological and developmental testing evaluation family by bare code number (for example 96130 and 96131 in the United States, which are time-based and count data integration and report writing), with structured administration under 96136 through 96139 and the relevant MBS item numbers in Australia, as payer policy with plan-specific rules. The format is a convention; the classification, the narrative, and the integration are the requirement.
Edition currency and the qualification boundary are the ADI-R's live defensibility questions. As of July 2026 the 2003 edition remains current on the WPS product page, with no successor released, though an ADI Third Edition is reported to be in publisher-funded validation, so an edition claim should carry its date. More consequential for interpretation is that the diagnostic algorithm was built around DSM-IV Autistic Disorder, which is why DSM-5-based research algorithms now exist, Kim and Lord's toddler algorithms for children under four and Lampinen and colleagues' algorithms for ages four to seventeen, framed by their authors as "an alternative method for summarizing ASD symptoms in a DSM-5-compatible manner" rather than a replacement for the published scoring. Read the classification with that gap in mind. Two boundaries belong in the write-up when they apply. First, who may use it: the ADI-R is qualification level C, and WPS states it "only qualifies individuals for test purchase, not for test use," which licensing boards and scope of practice govern; clinical-use training and the separate research-reliability standard run by the test authors are different things, and published sensitivity can exceed real-world clinic performance because of that gap. Second, administration mode: WPS holds that conducting the ADI-R remotely "has not been specifically validated," so a phone or videoconference interview is a stated limitation, even where a program such as BC accepts it.
ADI-R is a trademark of Western Psychological Services (WPS). BastionGPT is not affiliated with, or endorsed by, the publisher. This page reproduces no test items, stimuli, norms, or scoring materials.
There is no payer audit series for autism interview write-ups; the accountability here is ethical and psychometric, and it is pointed. The ADI-R has not been reviewed in the Buros Mental Measurements Yearbook, most likely because it "provides categorical results rather than scales or norms," so clinicians expecting a Buros review should be told none appears to exist. The peer-reviewed evidence is real but bounded: a 2021 meta-analysis put ADI-R sensitivity at .75 and specificity at .82, with specificity higher in research samples (.85) than clinical ones (.72), and the observation-based ADOS-2 outperforming it. The errors below are write-up errors, not administration errors. The BastionGPT Clinical Advisory Board sees the same ones most often in ADI-R report reviews:
BastionGPT is specifically trained, tuned, and clinically tested on psychological and psychoeducational evaluation reports.
See how clinicians use it day to day on the AI therapy notes page.
Many BastionGPT users report saving more than 90 minutes per day on documentation.
HIPAA-compliant with a signed BAA on every plan. Your data is never used to train models. BastionGPT drafts, you review and sign.
The ADI-R does not produce scores in the usual sense. WPS states that because it focuses on behaviors rare in unaffected individuals, it "provides categorical results rather than scales or norms," so there are no standard scores, percentiles, or severity metrics. What the developmental-history Diagnostic Algorithm yields is a categorical result: whether the pattern the caregiver reported met the instrument's threshold for autism. Report that classification and describe the history in your own words, and leave domain totals and cutoff values out, both because they are protected content and because the result is not a number to be graphed. This is a key difference from norm-referenced measures such as a cognitive battery, where standard scores and percentiles are the output.
Not in most places. No United States or Australian authority requires it: DSM-5-TR governs the diagnosis, and the American Academy of Pediatrics affirms that no single instrument is mandated (LAW and guidance). Payers and programs name it only as an example among validated tools, which is PAYER POLICY, not a requirement. The clear exception is jurisdictional: in British Columbia, a private diagnosis eligible for Autism Funding requires both the ADOS-2 and the ADI-R (PROGRAM POLICY), and the Canadian Paediatric Society notes only that "in some jurisdictions" they are required. So the honest write-up names the ADI-R as one validated caregiver-history instrument that adds standardized developmental data, and routes the requirement question to the specific payer or program, most often through the psychological evaluation report that carries the diagnosis.
They measure different things and are designed to be used together. The ADI-R is a caregiver interview about developmental history and current behavior; the ADOS-2 is a direct, standardized observation of the person. Because they draw on independent sources, they make additive contributions, which is why the pairing is often called the research gold standard, and why specificity tends to improve when they are combined. They also disagree: agreement between caregiver interview and direct observation is often only fair, and the disagreement is greater for older and more subtle presentations. A report that uses both should state where they converge and reconcile where they diverge, rather than averaging them or treating a disagreement as a failure. The ADI-R carries particular weight when the observation understates difficulty, as with camouflaging.
The 2003 edition is current as of July 2026, and WPS lists no successor on its product page, though an ADI Third Edition is reported to be in publisher-funded validation. The live issue is not the edition date but the diagnostic framework: the published algorithm was built around DSM-IV Autistic Disorder, so it predates the DSM-5 two-domain model. Researchers have developed DSM-5-based algorithms, Kim and Lord's for toddlers and Lampinen and colleagues' for ages four to seventeen, but their authors describe these as "an alternative method for summarizing ASD symptoms in a DSM-5-compatible manner," not the official scoring. The practical consequence for a write-up is to state that the classification supports the DSM-5-TR determination rather than mapping onto it one to one, and to date any edition claim.
The diagnostic algorithm depends on early developmental history, so it needs an informant who was present in childhood and can recall that period. For many adults, no such informant is available, and in that case the diagnostic algorithm cannot be validly completed. The right move is to document the limitation explicitly, shift evidentiary weight to current observation, self-report, and any available records, and state a clear caveat about diagnostic confidence rather than reporting a classification the data cannot support. Even where an informant is available, adult and older-adolescent histories often rest on distant recall, so note the informant's recall confidence. Some teams turn to adult-oriented structured interviews for this situation; whatever the approach, the report should name what the missing early history means for the conclusion.
WPS states plainly that conducting the ADI-R by phone or videoconference "has not been specifically validated," and recommends a secure platform that mimics face-to-face contact if it is done remotely. That is not the same as a prohibition: some programs accept it, and British Columbia explicitly accepts the ADI-R done virtually even though it does not accept a remote ADOS. The write-up rule is to disclose the mode: if the interview was conducted remotely, say so, describe the method, and note that the norms and validation base assume in-person administration. Treat remote administration as a documented limitation rather than a silent detail, and check the specific program's stance before relying on a remote interview for a funding decision.
Domain-level narrative and the classification, integrated into the formulation, and nothing more granular. Describe what the developmental history and current-behavior information showed across the social, communication, and restricted or repetitive domains, name which algorithm was used, and state whether the diagnostic classification was consistent with autism. Do not transcribe the interview questions, list item-level codes, or print domain totals or cutoff values. Reproducing that content breaches APA Ethics Standard 9.11 and WPS policy and erodes the instrument, and it does not make the report stronger. Then integrate the ADI-R findings with direct observation and records in the evaluation report, so the classification reads as one converging line of evidence rather than a standalone result.
This page does not reproduce them, and a report should not either. The ADI-R is a secure test: its items, prompts, and coding conventions are protected under APA Ethics Standard 9.11 and the publisher's agreements, and their value depends on informants not having seen them in advance. It is fair to describe the instrument in general terms, a 93-item, semi-structured caregiver interview covering early development, language, social development and play, and restricted or repetitive behavior across three domains, and to name what each domain assesses. It is not appropriate to publish the actual questions or the scoring keys, whether in a report, a template, or a web page. Clinicians who need the materials obtain them through WPS at the required qualification level.
BastionGPT drafts and checks the results section, not the clinical judgment. Paste a summary of the developmental-history findings and the classification, and it returns an organized narrative for your review, with no interview questions or item content required. It can cross-check an existing draft for the gaps reviewers flag, a classification written as a diagnosis, a current-behavior result presented as diagnostic, or an interview-versus-observation disagreement left unreconciled, and it can turn the findings into a plain-language summary for parents and referrers. It is HIPAA-compliant and keeps the classification as one input while you make the diagnosis. See how clinicians use it day to day on the AI therapy notes page.
The instrument facts and compliance claims on this page trace to these sources, last verified July 2026:
Educational content, not legal or billing advice. Sample notes are fictional. Follow your organization's policies and your board, payer, and jurisdiction requirements.