Overview
The Autism Diagnostic Interview, Revised (ADI-R; Rutter, LeCouteur, & Lord, 2003) is a comprehensive semi-structured interview is conducted with parents or caretakers who have knowledge about the individual’s current behavior and developmental history. The questions address the functional domains related to autism spectrum disorder (ASD): Language/Communication; Reciprocal Social Interactions; and Restricted, Repetitive, and Stereotyped Behaviors and Interests. It also queries clinically relevant behaviors such as aggression, self-injury, and possible epileptic features. The measure consists of 93 questions with extensive probes for each question. . Examiners score responses to each question using standardized criteria, which are transferred onto the scoring algorithm. The single Comprehensive Algorithm Form permits calculation and interpretation of any one of five, age-specific ADI-R algorithms (two Diagnostic Algorithms based on developmental history and used for formal diagnosis, and three Current Behavior Algorithms focusing on present functioning and used for treatment and educational planning). The ADI-R focuses on behaviors that are rare in nonaffected individuals, providing categorical results rather than scales or norms. Results can be used to support a diagnosis of ASD or to determine clinical needs of groups in which a high rate of ASD might be expected (e.g., severe language impairments, congenital blindness, institutional deprivation). The ADI-R is frequently used in the research community, which has supported evidence of the reliability and validity of its results. Accurate administration and coding of the ADI-R are highly standardized and valid assessment requires training; one training option is through the Training Package offered by WPS.
Summary
Age: 2 years - 100 years
Time to Administer: 1.5 - 2.5 hours
Method of Administration: Semi-structured interview; 93 items in three functional domains. Responses are coded in eight content areas. Yields algorithm scores that are compared with categorical cut-off scores.
Subscales: Notes
-Age Range: Children and adults with a mental age over 2 years
-Instrument includes information related to developmental history.
-Instrument requires special training to administer and score.
Autism Related Research
Barton, Robins, Jashar, Brennan, Fein (2013)
Age Range: 16.79 to 39.36 months
Sample Size: 422
Topics Addressed:
Diagnostic sensitivity and specificity
Outcome:Barton, Robins, Jashar, Brennan, Fein (2013)
Toddlers are vulnerable to stringent requirements for score thresholds for ADI-R domains A (social-communication symptoms) and B (restricted/repetitive symptoms). Receiver operating characteristic (ROC) curves mapped on cut-off sums for ADOS and ADI-R for participants. Optimal DSM-5 criteria (sensitivity= 0.93, specificity= 0.074) required meeting the ROC-determined cut-off criteria for 2/3 Domain A criteria, and 1 point for 1/4 Domain B criteria. These modifications will help ensure ASD is identified accurately in young children.
Conclusion: increased sensitivity and adequate specificity in this study was achieved when ADI-R domain level threshold was reduced for both domains A and B; and when ADI-R symptom-level threshold was increased in domain A.
De Bildt, Oosterling, van Lang, Kuijper, Dekker, Sytema, . . . de Jonge, (2013)
Age Range: 3-18 years
Sample Size: 1,204 Dutch children
Topics Addressed:
Diagnostic sensitivity and specificity
Outcome:De Bildt, Oosterling, van Lang, Kuijper, Dekker, Sytema, . . . de Jonge, (2013)
ADI-R well-discriminated ASD from non-ASD with ID. All other criteria were sensitive at the cost of specificity, with the risk of over-inclusiveness.
Conclusion: using the ADI-R and ADOS in combination to identify children with ASD and MR increases specificity than using ADI-R alone (i.e., reduced false positives).
Hus & Lord, (2013)
Age Range: 4-18 years
Sample Size: 2,334
Topics Addressed:
Implication for use of scores as a measure of ASD severity
Outcome:Hus & Lord, (2013)
Conclusion: with statistical adjustment for expressive language level (and in some cases, age) at the time of interview, ADI-Diagnostic and ADI-Current domain and total scores can be used as estimates of the severity of core ASD symptoms.
Lecavalier, Aman, Scahill, McDougle, McCracken, et al. (2006)
Age Range: 5–17 years
Sample Size: 226
Topics Addressed:
Factor structure, internal consistency, convergent validity
Outcome:Lecavalier, Aman, Scahill, McDougle, McCracken, et al. (2006)
Internal consistency (coefficient alpha) of domain scores = 0.54–0.84. Convergent validity (Spearman-ranked correlation coefficients) – Social and total ADI-R had highest correlations to other instruments, range = -0.29 to 0.35, depending on scale and domain.
Conclusion: The ADI-R contributes uniquely to ASD diagnosis but should undergo continual study with varying sample sizes and clinical presentations.
Risi, Lord, Gotham, Corsello, Chrysler, et al. (2006)
Age Range: 1.5–14 years
Sample Size: 1,297
Topics Addressed:
Diagnostic sensitivity and specificity
Outcome:Risi, Lord, Gotham, Corsello, Chrysler, et al. (2006)
Strict autism criteria used in combination with ADOS – 80% or higher for U.S. sample, 75% or higher for Canadian sample; lower for single use and use for other PDDs.
Conclusion: ADI-R and ADOS make independent, additive contributions to clinician judgment that result in more consistent and rigorous application of diagnostic criteria.
Saemundsen, Magnússon, Smári, & Sigurdardóttir (2003)
Age Range: 2–9.5 years
Sample Size: 54
Topics Addressed:
Concurrent validity
Outcome:Saemundsen, Magnússon, Smári, & Sigurdardóttir (2003)
Significant correlations found between the ADI-R domains and total score and the CARS total score, which supports concurrent validity of the two. –The CARS classified more cases with autism than the ADI-R.
Conclusion: The CARS represents a broader diagnostic concept of autism than the ADI-R and the authors suggested considering it as a screening tool.
Wiggins & Robins (2008)
Age Range: 1.5–3 years
Sample Size: 142
Topics Addressed:
Diagnostic agreement across autism measures used for young children
Outcome:Wiggins & Robins (2008)
Agreement improved with removal of Behavioral Domain of ADI-R (percent agreement with other measures):ADOS: AU class = 0.790, non-AU class = 0.701 CARS: AU class = 0.708 , non-AU class = 0.753. Conclusion: stereotyped interests and behaviors may not be as relevant to the ADI-R as other diagnostic criteria when evaluating toddlers for ASD.
LeCouter, Haden, Hammal, & McConachie (2008)
Age Range: 2–4 years
Sample Size: 101
Topics Addressed:
Concurrent validity
Outcome:LeCouter, Haden, Hammal, & McConachie (2008)
Agreement with ADOS: AU Social Interaction = 78%; AU Communication = 74%; Above/below AU cutoff = 81%, Above/below spectrum cutoff = 78%.
Conclusion: agreement is good between these instruments. They have a complementary effect in aiding diagnosis and confirm the importance of a multidisciplinary assessment process with access to information from different sources and settings.
Ventola, Kleinman, Pandey, Barton, Allen, Green, Robins, & Fein (2006
Age Range: 1.5-2.5 years
Sample Size: 45
Topics Addressed:
Concurrent validity
Outcome:Ventola, Kleinman, Pandey, Barton, Allen, Green, Robins, & Fein (2006
Cohen’s kappa:
ADOS and clinical judgement = 0.593
ADOS and CARS = 0.619
CARS and clinical judgment = 0.691
ADI-R and ADOS = 0.066
ADI-R and CARS = 0.095
ADI-R and clinical judgment = 0.153.
Conclusion: For assessing toddlers, ADOS, CARS, and clinical judgment agreed with each other but not with the ADI-R in that many children classified with ASD by the other measures were not classified with autism by the ADI-R because they did not display enough repetitive behaviors and stereotyped interests. The authors noted that modification of ADI-R criteria might be more sensitive if the requirement for repetitive behaviors was less stringent.
Mazefsky & Oswald (2006)
Age Range: 1.75–8 years
Sample Size: 78
Topics Addressed:
Discriminative validity of ADOS, ADI-R, and GARS
Outcome:Mazefsky & Oswald (2006)
The ADOS and ADI-R led to approximately 75% agreement with team diagnoses.
Conclusions: The GARS was generally ineffective at discriminating between children with various team diagnoses and consistently underestimated the likelihood of ASD.
Gray, Tonge, & Sweeney (2008)
Age Range: 1.5–4.5 years
Sample Size: 209
Topics Addressed:
Diagnostic validity of ADI-R and ADOS for young children with and without autism
Outcome:Gray, Tonge, & Sweeney (2008)
Conclusion: Those with ASD obtained significantly higher scores than those without, on both the ADI-R and ADOS.
Frazier, Youngstrom, Kubu, Sinclair, & Rezai (2008)
Age Range: 2-46 years
Sample Size: 1,170
Topics Addressed:
Factor structure of ADI-R algorithm items
Outcome:Frazier, Youngstrom, Kubu, Sinclair, & Rezai (2008)
Conclusions: A two-factor structure was supported, with stereotyped language and restricted/repetitive/stereotyped behavior loading on one factor and impairments in social interaction and communication loading together on a second factor. Factor structures fit equally well in all age groups, supporting its usefulness across age ranges.
Moss, Magiati, Charman, & Howlin (2008)
Age Range: Time 1: 2.3–4.5 years | Time 2: 9.1–12.1 years
Sample Size: 35
Topics Addressed:
Test-retest reliability
Outcome:Moss, Magiati, Charman, & Howlin (2008)
Most children (80%) continued to meet ADI-R algorithm cut-off for autism at follow-up, although significant decreases in ADI-R domain and item scores were also found.
Conclusion: while classification of children according to ADI-R criteria generally remained stable between preschool and elementary school age, many children demonstrated significant t improvements in symptom severity.
Cicchetti, Lord, Koenig, Klin, & Volkmar (2008)
Age Range: 3.5 years
Sample Size: 1
Topics Addressed:
Interrater reliability
Outcome:Cicchetti, Lord, Koenig, Klin, & Volkmar (2008)
94-96% with weighted kappas between 0.80 and 0.88.
Conclusion: reliability of the ADI-R can be meaningfully assessed, both in terms of statistical and clinical meaningfulness, when multiple clinical examiners evaluate a single case.
Fusar et al. (2017)
Age Range: 18 years or older
Sample Size: 113
Topics Addressed:
Sensitivity and specificity of the ADOS-2 and ADI-R in diagnosing adult ASD
Outcome:Fusar et al. (2017)
Findings regarding the ADOS-2 are consistent with previous studies of discriminant validity of ADOS-2 Module 4 in samples of adults with average or above average intelligence, which together cautiously suggest it could be a reliable instrument for first evaluations of adults. Agreement between ADOS-2 and ADI-R scores was fair, but agreement between ADI-R and clinical diagnosis was poor, correctly classifying into the spectrum only 55% of the sample. ADI-R presented good specificity but lacked sensitivity.
Conclusion: The ADOS-2 Module 4 accurately classifies adults with ASD who do not have comorbid intellectual disability; however, the ADI-R may not capture essential autism features in older people. This may be, in part, related to concerns about retrospective reporting for developmental profile required within the ADI-R. Overall, authors emphasize the importance of clinician training and multifaceted data collection when assessing adults on the spectrum.