CHAPTER III (continued)

Screening tests for autism are often used if there is a heightened concern about possible autism, or as part of a general assessment of a child's development. Screening for autism is often done prior to a more specific in-depth assessment for autism. Screening may be used as a part of the assessment process as a method intended to lead to a yes/no decision that autism is either unlikely or is possible and requires further evaluation. Most screening tests are designed to be brief and easy to administer.

Many simple screening tests for autism are available to the public through a variety of sources and claim to be useful in identifying children with autism. Most of these tests have not been evaluated using standard research methods. The literature review found only one screening test for autism in young children, the Checklist for Autism in Toddlers (CHAT), that had been evaluated in studies meeting criteria for adequate evidence about efficacy.

The Checklist for Autism in Toddlers (CHAT) is a brief screening instrument that is intended to detect possible autism in toddlers. Since it is a screening test, the CHAT provides a first level of evaluation leading to a yes/no decision that, at the current time, autism is either unlikely or is possible (and requires further evaluation). The CHAT was published in 1992 (Baron-Cohen, et. al., 1992). The CHAT takes only about five to ten minutes to administer and score. Specific training is not required, and it can be administered by a variety of individuals. The CHAT is designed to be used with toddlers as young as 18 months of age.

The CHAT consists of nine yes/no questions to be answered by the child's parent. These questions ask if the child exhibits specific behaviors, including: social play, social interest in other children, pretend play, joint attention, pointing to ask for something, pointing to indicate interest in something, rough and tumble play, motor development, and functional play. The CHAT also includes observations of five brief interactions between the child and the examiner, which enable the clinician to compare the child's actual behavior with the parental reports.

Evidence Ratings: [A] = Strong [B] = Moderate [C] = Limited [D1] = Opinion/No evidence meeting criteria [D2] = Literature not reviewed

Recommendations

1. It is important to identify children with autism as early as possible. The CHAT is a useful first-level screening method for children from 18 to 36 months of age in which there is any level of concern about possible autism. [A]
2. If screening using the CHAT suggests possible autism, further assessment is needed to determine a diagnosis. [A]

3. If screening using the CHAT suggests autism is unlikely, it is still important to:

• assess the child for other developmental or medical problems that may have caused the initial concern
• continue regular periodic surveillance for problems that may be related to the cause of the initial concern [A]

4. It is important to remember that not all children with autism can be identified early. Because the time of onset and severity of symptoms vary, it is recommended that screening be repeated at various age levels when concerns for autism persist. [D1]

Several standardized tests and checklists have been developed to help assess the behavior of children with possible autism. These tests are also intended to further evaluate children in whom autism is considered possible (due to parent concerns, clinical clues, and/or positive screening test results).

These assessment instruments can be used in various ways in assessing children with possible autism. Sometimes these instruments can be used to determine if autism is likely, so that a decision can be made to seek a specific diagnosis. At other times, some of these assessment instruments may be used as part of the formal diagnostic process. Finally, in certain instances some of these instruments may be used to rate the severity of symptoms, which may be useful in assessing interventions, periodic monitoring of the child's progress, and assessing outcomes.

This section is divided into the following six parts:

• Autism Behavior Checklist (ABC): a behavior checklist completed by a parent.
• Autism Diagnostic Interview-Revised (ADI-R): a structured interview.
• Childhood Autism Rating Scale (CARS): a test combining parent reports and direct observation by the professional.
• Pre-Linguistic Autism Diagnostic Observation Schedule (PL-ADOS), a test using direct observation of the child's behavior as elicited by the examiner.
• Reviews of other evidence about behavior rating scales.
• Reviews of other evidence about structured interviews.

The first four tests reviewed in this section (ABC, ADI-R, CARS, and PL-ADOS) are individual autism assessment instruments (or tests) that have been specifically designed to assess children with possible autism and which are currently available for use by U.S. clinicians. All of these tests rely on either historical information about the child's behavior (usually provided by a parent), direct observation of the child by a professional, or a combination of these methods.

Tests that rely on historical information may be in the form of behavior checklists or structured interviews. Behavior checklists (such as the ABC) are lists of questions to be completed by parents and later scored by a professional. Structured interviews (such as the various versions of the ADI) are composed of a prescribed set of questions and interview protocol that the professional uses to question the parents.

Tests that rely on direct observation of the child by a professional (such as the PL-ADOS) often prescribe specific ways for the examiner to elicit responses from the child. These tests also have a standardized method for scoring the observed behaviors. The CARS was the only autism assessment instrument reviewed that combines both historical information from a parent and direct observation of the child by the professional. The CARS also provides a total score that can rate the severity of behavior.

Tests reviewed that are not readily available in the US

This section also includes additional evidence about the efficacy of two autism assessment instruments that are not readily available to U.S. clinicians. The instruments evaluated include several versions of the Behavioral Summarized Evaluation (BSE), which is a behavior rating scale, and the Parent Interview for Autism (PIA), which is a structured interview. Those studies which used the BSE (and met criteria for adequate evidence about efficacy) were done in France or Italy. The BSE is not commonly used in the U.S. as an assessment instrument. Although the PIA was developed in the U.S., it is described only in a few research studies, and is apparently not currently widely available to clinicians.

The panel considered the evidence regarding the PIA and BSE to be useful for supporting some general recommendations about the use of autism assessment instruments. However, since the BSE and PIA are not readily available in the U.S. at this time, the panel chose not to make any specific recommendation regarding these specific testing instruments.

The Autism Behavior Checklist (ABC) is a list of questions about a child's behaviors. The ABC was published in 1980 (Krug et al., 1980) and is part of a broader tool, the Autism Screening Instrument for Educational Planning (ASIEP) (Krug et al., 1978). The ABC is designed to be completed independently by a parent or a teacher familiar with the child who then returns it to a trained professional for scoring and interpretation. Although it is primarily designed to identify children with autism within a population of school-age children with severe disabilities, the ABC has been used with children as young as 3 years of age.

The ABC has 57 questions divided into five categories: (1) sensory, (2) relating, (3) body and object use, (4) language, and (5) social and self-help.

Evidence Ratings: [A] = Strong [B] = Moderate [C] = Limited [D1] = Opinion/No evidence meeting criteria [D2] = Literature not reviewed

Recommendations

1. The ABC appears to have limited usefulness in identifying children with autism who are under the age of 3. [A]
2. When used in conjunction with other diagnostic instruments and methods, the ABC may have some usefulness as a symptom inventory to be completed by parents or teachers. Clinicians could utilize this inventory in structuring their evaluation. [D1]

The Autism Diagnostic Interview-Revised (ADI-R) is a semi-structured interview for a clinician to use with the child's parent or principal caregiver. The original version of this test, the Autism Diagnostic Interview (ADI) was published in 1989 (LeCouteur et al., 1989) and was correlated to the ICD-10 definition of autism. The original ADI was intended primarily for research purposes, providing behavioral assessment for subjects with a chronological age of at least five years and a mental age of at least two years. The ADI-R (Lord et al., 1994) is a shorter version of the ADI, which has been developed for clinical use. It is intended to be briefer and more appropriate for younger children than the ADI. The ADI-R takes from 11/2 to 2 hours to administer and can be used with children as young as two years of age (with a mental age greater than 18 months). The ADI and the ADI-R focus on getting maximal information from the parent about the three key areas defining autism: (1) reciprocal social interaction; (2) communication and language; and (3) repetitive, stereotyped behaviors.

Recommendations

1. The ADI-R may be useful as part of a multidisciplinary intake assessment in diagnosing young children with possible autism. [C]
2. Because of the time needed to administer the ADI-R, and the extensive training needed, this test may not be a practical assessment method in all clinical situations. [D1]
3. A structured parent interview, such as the ADI-R, is a method that can help maximize parental recall but is not a substitute for direct observation of the child by a professional assessing the child. Therefore, it is important to supplement structured parent interviews with direct observation of the child. [D1]
   

The Childhood Autism Rating Scale (CARS) is the most widely used standardized instrument specifically designed to aid in the diagnosis of autism for use with children as young as 2 years of age. Published in 1980 (Schopler et al., 1980), the CARS was originally correlated to the DSM-III and then to the DSM-III-R. The CARS is intended to be a direct observational tool used by a trained clinician. It takes about 20-30 minutes to administer.

The 15 items of the CARS include: Relationships with People, Imitation, Affect, Use of Body, Relation to Non-human Objects, Adaptation to Environmental Change, Visual Responsiveness, Auditory Responsiveness, Near Receptor Responsiveness, Anxiety Reaction, Verbal Communication, Nonverbal Communication, Activity Level, Intellectual Functioning, and the clinician's general impression.

Evidence Ratings: [A] = Strong [B] = Moderate [C] = Limited [D1] = Opinion/No evidence meeting criteria [D2] = Literature not reviewed

Recommendations

1. The CARS may be useful as part of the assessment of children with possible autism in a variety of settings: early intervention programs, preschool developmental programs, and developmental diagnostic centers. [A]
2. Among the autism assessment instruments reviewed, the CARS appears to possess an acceptable combination of practicality and research support, despite the limited research on its use in children under 3 years of age. [A]
   
3. Because it gives a symptom severity rating, the CARS may be useful for periodic monitoring of children with autism and for assessing long-term outcomes. [D1]
4. It is very important that professionals using the CARS have experience in assessing children with autism and have adequate training in administering and interpreting the CARS. [D1]
   
5. An autism assessment instrument that is practical, is supported by research, and includes a severity rating (such as the CARS) may be useful for collecting consistent information to assist with estimating the prevalence of autism and assess functional outcomes (especially if tied to other information about interventions and service delivery). [D1]
   

Pre-Linguistic Autism Diagnostic Observation Schedule (PL-ADOS)

The Autism Diagnostic Observation Schedule (ADOS; Lord et al., 1989) consists of eight tasks, four focusing on social behaviors and four on communicative behaviors. The test was intended primarily for older, higher-functioning, verbal autistic children. The Pre-Linguistic Autism Diagnostic Observation Schedule (PL-ADOS; DiLavore et al., 1995) is a version of the ADOS modified to diagnose young children (under the age of 6 years) who are not yet using phrase speech. It is a semi-structured assessment of play, interaction, and social communication and takes about 30 minutes for a trained clinician to administer.

Recommendations

1. The PL-ADOS may be useful as part of a multidisciplinary intake assessment in diagnosing young children with possible autism. [C]
2. Since extensive training is needed to learn how to administer the PL-ADOS, it may not be a practical assessment method in certain clinical situations. [D1]
   

Because the BSE may not be readily available to U.S. clinicians, no specific recommendations on the use of this test were made. However, evidence from scientific studies about the use of various versions of the BSE for identifying young children with autism is used as the basis for supporting some general recommendations about the use of autism assessment instruments that provide a symptom severity rating.

The Behavioral Summarized Evaluation (BSE) (Barthelemy et al., 1992) is a rating scale developed in France and designed to measure changes in behavior in autistic children and adolescents. The BSE is intended to be completed by someone having daily contact with the child. Its primary use has been as a weekly or biweekly measure of clinical status of autistic children participating in various intervention studies. However, the original version of the BSE, a revised version (BSE-R) (Barthelemy et al., 1997) and a downward extension of the BSE known as the Infant Behavioral Summarized Evaluation (IBSE) (Adrien et al., 1992) have all been analyzed for their ability to identify autism in children.

Evidence Ratings: [A] = Strong [B] = Moderate [C] = Limited [D1] = Opinion/No evidence meeting criteria [D2] = Literature not reviewed

Recommendations

1. An autism assessment instrument that provides a symptom severity rating may be useful for periodic monitoring of children with autism and for assessing outcomes. [D1]
2. A behavior rating scale completed by someone familiar with the child (such as a parent, early childhood professional or teacher), when interpreted by a professional with expertise in assessing children with autism, may provide information that is useful in helping to identify and diagnose children with autism. [A]
   

The Parent Interview for Autism (PIA) is an instrument that was used in a research study and might not be readily available to practicing clinicians. Therefore, the panel chose not to make specific recommendations on using the PIA, but reviewed evidence from the study to make general recommendations about the use of structured parent interviews for young children with possible autism.

The Parent Interview for Autism (PIA; Stone and Hogan, 1993) is intended to elicit relevant information about children functioning at the preschool level and below. The PIA is a respondent-based interview that requires parents to make judgments about the frequency of occurrence of specific behaviors. PIA items are administered verbally, so that parents' questions can be clarified. Administration of the PIA takes approximately 30-45 minutes.

Recommendations