Guidelines for Speech-Language Pathologists in Diagnosis, Assessment, and Treatment of Autism Spectrum Disorders Across the Life Span

The Important Role of the SLP

Given the importance of social communication in the diagnosis of ASD, the SLP can play an important role in both screening and diagnosis. The Child Neurology Society and American Academy of Neurology formed a multidisciplinary consensus panel to determine practice parameters for screening and diagnosis of ASD (Filipek et al., 1999). The panel included representatives from the disciplines of psychiatry, neurology, pediatrics, psychology, speech-language pathology, audiology, and occupational therapy, as well as from autism advocacy organizations, with liaisons from the National Institutes of Health. The panel concluded that all professionals involved in early health care, including SLPs, need to be able to recognize the symptoms of ASD and use autism-screening tools to make decisions about appropriate referrals for further evaluations (Filipek et al., 1999). This panel emphasized the importance of interdisciplinary collaboration in assessing and diagnosing ASD, due to the complexity of these disorders, the varied aspects of functioning affected, and the need to distinguish it from other disorders or medical conditions. Further, Filipek et al. stressed that professionals involved in diagnosis of ASD must be knowledgeable and experienced in using guides such as the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV; American Psychiatric Association, 1994, 2000) along with results of various diagnostic assessment tools to make clinical judgments about these types of disorders. A comprehensive interdisciplinary assessment is important not only for diagnosis but also for intervention planning. Thus, ideally the diagnostic role of the SLP would be as a key member of an interdisciplinary team with the appropriate individual and collective expertise in ASD. In some cases, however, there may be no appropriate team accessible to a family. As Filipek et al. (1999) stated, “Language pathologists are independent health care providers who have responsibilities at the levels of screening (Level 1), diagnosis and evaluation (Level 2) of autism” (p. 461). The SLP who has been trained in the clinical criteria for ASD, as well as in the use of reliable and valid diagnostic and assessment tools for individuals with ASD, and who is experienced in diagnosis of developmental disorders, may be qualified to diagnose these disorders as an independent professional. It would be incumbent upon the SLP to ensure that diagnostic findings are interpreted in relation to measures of nonverbal developmental level or IQ and to make appropriate referrals to other professionals to obtain a thorough assessment of the individual's needs and determination of any comorbid diagnoses that lie outside the expertise of the SLP. The SLP plays a critical role in referring children suspected of possible ASD to an audiologist to confirm or rule out a hearing loss. Furthermore, the SLP should play a primary role in the diagnosis of speech and language impairments that can co-occur with ASD, including, but not limited to, features of specific language impairment, apraxia, and dysarthria. The SLP should be aware that some state laws or regulations may restrict the scope of practice of licensees, however, and prohibit the SLP from providing such diagnoses.

Designation of a student with a disability within school settings must be made within the team decision-making process as designated by the Individuals with Disabilities Education Improvement Act of 2004 (IDEA, 2004). Within a public school setting, eligibility for services under the disability category of autism is based on the definition provided in the IDEA, as provided below:

Autism means a developmental disability significantly affecting verbal and nonverbal communication and social interaction, generally evident before age 3, which adversely affects a child's educational performance. Other characteristics often associated with autism are engagement in repetitive activities and stereotyped movements, resistance to environmental change or change in daily routines, and unusual responses to sensory experiences. The term does not apply if a child's educational performance is adversely affected primarily because the child has an emotional disturbance as defined by IDEA criterion.

A child who manifests the characteristics of “autism” after age 3 could be diagnosed as having “autism” if the criteria in the preceding paragraph are met. (34 C.F.R. § 300.7 [c] [1])

Individuals diagnosed with an ASD or pervasive developmental disorder by means of other sources of clinical criteria, such as the DSM-IV-TR (American Psychiatric Association, 2000), are likely to be eligible for special education services under the category of autism as defined above, due to the common challenges and deficits in social communication functioning across the various disorders on the autism spectrum.

Return to Top

---

Screening for ASD

Early indicators of ASD are observable by age 12 months (Baranek, 1999; Osterling & Dawson, 1994, 1999; Wetherby et al., 2004; Zwaigenbaum et al., 2005), and ASD can be reliably diagnosed as early as age 24 months by experienced and knowledgeable diagnosticians (Filipek et al., 1999). The main characteristics that differentiate ASD from other developmental disorders in young children include difficulties in eye gaze, orienting to one's name, pointing to or showing objects of interest, pretend play, imitation, nonverbal communication, and language development. It is important to consider cultural diversity of social communication when examining these areas. In addition, loss of language or social skills at any age should be considered grounds for screening (Filipek et al., 1999). Because children with ASD are often initially suspected of having a hearing problem, audiologists are in a critical role to spot possible signs of ASD in children whose hearing they are testing and to make appropriate referrals for screening and diagnosis of ASD.

Screening for ASD may utilize broadband screeners designed to detect developmental delays in the general pediatric population or autism-specific screening tools designed for either the general population or high-risk populations such as children referred to the early intervention system. Any screening tool should have strong psychometric features to support its accuracy in identifying at-risk children who need further evaluation. The following proportions provide particularly important information about the accuracy of a screening tool:

• sensitivity, or true positives—the percentage of children identified as at risk (i.e., receiving a positive screen or evaluation result) who failed the follow-up testing or received a diagnosis of ASD. The percentage of true positives added to the percentage of false negatives (i.e., the proportion of children identified as no risk, by receiving a negative screen or evaluation result, who failed the follow-up testing or received a diagnosis of ASD) equals 100%. Thus, a lower false negative rate means a higher true positive rate.

• specificity, or true negatives—the proportion of children identified as no risk (i.e., receiving a negative screen or evaluation result) who passed the follow-up testing or for whom ASD was ruled out. The percentage of true negatives added with the percentage of false positives (i.e., the proportion of children identified as at risk, by receiving a positive screen or evaluation result, who passed the follow-up testing or for whom ASD was ruled out) equals 100%. Thus, a lower false positive rate means a higher true negative rate.

• positive predictive value—the proportion of children identified as at risk who fail the follow-up testing out of the total number of children identified as at risk.

• negative predictive value—the proportion of children identified as no risk who pass the follow-up testing out of the total number of children identified as no risk.

To be considered psychometrically sound, a screening tool would minimally need to report sensitivity and specificity. Meisels (1989) recommended that cutoffs for both sensitivity and specificity be set at no less than 80% for developmental screening of young children; however, he noted that a “75% sensitivity ratio is considerably less favorable than a 75% specificity proportion” (p. 579). It is also important to consider the positive predictive value; however, this is related to the base rate of a disorder. That is, the higher the prevalence rate of the disorder, the greater the probability that a positive result will be correct and the higher the positive predictive value. In screening a general population for relatively low incidence disorders such as ASD, even an instrument with a sensitivity and specificity of .80 will yield a poor positive predicative value. In other words, the lower the prevalence rate of a disorder, the lower the probability that an individual has the disorder given a positive screen result. As Clark and Harrington (1999) point out, screening tools are intended to identify individuals who may be at risk for a disorder rather than to serve as “gold standards” for determining a diagnosis. Thus, professionals must not assume that failing a screening means that an individual has an ASD and must instead complete thorough assessments before providing a diagnosis.

There is currently very limited research on the accuracy of broadband screeners to identify young children at risk for ASD. The First Signs program (www.firstsigns.org), a national nonprofit organization whose goal is to improve early identification of ASD, has assembled a set of psychometrically sound broadband screeners based on parent report that include the Ages and Stages Questionnaire (Squires, Potter, & Bricker, 1999), the Parents' Evaluation of Developmental Status (Glascoe, 1997), and the Communication and Symbolic Behavior Scales Developmental Profile (CSBSDP) Infant Toddler Checklist (ITC; Wetherby & Prizant, 2002). All three of these broadband screeners have sensitivity and specificity at or near 80% for identifying children with developmental delays from a general pediatric population. The ITC is the only broadband screener that has preliminary validation data showing it has high sensitivity and specificity (both 88.9%) for catching toddlers at risk for ASD and other developmental delays using a prospective sample of more than 3,000 children (Wetherby et al., 2004).

Autism-specific screeners use parent report and/or interactive observational measures. Because of the challenge of identifying very young children with ASD, there is limited research on the accuracy of autism-specific screeners. Some of the tools have been validated on children referred to health care or education agencies because autism was suspected or from children who already have been identified as having developmental delays or disabilities. However, it is important also to validate autism-specific screeners, either as an initial screener or follow-up to a broadband screener, on a general population sample, since this is ultimately how they will be used clinically. This is particularly critical for higher functioning children who may be more easily missed and not be referred into the system. Following are autismspecific screening tools that have some published psychometric information, including sensitivity and specificity.

Checklist for Autism in Toddlers (CHAT; Baird et al., 2000; Baron-Cohen, Allen, & Gillberg, 1992; Baron-Cohen et al., 1996). The CHAT consists of 9 items reported by parents and 5 items observed by a health professional at the 18-month developmental checkup. Baron-Cohen and colleagues screened more than 16,000 children using the CHAT and identified 19 children at 18 months who were later diagnosed with ASD based on failure of the following 3 key items: (a) protodeclarative pointing, (b) pretend play, and (c) gaze monitoring. However, at follow-up at age 7 years, 94 cases of ASD were identified. These findings indicate that the CHAT has a specificity of 98% but a sensitivity of 38% (Baird et al., 2000) and missed many children at 18 months who were later diagnosed with ASD. While the validity of the CHAT is disappointing, it indicates that some children with ASD can be identified at 18 months and provides clues to early indicators of ASD, based on the children they were able to identify early. The poor sensitivity indicates that it should not be relied on as an accurate screener and likely does not merit the time in a pediatric practice.

Modified Checklist for Autism in Toddlers (M-CHAT; Robins, Fein, Barton, & Green, 2001). The M-CHAT is an expanded version of the original CHAT. The M-CHAT has 23 questions using the original 9 items from the CHAT as a basis. The M-CHAT is currently in its final stages of testing and validation and has not yet been validated for general population screening. Preliminary results, however, have suggested that it has improved sensitivity compared with the CHAT. The M-CHAT has demonstrated validity in identifying the majority of children with ASD and developmental delay from age 24 months in a study in which most of the at-risk children had been referred already to early intervention programs due to developmental concerns. Follow-up data for children who were not found to be at risk on the M-CHAT have not been published yet; thus, the extent to which the M-CHAT may miss children who will later be diagnosed with ASD is still unknown. The M-CHAT is available in both Spanish and English.

Social Communication Questionnaire (SCQ; Rutter, Bailey, Lord, & Berument, 2003). The SCQ is a 40-item parent report screening tool for ASD and is a component of the Autism Diagnostic Interview, described below. The SCQ is an autism-specific screening tool; that is, the intended use of the SCQ is for screening individuals previously identified with developmental concerns. The SCQ yields a total score with 0 indicating no risk for ASD and 15 or higher indicating risk for ASD. The SCQ has been validated as a screening tool for ASD in children age 4 and older, as well as adults, and has good sensitivity and specificity. Preliminary research indicates that the specificity and sensitivity of the SCQ are slightly poorer for younger children based on 112 children between age 2 and 5 compared with 181 individuals over age 5 (Corsello, Cook, & Leventhal, 2003). Therefore, caution should be taken in using this with younger children; however, further research is needed on children under 4.

Pervasive Developmental Disorders Screening Test, Second Edition (PDDST-II; Siegel, 2004). Research on the PDDST-II has been presented at several conferences but has not yet appeared in the peer-reviewed literature. The PDDST-II is a clinically derived self-administered parent questionnaire divided into three stages. PDDST-II Stage 1 is intended for use in the primary care setting with items ordered developmentally from birth to age 36 months. The cutoff score for Stage 1 was derived by comparing results for 197 very low birth-weight children with those for 380 children referred due to concerns about possible ASD (some of whom were later diagnosed with ASD, and some of whom did not receive ASD diagnoses). The sensitivity and specificity were 85% and 71%, respectively, in identifying those children in need of further evaluation for possible ASD. The questionnaire has not been validated with a general population sample, however. The PDDST-II Stage 2 screener is intended to aid in differentiating a possible diagnosis of ASD from other disorders such as developmental language delay, mental retardation, or ADHD. In a study involving 260 children with ASD and 120 children with other developmental disorders, different cutoff scores yielded sensitivity ranging from 69% to 88% and specificity ranging from 25% to 63%. The purpose of the PDDST-II Stage 3 is to assist in differentiating children with autistic disorder from those with other pervasive developmental disorders and to provide information on symptom severity. The author is continuing research and development on this instrument, but the preliminary results suggest it has utility for screening and diagnosing young children.

Screening Test for Autism in Two-Year-Olds (STAT; Stone, Coonrod, & Ousley, 2000; Stone, Coonrod, Turner, & Pozdol, 2004; Stone & Ousley, 1997). The STAT is a direct observational scale designed for children from age 24 to 35 months. The purpose of the STAT is to serve as a Level 2 tool to screen for autism among young children referred for developmental concerns. The 20-minute play interaction includes 12 activities that sample 3 main developmental areas: play, motor imitation, and nonverbal communication. Stone et al. (2004) reported a sensitivity of 92% and specificity of 85% for this instrument in identifying children between the ages of 24 and 35 months with autistic disorder. The sensitivity of the instrument in screening for other variants of ASD was lower. As the authors point out, underidentification of children with milder symptoms is an issue for other screening tools and procedures as well. The authors also reported results indicating that the STAT has good interobserver and test-retest reliability, as well as concurrent validity with clinical diagnoses of autistic disorder and results of independent assessment on the Autism Diagnostic Observation Schedule (ADOS; Lord, Rutter, DiLavore, & Risi, 1999; see below).

Systematic Observation of Red Flags (SORF; Wetherby et al., 2004). The SORF consists of an observational rating of 29 potential “red flags” for ASD that can be scored from video of the CSBS DP Behavior Sample (Wetherby et al., 2004). The Behavior Sample uses a standard set of systematic procedures designed to encourage spontaneous behaviors that range in degree of structure provided. Thirteen of the items on the SORF have been shown to discriminate toddlers who later are diagnosed with ASD from toddlers younger than age 2 with typical development and those with developmental delay in which ASD had been ruled out, with an accuracy of 94.4% (Wetherby et al., 2004). This indicates strong sensitivity and specificity on a preliminary sample of 18 children later diagnosed with ASD identified from a general population screen of about 3,000 children. The following red flags distinguished the children with ASD from children who were developmentally delayed in which ASD was ruled out and who were typically developing in the second year of life: (a) lack of appropriate gaze; (b) lack of warm, joyful expressions with gaze; (c) lack of sharing enjoyment or interest; (d) lack of response to name; (e) lack of coordination of gaze, facial expression, gesture, and sound; (f) lack of showing; (g) unusual prosody; (h) repetitive movements or posturing of body, arms, hands, or fingers; and (i) repetitive movements with objects. The following red flags distinguished the children with ASD and developmental delay from the typically developing children: (a) lack of response to contextual cues; (b) lack of pointing; (c) lack of vocalizations with consonants; and (d) lack of playing with a variety of toys conventionally.

Return to Top

---

Diagnosis of ASD

An early accurate diagnosis can assist in earlier access to needed services and appropriate treatment. Diagnosis also can provide a common language across multidisciplinary teams and can lead to some sense of relief for families and caregivers who are provided with a framework within which to understand their child's difficulties. Any diagnosis of ASD, particularly of young children, should be periodically reviewed, as diagnostic categories and conclusions may change as the child develops.

The following information should be gathered in a diagnostic evaluation of children at risk for ASD: (a) review of relevant background information to guide the diagnostic evaluation and the selection of appropriate tools; (b) caregiver interview to gather health, developmental, and behavioral history of the child and medical and mental health history of the family; and (c) direct behavior observation. A medical evaluation should be completed based on the recommendations of the American Academy of Pediatrics on the role of the pediatrician in the diagnosis and management of ASD (American Academy of Pediatrics, 2001a, 2001b). A medical evaluation for a child at risk for ASD should consist of a general physical and neurodevelopmental examination, including evaluations of vision and hearing. In addition, if there is a family history of mental retardation or genetic conditions sometimes associated with ASD (e.g., fragile X, tuberous sclerosis), or if the child exhibits physical features suggestive of a possible genetic syndrome, then a recommendation for genetic testing would be appropriate. If the child exhibits symptoms such as lethargy, cyclic vomiting, pica, or seizures, then selective metabolic testing may be appropriate (Filipek et al., 1999).

A diagnostic evaluation to confirm or rule out a diagnosis of autism or ASD should be performed only by professionals who have specific expertise in the evaluation and treatment of autism (Filipek et al., 1999; NRC, 2001). Diagnostic tools for ASD include parent or caregiver report (i.e., interview or questionnaire) as well as diagnostic observation instruments. Following are diagnostic tools for ASD that have some published psychometric information including evidence of reliability and validity.

The ADOS (Lord et al., 2000) is a semistructured observational assessment in four modules that includes activities designed to evaluate communication, reciprocal social interaction, play, stereotypic behavior, restricted interests, and other abnormal behaviors in individuals with ASD across the age range from preschool to adulthood and covering language skills from nonverbal to conversational. The ADOS consists of four modules developed for individuals with varying levels of linguistic ability; each module includes a set of activities that press for different behaviors that contribute to a diagnosis of autism or ASD. Administration requires 30–45 minutes, thus making the ADOS a feasible diagnostic tool in many clinical settings. As indicated in a recent review of autism diagnostic tools (Lord & Corsello, 2005), research has documented excellent interrater reliability for ADOS total scores (following substantial training and practice) and excellent internal consistency for the two major domains of social communication behaviors and restricted-repetitive behaviors. The ADOS scores also have a high degree of accuracy when compared with expert clinical diagnoses using DSM-IV criteria. Lord et al. (2000) reported that the sensitivity of the four modules ranged from 90% to 97%, and the specificity ranged from 87% to 93% in identifying individuals with ASD versus individuals with other DD who were outside the autism spectrum. The ADOS does not perform particularly well in discriminating between individuals diagnosed with autistic disorder versus pervasive developmental disorders not otherwise specified (PDD-NOS), however. Lord and Corsello (2005) note that ADOS scores slightly overinclude young children with mental retardation in the autism spectrum and slightly underinclude older children and adults with milder impairments and relatively high verbal skills.

The Childhood Autism Rating Scale (CARS; Schopler, Reichler, & Renner, 1988) is a 15-item structured interview and observation instrument that is suitable for use with children above age 24 months only. Each of the 15 items uses a 7-point rating scale to indicate the degree to which the child's behavior deviates from an age-appropriate norm. It also distinguishes between mild-to-moderate and severe presentations of ASD. The CARS is recognized widely and used as a reliable instrument for the diagnosis of ASD (e.g., Morgan, 1988; Sevin, Matson, Coe, Fee, & Sevin, 1991). The examiner rates the child based on observation of behaviors exhibited during other assessment activities. The CARS takes approximately 15 minutes to complete but should be based on opportunities to observe the child across at least a 1-hour time period and across a variety of types of activities and interactions. It can be used reliably with relatively limited training and is available in a number of languages other than English. In a recent review of diagnostic instruments for autism, Lord and Corsello (2005) conclude that the evidence across several studies indicates the CARS may overidentify children as falling into the autism spectrum, particularly when the children have low verbal skills and/or low cognitive levels. For this reason, they do not recommend the exclusive use of the CARS in identifying or classifying participants for research purposes. Although the CARS was originally developed to correspond to the DSM, Third Edition (DSM-III; American Psychiatric Association, 1980), a recent study reported 100% agreement between clinical diagnoses of autistic disorder (vs. other disorders including Asperger's and PDD-NOS) using DSM-IV criteria and a classification that was based on the CARS cutoff score (Rellini, Tortolani, Trillo, Carbone, & Montecchi, 2004), suggesting that the CARS continues to have clinical utility in identifying children with autistic disorder, but not for diagnosing children with other ASD.

The Autism Diagnostic Interview—Revised (ADI-R; Rutter, LeCouteur, & Lord, 2003) is a comprehensive structured parent interview that probes for symptoms of ASD in the areas of social relatedness, communication, and ritualistic or perseverative behaviors. With this tool, information about the developmental history of the individual is combined with information about current behaviors and used in a scoring algorithm to indicate whether the person meets the DSM-IV criteria for autism or a related disorder. Administration takes approximately 2 to 3 hours and requires specific training and validation procedures. Due to the length of the ADI-R, it is used relatively infrequently in purely clinical settings. In reviewing research on the instrument, Lord and Corsello (2005) noted that with training, high interrater reliabilities have been achieved for each of the three subscales of the ADI-R (social reciprocity, communication, and restricted-repetitive behaviors). Internal consistency is also excellent within the three subscale domains. The tool also has performed extremely well in differentiating individuals with autism and those with other DD. Like some of the other autism diagnostic instruments, the ADI-R tends to slightly overinclude individuals with mental ages less than or equal to 18 months. The evidence regarding individuals who are high-functioning and verbal has been inconsistent with respect to whether the tool is under- or overinclusive.

The Parent Interview for Autism (PIA; Stone & Hogan, 1993) is a structured interview designed to gather relevant diagnostic and symptom severity information from parents of young children (under age 6) suspected of having ASD. The PIA targets 11 relevant areas, including social behavior, communication, repetitive activities, and sensory behaviors. Internal consistency and test-retest reliability are adequate, and concurrent validity with the DSM-III, Revised (American Psychiatric Association, 1987) and CARS (Schopler et al., 1988) was demonstrated; however, concurrent validity examining clinical diagnoses or tools using current diagnostic criteria has not been published. More recently, Stone and colleagues (Stone, Coonrod, Pozdol, & Turner, 2003) have reported similar psychometric properties for a shorter clinical version of the PIA used with a sample of children under the age of 3.

The Gilliam Autism Rating Scale (GARS; Gilliam, 1995) is a checklist designed to be used by parents, teachers, and professionals to help identify and estimate the severity of symptoms of ASD, but it is only normed from age 3. There is an optional subtest, however, that describes development in the first 3 years of life. The tool provides a global rating of ASD symptomatology using 56 items that are grouped into four subtests. The GARS has been standardized such that the total score yields an “autism quotient” with a mean of 100 and a standard deviation of 10. In the author's published research, quotients above 90 were associated with a high risk of autism, scores in the 70 to 90 range were associated with a moderate risk, and scores below 70 with a limited risk for autism. Despite strong psychometric properties reported in the initial research on the GARS, a recent independent study found that the GARS had a sensitivity of only 48% in identifying children with autism who had been diagnosed using DSM-IV criteria, with diagnoses verified using the ADOS and the ADI-R (South et al., 2002). As noted above, underdiagnosis of the disorder is of considerable concern in clinical settings. South and colleagues noted that a revision of the GARS was in progress at the time of their report.

A clinical diagnosis of one of the subcategories of PDD can be made using the DSM-IV criteria based on information gathered from the diagnostic evaluation along with one or more ASD diagnostic tools. The current “gold standard” measures for the diagnosis of ASD in research protocols are the ADI-R and the ADOS (Filipek et al., 1999; Lord & Corsello, 2005) because of their strong psychometric features. It is important for the team working with the child to continue to provide information and support to parents after a diagnosis has been made. The goal of this support is to help families understand the nature of the disorder, where their child falls on the spectrum of ASD, how to access appropriate support and services, and how to choose between the options available. Ongoing assessment also is important to monitor intervention effects on the changing needs of the child and possible changes in diagnosis.

Return to Top

---

Similar Diagnostic Categories and Differential Diagnosis

Although there is general conceptual agreement regarding the core deficits of ASD, the variability seen in this spectrum of disorders presents a number of challenges to reliable diagnoses of ASD and other disorders that may overlap or be confused with ASD. For example, different diagnostic systems tend to converge in specifying criteria for the diagnosis of autistic disorder, but as Volkmar, Lord, Bailey, Schultz, and Klin (2004) pointed out, there are multiple widely circulated definitions of Asperger's disorder. This situation undoubtedly contributes to divergent findings in research, stemming from the varying diagnostic criteria used to select participants. Of particular relevance to the SLP, a primary difference in criteria for Asperger's disorder versus autistic disorder in the DSM-IV is the stipulation that individuals diagnosed with Asperger's disorder will show no clinically significant delay in language development (defined as having single words by age 2 and phrase speech by age 3). But as Filipek et al. (1999) pointed out, the language in Asperger's disorder “is clearly not typical or normal” (p. 447). Mayes, Calhoun, and Crites (2001) found that after grouping children with Asperger's disorder and autistic disorder who had normal intelligence based on a positive or negative history of early language delay, the groups were not different from each other on any of 71 variables investigated. Thus, they questioned the meaningfulness of this criterion for differential diagnosis. More recently, Howlin (2003) has reported similar findings and further noted that both groups exhibited low language scores as adults, calling into question the assumption that language development in Asperger's disorder (aside from pragmatic development) is essentially normal.

Despite existing criteria for differential diagnoses of Asperger's disorder and autistic disorder, Macintosh and Dissanayake (2004) concluded from their literature review that there is insufficient evidence that Asperger's disorder is a syndrome distinct in meaningful ways from high-functioning autism. Some researchers have reported differing cognitive profiles between individuals with Asperger's disorder and those with high-functioning autism. Asperger's disorder is more often associated with a higher verbal IQ than performance IQ, contrasting with the opposite pattern in high-functioning autism (e.g., Klin, Volkmar, Sparrow, Cicchetti, & Rourke, 1995; Volkmar et al., 2004). Thus, in these studies, individuals with Asperger's disorder reportedly demonstrated strengths in verbal language and verbal memory despite relative weaknesses in visually based tasks involving nonverbal concept formulation, the perception of visuospatial information, and memory of visual images. The converse was reported in individuals with high-functioning autism. Findings have been inconsistent, however. In two studies using DSM-IV criteria for differential diagnosis of Asperger's disorder and high-functioning autism, researchers found that as a group, the individuals with Asperger's disorder had higher verbal IQs and full-scale IQs than the individuals with high-functioning autism (Ghaziuddin & Mountain-Kimchi, 2004; J. N. Miller & Ozonoff, 2000). At the individual level, however, varying cognitive profiles were found in each group, and findings of normal language development histories, motor delays, and high verbal relative to performance IQs were not specific to the groups with Asperger's disorder.

The question of whether individuals with Asperger's disorder represent a distinct subgroup from individuals of normal intelligence with autistic disorder continues to generate considerable interest and controversy. One possibility is that the current diagnostic criteria do not serve to reliably discriminate between the two groups and that improvements in identifying and operationalizing the key criteria will yield meaningful subgroups (Klin, McPartland, & Volkmar, 2005), whereas another possibility is that individuals currently diagnosed with Asperger's disorder represent the highest IQ end of a continuous spectrum that cannot be subdivided meaningfully (J. N. Miller & Ozonoff, 2000). The existing research serves to inform us clinically of the varying neuropsychological and developmental profiles existing among persons with ASD, and it reminds us of the importance of careful individualized assessments for the purposes of intervention planning.

Rett's disorder offers a contrasting situation (Jellinger, 2003; A. M. Kerr, 2002; A. M. Kerr, Belichenko, Woodcock, & Woodcock, 2001). The disorder is observed overwhelmingly in females, in contrast to autism, Asperger's disorder, childhood disintegrative disorder, and PDD-NOS, all of which occur at higher rates in males. Rett's disorder involves regression in social, cognitive, and psychomotor development after an early period of development that appears normal to caregivers (although retrospective analyses of infant videotapes suggest there may be subtle indicators of the disorder during the first year, according to A. M. Kerr, 2002). The social regression along with the appearance of repetitive behaviors increases the chance of a misdiagnosis of autism early on; across time, however, social interest and interactions develop that are consistent with the overall developmental level of the children (although this is frequently quite low). Identification of a common genetic anomaly in the vast majority of children meeting the behavioral criteria for diagnoses of Rett's disorder in 1999 has led to opportunities for more refined study of the developmental course of children with a known etiology for Rett's disorder. Although most of the diagnosed individuals present with severe developmental delays across all areas, research since 1999 has demonstrated a broad range of developmental outcomes among girls with the characteristic genetic anomaly, with some individuals developing functional single word or phrase speech, and occasional individuals functioning within the normal range (A. M. Kerr, 2002). The degree of impairment appears to be related to the number of cells in which the mutated gene is active.

A number of diagnostic labels overlap with ASD to some extent, with the overlap and diagnostic uncertainty being the greatest for spectrum disorders other than autistic disorder itself (Volkmar et al., 2004). These diagnostic labels include semantic-pragmatic disorder (Bishop & Rosenberg, 1987; Rapin & Allen, 1983) or pragmatic language impairment (Bishop, 1998), nonverbal learning disorder (NLD; Volden, 2004), hyperlexia (Grigorenko, Klin, & Volkmar, 2003), sensory integration disorder/dysfunction or sensory processing disorder (L. J. Miller, Cermak, Lane, Anzalone, & Koomar, 2004), and multiple complex developmental disorders (Buitelaar & van der Gaag, 1998; D. J. Cohen, Paul, & Volkmar, 1986). Widespread consensus does not exist on the existence of, or criteria for, these disorders as representing distinct diagnoses, which makes interpreting the available literature more challenging. Fitzgerald (1999) argues that several of these categories represent unnecessary “diagnostic splitting” resulting to some extent from a lack of communication among different professional groups, such as SLPs and psychiatrists. Further, Fitzgerald suggests that families often will be better served by receiving a diagnosis of ASD, at least in those situations in which the individual meets criteria for one of the disorders within the autism spectrum, in order to point to the need for and access to recognized services.

Botting and Conti-Ramsden (1999) maintain that at least in Great Britain, the diagnosis of semanticpragmatic disorder (or more recently, pragmatic language impairment) is used for children who do not meet the criteria for a diagnosis of autism; however, they acknowledge an ongoing debate about whether these children are viewed more appropriately as representing a subgroup of children with language impairments or a subgroup of children with PDD. At least 6 of the 10 children with pragmatic language impairment included in their study did not meet diagnostic criteria for Asperger's disorder or autistic disorder. Thus, the only possibly appropriate ASD diagnosis for these children would be PDD-NOS, and the investigators questioned the benefit of that diagnosis with respect to planning intervention or providing families with access to services.

Nonverbal learning disorder or disability (Volden, 2004) is characterized by deficits in such areas as arithmetic, tactile and visuospatial perception, and motor coordination. Individuals with this disorder also have been described as having good rote verbal memory and problems with social pragmatic skills (Myklebust, 1975; Rourke, 1989). As Volden points out, the neuropsychological profile of individuals with NLD bears striking similarities to that reported for many individuals with Asperger's disorder. There has been insufficient research to resolve questions about overlap between these diagnoses, but it is likely at least some individuals with diagnoses of NLD would be included appropriately on the autism spectrum.

Hyperlexia is another diagnostic term that has been applied with varying criteria (Grigorenko et al., 2003) to refer to children with precocious printed language decoding abilities. In some definitions, hyperlexia is identified based on a discrepancy between print recognition and print comprehension abilities. Other proposals have identified hyperlexia based on precocious print recognition in the context of significant language and/or cognitive impairments, and still others have combined the criteria and specified that hyperlexia involves a discrepancy between superior print recognition and both language or cognitive level and print comprehension abilities. Hyperlexia has been reported among children with DD other than ASD, but particularly when the criteria used for hyperlexia include the development of precocious decoding skills relative to cognitive level, research demonstrates a high likelihood that children with hyperlexia will meet criteria for an ASD diagnosis (Grigorenko et al., 2002).

Sensory integration disorder or dysfunction and sensory processing disorder are diagnostic labels stemming from Ayres's (1975) work related to children with learning disabilities. Symptoms of sensory integration disorder include both oversensitivity and underreactivity to various stimuli, distractibility, social emotional problems, physical clumsiness, hyper- or hypoactivity, impulsiveness, and other difficulties in self-regulation of arousal level, and concomitant delays in speech, language, and motor skills and academic achievement (Anzalone & Williamson, 2000). Limited peer-reviewed research is available regarding differential diagnosis or characterization of the population of individuals with sensory integration dysfunction, but clinically it is common to encounter children with diagnosed or suspected ASD who also have been given a diagnosis of sensory integration disorder. Although few of the symptoms of sensory integration disorder appear among the diagnostic criteria for ASD, these symptoms nevertheless are reported widely among children with ASD (Baranek, David, Poe, Stone, & Watson, in press; Piek & Dyck, 2004). More detail about sensory and feeding issues for individuals with ASD with supporting references is provided in the companion technical report (ASHA, 2006b).

Multiple complex developmental disorders (Buitelaar & van der Gaag, 1998; D. J. Cohen et al., 1986) grew out of studies of children diagnosed with PDD-NOS and other severe developmental disorders who did not meet the criteria for autistic disorder. Cluster analyses “identified a group of children characterized by social problems, bizarre and disorganized thinking, recurrent anxieties, inappropriate affect, and mood lability” (Buitelaar & van der Gaag, 1998, p. 912). Researchers have proposed that this is a distinct diagnostic subgroup that represents not a milder variant of autism, but rather a different group with different distinguishing diagnostic features. Compared with children with autism, these children show more symptoms of aggression, more anxiety, and more psychotic thoughts and suspiciousness, whereas children with autism show more disturbances in social interaction, communication, and stereotyped and rigid behaviors (van der Gaag et al., 1995). Unlike children with autism, children meeting the criteria for multiple complex developmental disorders are at risk for developing schizophrenia in adulthood; they also show different biologically based responses to psychosocial stress, suggesting that this subgroup may have a different biological etiology than children with autism (Jansen, Gispende Wied, van der Gaag, & van Engeland, 2003).

Return to Top

---

Determining Eligibility and the Challenges of Identifying Higher Functioning Individuals

High-functioning individuals with ASD pose particular challenges both for identification and for determining eligibility for services. By definition, individuals with high-functioning autism or Asperger's disorder have either verbal or nonverbal measured intelligence within normal limits. Many are not diagnosed until later school age, adolescence, or even adulthood (Howlin & Asgharian, 1999). One reason for such late diagnosis is that they often appear to succeed in some or most academic subjects, particularly in the early school years. This often masks the significant challenges faced by these students and thus may delay a referral for special education services. Long-term outcome research for individuals with high-functioning autism or Asperger's disorder, however, has shown that social communication deficits significantly affect their ability to adjust to new social demands in later academic and community settings and, therefore, achieve vocational goals (Gilchrist et al., 2001; Mueller, Schuler, Burton, & Yates, 2003; Tsatsanis, Foley, & Donehower, 2004). These findings suggest that it is important to provide early intervention to address the gap between cognitive potential and social adaptive functioning.

Although current research outcomes support the provision of individualized educational programming, SLPs often find it challenging to demonstrate that a child or older individual with high-functioning autism or Asperger's disorder is eligible for services. The NRC has recognized this challenge and has recommended that all children with ASD, including autistic disorder, Asperger's disorder, and PDD-NOS, be deemed eligible for special education services under the category autistic spectrum disorders, as opposed to other educational categories often used by school systems (e.g., other health-impaired, social emotional disorder; NRC, 2001, p. 213). Nevertheless, it may be necessary to conduct a comprehensive evaluation to demonstrate the expected gap between an individual's cognitive potential and his or her social adaptive functioning. Furthermore, school systems need to make eligibility decisions through the team process of the individualized educational program. The team should decide the optimal eligibility decision for the child. Formal testing may be useful for assessing the structure and form of language, whereas these evaluation tools may not provide an accurate assessment of an individual's use of language (i.e., pragmatics). Determining an individual's social and communication competence, therefore, necessitates evaluation across a range of social settings and not just one-on-one structured formal testing sessions. A variety of strategies should be used for gathering information. These may include direct standardized assessments, naturalistic observation across contexts, and caregiver/teacher interviews or questionnaires. Standardized assessment tools such as the Clinical Evaluation of Language Fundamentals, Fourth Edition (Wiig, Secord, & Semel, 2004) and the Test of Language Competence—Expanded Edition, Levels 1 and 2 (Wiig & Secord, 1989) should be complemented with standardized measures of social adaptive functioning such as the Vineland Adaptive Behavior Scales—Expanded Edition (Sparrow, Balla, & Cicchetti, 1984) and the Vineland Adaptive Behavior Scales—Classroom Edition (Sparrow et al., 1984). Finally, measures that provide a means to assess an individual's spontaneous communicative bids within natural conversational exchanges such as the Damico Clinical Discourse Analysis (Damico, 1985), the Children's Communication Checklist (Bishop, 1998, 2001), and the Pragmatic Rating Scale (Landa et al., 1992) also may be useful for documenting social communication deficits in higher functioning individuals with ASD.

Return to Top

---

Summary of Recommendations

The SLP plays a critical role in screening and early detection of individuals at risk for ASD and makes referrals to experienced professionals for diagnosis and intervention services. SLPs who acquire and maintain the necessary knowledge and skills can diagnose ASD, typically as part of a diagnostic team or in other multidisciplinary collaborations, and should make appropriate referrals to rule out other conditions and facilitate access to comprehensive services. The SLP who has been trained in the reliable and valid use of diagnostic and assessment tools as well as in the clinical criteria for ASD may be qualified to diagnose these disorders as an independent professional. Individuals with ASD may be eligible for or demonstrate a need for speech-language pathology services due to the pervasive nature of the social communication impairment, regardless of age, cognitive abilities, or performance on standardized testing of formal language skills.

As mandated by the IDEA (2004), SLPs should avoid applying a priori criteria (e.g., discrepancies among cognitive abilities and communication functioning, chronological age, or diagnosis) and make individualized decisions on eligibility for services. Further, in public school settings, a student's diagnosis and eligibility for services must be determined by a team rather than a single individual. Because formal assessment tools may not accurately detect problems in the social use of language and communication, eligibility for special education services may need to be based on clinical judgment and more informal, observational measures.

Framework for Considering the Evidence Base of Treatment

The NRC formed the Committee on Educational Interventions for Children with Autism at the request of the U.S. Department of Education, Office of Special Education Programs. The charge for the NRC committee was to integrate the scientific, theoretical, and policy literature and create a framework for evaluating the scientific evidence concerning the effectiveness of educational interventions for young children with ASD. A report was published in October of 2001 summarizing the findings and recommendations of this committee (NRC, 2001). The NRC report delineated specific conclusions and recommendations about diagnosis and assessment, role of families, goals for educational services, characteristics of effective interventions based on empirical studies, public policies, personnel preparation, and needed research.

To achieve a systematic and rigorous assessment of research studies, the NRC committee established guidelines for evaluating the quality of the scientific evidence based on the following three criteria (NRC, 2001):

• internal validity or control for nonspecific factors, such as maturation, expectancy, experimenter bias

• external validity or control for selection biases addressed in random assignment, adequate sample size, and well-defined populations

• generalization of treatment outcomes documented in a natural setting outside of experimental intervention or with functional outcomes

The NRC committee rated each research study using a 4-point scale for the level of rigor with respect to internal validity, external validity, and generalization with 1 being the highest evidence and 4 being insufficient design or no evidence (NRC, 2001, p. 15). The NRC committee integrated this information with an eye toward convergence of evidence, particularly from independent sources using different methodologies in an effort to characterize features of appropriate, effective educational interventions for children with ASD.

The ASHA Ad Hoc Committee on Autism Spectrum Disorders used the NRC report and adopted the NRC guidelines to evaluate scientific evidence for these guidelines and family of related documents on ASD. The NRC report included a thorough literature review on children from birth to age 8, published in peer-reviewed journals through 2000. The ASHA committee reviewed literature with an effort to identify research published since the NRC report or that which was not included in the NRC report. To assist this committee, the ASHA National Center for Evidence-Based Practice in Communication Disorders conducted a literature search to identify empirical treatment studies on speech, language, and/or communication in children with ASD that have been published since 2000, in adolescents or adults with ASD that have been published over the past decade, and any studies pertaining to multicultural issues and ASD. Studies were included in this review only if they were published in English in peer-reviewed journals and reached a Level 1, 2, or 3 rating on internal and external validity based on the NRC 4-point scale. Chapters in books were used only as resources for summaries of findings based on peer-reviewed publications.

Within the field of ASD, there is a wide variety of approaches to intervention that vary with regard to availability of empirical evidence, longevity, popularity, and influence of marketing efforts. Because of the sheer number and variety of the approaches available, criteria were established for the inclusion and exclusion of intervention information within these guidelines. These criteria focused on available empirical information and were based on criteria established by the NRC (2001). Intervention approaches were discussed in these guidelines if

• empirical research published in a peer-reviewed journal was available that included individuals with ASD;

• intervention outcomes were measured in relationship to the core challenges of ASD; and

• empirical research published in a peer-reviewed journal was available for all or some of its methodological components.

Approaches that had no evidence were excluded. Approaches that had strong refuting evidence, such as facilitated communication and auditory integration training, were also excluded. Members are referred to ASHA position statements for further information on these two methods (ASHA, 1994, 2004a).

Treatment approaches were evaluated for internal validity, external validity, and generalization according to guidelines described by the NRC (2001). Results are summarized in these guidelines in an effort to characterize the strength of empirical evidence for different intervention approaches considering the findings of individual research publications, methodological challenges in clinical studies, selection biases, and the difficulties in measuring meaningful, generalizable outcomes. Like the NRC committee, the ASHA Ad Hoc Committee on ASD recognized that within empirical research, there is a range of emphases and designs depending on the questions being answered, so no attempt was made to prioritize specific interventions or programs. The goal was to integrate intervention information in an effort to identify points of convergence of findings and thus to provide a framework to guide the clinician in making individual intervention decisions.

Return to Top

---

Active Ingredients of Effective Programs

Three major research conclusions emerge from current empirically supported intervention strategies for individuals with ASD and should form the basis for clinical decision making. First, there is empirical support demonstrating the effectiveness of a range of approaches for enhancing communication skills of individuals with ASD along a continuum from behavioral to developmental (Dawson & Osterling, 1997; NRC, 2001; Prizant & Wetherby, 1998; Rogers, 1998). Furthermore, there are no group design studies directly comparing the effectiveness of different approaches using randomly assigned, matched control samples with sufficient sample sizes and adequate statistical power. Therefore, evidence that any one approach is more effective than another approach is not available to date.

Second, intervention research is not yet available to predict which specific intervention approaches or strategies work best with which individuals with ASD. No one approach is equally effective for all individuals with ASD, and not all individuals in outcome studies have benefited to the same degree (NRC, 2001). Group design treatment studies are needed to identify characteristics of individuals with ASD that predict response to treatment (Yoder & Compton, 2004). For clinicians to determine whether an individual with ASD is benefiting from a particular treatment program or strategy, measurement of that individual's progress using systematic methods, such as in single-subject research design, is recommended.

Third, knowledge about the effectiveness of treatment is limited by the outcome measures used in research. The most common reported outcome measures in comprehensive interventions for children with ASD are changes in IQ scores and post-intervention placement (NRC, 2001). The NRC concluded that these measures may not be ecologically valid, because they do not measure meaningful changes within natural learning environments, do not address the core deficits in ASD, and are particularly problematic for young children. The NRC (2001) recognized the need for more meaningful outcome measures and recommended measures that include (a) gains in initiation of spontaneous communication in functional activities and (b) generalization of gains across activities, interactants, and environments. The broad impact of the social communication challenges and problems with generalization for individuals with ASD underscores the critical importance of ecologically meaningful outcome measures. The NRC concluded that learning in natural learning environments appears to be the most effective intervention approach. Not only do such environments invite higher rates of initiation and generalization, they also enhance the ecological validity of the intervention because the behaviors involved are more likely to translate into a better quality of life and increase social acceptance.

Based on a systematic review of research on educational interventions for children with ASD from birth through age 8, the NRC concluded that there is a convergence of evidence that the following characteristics are essential active ingredients or components of effective interventions for children with ASD:

Entry into intervention programs as soon as ASD is suspected. Children who participate in intensive intervention beginning by age 3 have a significantly better outcome than those beginning after 5. Intervention beginning before age 3 appears to have an even greater impact. These intervention findings indicate the pressing need to identify and provide intervention for children with ASD as early as possible (NRC, 2001).

Active engagement in intensive instructional programming for a minimum of 5 hours per day, 5 days a week. Instruction is used broadly to refer to the structure that supports learning and can occur in any environment—the home, school, or community settings. Children with ASD need instructional strategies that ensure that they are actively engaged during activities, and the intensity of programming needs to be sufficient to provide 5 hours per day of active engagement (NRC, 2001). SLP direct services should contribute to the 25 hours per week of active engagement but would likely compose only a small portion of these hours. SLP consultative services should be aimed at helping the communicative partner (i.e., teacher, parent, peer, or sibling) provide the supports and teaching strategies to enhance active engagement in natural learning environments.

Repeated, planned teaching opportunities. Instructional opportunities should be organized in a series of brief time intervals and include sufficient amount of adult attention to meet individualized goals. Goals and objectives should be targeted in systematically planned, developmentally appropriate learning activities (NRC, 2001).

Inclusion of a family component, including parent training. Concerns, priorities and perspectives of the family need to actively shape educational planning. All of the comprehensive intervention programs with the best treatment outcomes included a strong family component. Family members should be supported to be effective members of the educational team and provided with the opportunity to learn strategies for teaching their child new skills and reducing problem behaviors (NRC, 2001).

Low student:teacher ratios. A child must receive sufficient individualized attention on a daily basis so that instructional strategies can be implemented effectively. In order to accomplish this, a low student:teacher ratio is needed, with no more than two young children with ASD per adult in the classroom as a guideline; however, this may vary depending on the functioning level of the individuals with ASD (NRC, 2001).

Mechanisms for ongoing assessment and program evaluation with corresponding adjustments in programming. An individual's progress in meeting objectives should be measured on an ongoing basis to refine the instructional program. Lack of documented progress over a 3-month period should be an indicator that changes in one or more aspects of programming need to be adjusted in some way, such as increasing intensity by adding instructional time or lowering student:teacher ratio, modifying the curricula or instructional strategies, or targeting different objectives (NRC, 2001).

Six kinds of instruction should take priority for individuals with ASD: (a) functional, spontaneous communication; (b) social instruction in various settings throughout the day; (c) play skills with a focus on play with peers and peer interaction; (d) new skill acquisition and generalization and maintenance in natural contexts; (e) functional assessment and positive behavior support to address problem behaviors; and (f) functional academic skills when appropriate (NRC, 2001).

There are many different intervention approaches and strategies that have been developed and implemented for individuals with ASD. Programs differ in how goals are prioritized and the techniques used to target goals. Some programs rely heavily on singular strategies, while others are more comprehensive or eclectic. There are many “name brand” programs; however, there can be much variation in the way these programs are implemented. More important than the name of the program is how the environment and instructional strategies support individualized goals and objectives for the individual with ASD and his or her family and other communication partners (NRC, 2001).

Return to Top

---

Application of Active Ingredients to Decision Making for the SLP

It is challenging for SLPs to make informed decisions about intervention approaches and strategies for individuals with ASD, especially in light of the wide variety of approaches to intervention within the field of ASD and the variance in regard to availability of empirical evidence, longevity, popularity, and influence of marketing efforts. The following questions should be considered by clinicians to aid this decision-making process based on efficacy research and core characteristics of ASD:

1. Can the intervention approach harm the child with ASD?

2. Is there empirical evidence to support or refute the intervention approach?

3. What is the impact of the intervention on the caregiver?

4. How will you know the treatment is working?

5. Does the intervention program match the family vision and the developmental level and learning style of the individual with ASD?

6. Does the intervention program include the family?

7. Does the intervention program support the development of spontaneous communication?

8. Does the intervention program consistently support social development, play, and interaction with peers as appropriate?

9. Does the intervention program promote generalization and maintenance in natural contexts?

10. Does the intervention program incorporate functional assessment and positive behavior support to address problem behaviors?

11. Does the intervention program provide sufficient intensity for meaningful progress?

12. Does the intervention program address functional academic skills, if appropriate?

13. Does the intervention program address self-advocacy skills and independence for activities of daily living for adults with ASD?

The IDEA (2004) added a new provision that the individualized education program should include a statement of the special education and related services, based on peer-reviewed research to the extent possible (§1414(d)(1)(A)(i)(IV)). In public school settings, the SLP should play an active role as a team member to stay abreast of peer-reviewed research and incorporate strategies based on evidence-based practice into educational programs for students with ASD.

Return to Top

---

Summary of Recommendations

Based on a review of empirical research on specific intervention strategies and comprehensive intervention programs using a package of strategies, the NRC (2001) committee concluded that there are a number of critical features that are the active ingredients in effective intervention programs for children with ASD, delineated above. SLPs should be an active member of educational teams that work collaboratively to incorporate these critical features into educational programs of individuals with ASD as well as build the capacity within school districts to incorporate these features systemwide.

Assessment Goals and Strategies

Following screening, diagnosis, and eligibility consideration, ongoing assessment is critical to guide program planning. Assessment of an individual with ASD should be an ongoing process for achieving a number of essential goals. These include: (a) to determine an individual's current profile of social communication skills, (b) to identify learning objectives that are priorities within natural communication contexts, and (c) to examine the influence of the communication partner and the learning environment on the individual's competence as a communicator (Meisels, 1996). Dynamic assessment is a term used for assessment protocols in which support is systematically provided to determine what factors influence and enhance an individual's current skills and ability to complete tasks that would otherwise be too difficult for her or him to accomplish independently. Dynamic assessment procedures should be implemented, as these procedures identify not only those skills that an individual has achieved but also those skills that may be emerging and what contextual supports enhance communication skills, such as augmentative and alternative communication (AAC) and modeling provided by communication partners (Mirenda, 2003; Olswang, Bain, & Johnson, 1992; Schuler, 1989; Vygotsky, 1978).

With these assessment goals in mind, assessment strategies should not rely solely on standardized, norm-referenced tools, as information should be gathered across natural social contexts (Schuler, Prizant, & Wetherby, 1997). Additionally, primary caregivers and communication partners (e.g., family members, teachers, clinicians, and peers) should be incorporated as active participants and informants in this assessment process (Prizant & Bailey, 1992; Schuler, 1989). Observing an individual in his or her natural social contexts, gathering information from that individual's communication partners through questionnaires and/or interviews, and staging communication contexts that assess an individual's spontaneous use of communication and language provide critical measures of an individual's strengths and functional needs across meaningful contexts (Schuler, 1989; Wetherby & Prizant, 1993).

Likewise, these assessment strategies provide a means to examine the need to support communication partners in their efforts to respond to the individual's subtle bids for communication, interpret the functions of problem behavior, and modify the environment to foster social engagement. The idiosyncratic language used by individuals with ASD makes it difficult for communication partners to be responsive and adjust the quantity and quality of their language models. Consequently, language modeled for individuals with ASD is often too complex or too simple, limiting opportunities for communication growth. Likewise, a communication partner may misinterpret an individual's subtle bids for communication or the functions of problem behavior, a pattern that may limit the individual's exposure to modeling for coping and expressing intentions in more socially appropriate ways.

Return to Top

---

Prioritizing Intervention Goals

Based on the ongoing assessment of an individual's strengths and needs as well as the strengths and needs of his or her communication partners, intervention goals and strategies can be prioritized. Developmental sequences and processes of language development should provide a framework for determining baselines and implications for intervention goals (Schuler et al., 1997). Family priorities also should be considered paramount when selecting intervention goals, as meaningful outcomes are strongly correlated with communication competence across functional social contexts (e.g., home, school, vocational, and community settings). An individual with ASD will demonstrate greater social communication competence when goals are prioritized to ensure effective communication in meaningful contexts and across natural communication partners (Wetherby, Schuler, & Prizant, 1997).

The most critical domains for prioritizing intervention goals should be derived from the core features of ASD and the core challenges that affect social adaptive functioning within the ever-changing social contexts of an individual's natural routines. As outlined previously in this document, core challenges are noted in aspects of joint attention (e.g., social orienting, establishing shared attention, monitoring emotional states, and considering another's intentions), social reciprocity (e.g., initiating bids for interaction, maintaining interactions by taking turns, and providing contingent responses to bids for interaction initiated by others), language and related cognitive skills (e.g., understanding and using nonverbal and verbal communication, symbolic play, literacy skills, and executive functioning), and behavior and emotional regulation (e.g., effectively regulating one's emotional state and behavior, maintaining social engagement, and attending to salient aspects of the social environment). Thus, intervention goals should be prioritized under these domains following a developmental framework.

These core challenges take different forms as an individual matures and responds to intervention. Application of a developmental framework ensures that appropriate goals are being addressed prior to the emergence of symbolic language, at emerging language stages, and at more advanced stages of conversational discourse (NRC, 2001; Prizant, Wetherby, Rubin, & Laurent, 2003). Table 1 provides sample intervention goals for prelinguistic, emerging language, and more advanced language stages. Actual goals for an individual should vary based on those aspects of development that are consistent with family priorities and an individual's functional needs within his or her current social contexts. Goals should incorporate the functional use of the individual's full communication abilities using a multimodal communication system. Decisions about the integration of modes of communication (e.g., spoken language, gestures, sign language, picture communication, speech generating devices [SGDs], and/or written language) should be individualized according to specific capabilities and contexts of communication, as well as cultural issues.

Return to Top

---

Summary of Recommendations

Integral to the diagnostic criteria, all individuals with ASD have core challenges in the area of social co