Screening for autism and pervasive developmental disorders in very young children

Introduction

Autism is a behavioral syndrome characterized by a triad of impaired social interaction, communication, and imaginative development, with repetitive stereotyped behaviors. The definition of autism has evolved over the years and was broadened with the introduction in the 1980s of the terms pervasive developmental disorder and Asperger syndrome. In the Diagnostic and Statistical Manual of Mental Disorders IV-TR (DSM-IV-TR),[1] pervasive developmental disorder is the umbrella term covering autistic disorder, Asperger’s disorder, and pervasive developmental disorder NOS (not otherwise specified). The latter is a term for atypical autism, that is, either milder symptoms or characteristics in only one or two of the three areas necessary for a diagnosis.

The pervasive developmental disorders also include Rett’s disorder, a neurological condition, and childhood disintegrative disorder, a poorly defined variant of autism. Both affect so few children that they will not be considered here. Asperger disorder will also not be discussed further since it is a mild form of high-functioning autism that manifests later, so would not be considered in very young children. At present, there is no clear distinction between autistic disorder and pervasive developmental disorder NOS, nor is there clear diagnostic criteria for them. Thus, the term autism spectrum disorders has gained support since it reflects the range in both type and severity of symptoms seen in children with autism. In this article the terms pervasive developmental disorder (PDD) and autism spectrum disorders (ASD) are used interchangeably to refer to the spectrum.

It has been suggested that these disorders be reclassified using terms such as social communication spectrum disorder [2] or multisystem developmental disorder. The latter, proposed by Zero to Three: National Center for Infants, Toddlers and Families, suggests that in some cases autistic socialization and communication difficulties are secondary to "regulatory dysfunctions" and the categories of pervasive developmental disorder/ autism spectrum disorders should only be used if the "deficits appear fixed." These attempts at new definitions reflect the difficulty of describing children who may not be clearly autistic, who are very young and have not grown into their autism, or for whom the relationship difficulties are related to difficulties in motor planning, auditory and visual processing deficits, or in regulating emotions. This description may be applicable to a small group of children with autistic symptoms. The model and concept of multisystem developmental disorder lack a research base, however, and have not gained widespread support.[3]

Prevalence figures vary according to diagnostic criteria used. But because autism affects as many as 1 in 500 children, it can be considered a common disorder of childhood.[4] This means that over 100 babies may be born in BC each year with autism spectrum disorders. Boys are affected 4 to 5 times more often than girls, and more than 70% of autistic children also have some degree of mental retardation.

Symptoms of ASD/PDD

Children with autism show characteristics in three core areas:

• Qualitative impairment in social interaction

• Qualitative impairment in communication and imagination

• Restricted, repetitive, and stereotyped patterns of behavior, interests, or activities

See Table 1 for more detail.

The lack of social relatedness is considered to be the main symptom and handicap in autism. This trait is often present very early, with poor eye contact, disinterest or dislike in being held, and being "in his or her own world." A lack of showing and sharing interests is often present and there is less desire to play with other children. The communication difficulties are seen with delayed development of speech and language. Even when language develops it is often rote, repetitive, and lacking communicative intent. The repetitive, rigid behaviors may be seen in flapping, rocking, or finger posturing, but also may be seen in having only a small number of activities and interests or in repetitive play such as lining up objects. In very young children, behavioral rigidity may result in tantrums and distress.

Early identification

There has been increasing interest in identifying autism and related disorders as early as possible to allow for early intervention. The influence of the environment on brain structure and function is greatest during the first few years of life, and intervention may prevent the development of maladaptive behaviors.[5,6] Further, early intervention provides support for parents. Indeed, the confusion and stress experienced by families is often replaced by remarkable coping once a diagnosis is reached and parents are provided with recommendations for intervention.

Early intervention programs differ in their relative emphasis on language, social skill development, or behavioral management, but common characteristics of successful programs are that they are offered early and intensively.[7] Thus, when a parent approaches a physician with a concern about a child’s development, "wait and see" is not the best response. There are, of course, anxious parents of normally developing children who may present with persistent concerns about their child’s development. How can the physician differentiate between this group, and those who are understandably anxious about their child’s abnormal development?

Recent reports have made recommendations for the early identification of autism.[4,8] They all concur that monitoring language and behavior in very young children, assisted by screening tools or checklists, then referral to an experienced team, constitute best practice.

Identification and diagnosis of autism in very young children has been considered to be difficult, but there is increasing evidence that it is possible. Children with autism can be distinguished from normal children or from those with mental handicaps or language disorders at 2 to 3 years of age by experienced clinicians.[9-11]

Diverse factors complicate early identification. These include the broad range in normal developmental milestones, the variability in the emergence of autistic behavior in children, the lack of appropriate referrals by professionals to whom parents express concern, the family’s lack of knowledge of abnormal child development, and language barriers in new immigrant families.

Many parents report retrospectively that they suspected problems in their children, often in infancy but usually by 1 to 11/2 years. About 25% report normal development up to 18 months, with increasing concern when language is slow to develop.

Diagnostic approaches

What should the physician do when a parent comes with concerns? First, it is important to check the developmental milestones of the child and ensure that vision and hearing are normal (see Table 2). The focus must be shifted from looking for typical speech/language problems to identifying abnormal perceptual and social dysfunctions.[10] The items that distinguish children with autism from others with developmental delays focus less on early speech and language delays or repetitive behavior and more on abnormal play and social isolation, including poor eye contact and lack of sharing interests or lack of imitation. Motor stereotypies and language delays are common in all developmental disabilities, and behaviors such as the need for sameness or sensory disturbances either may develop later or are not common enough in autism to be diagnostic at a young age.[7,12]

Thus, to identify autism in very young children, and assist in distinguishing them from children who are only developmentally delayed, look for differences in four key areas of behavior:

• Social isolation, described as "ignoring others," "doesn’t respond to name," "fails to attend to voice," "acts as if deaf," "does not imitate."

• Abnormal play, either "absence of pretend play," "playing only with parts of objects" or "playing only with a few things," "inappropriate relating to toys."

• Lack of joint attention, or a lack of "showing," "sharing," "commenting," "directing other’s attention," or "protodeclarative pointing." (The goal of protodeclarative pointing is to indicate objects of interest rather than to obtain an object.)

• Difference in eye contact described as "abnormal gaze" or "difficulty getting or sustaining eye contact."

If any child has a combination of a lack in pretend or social play, joint attention, and pointing out interests at 18 months, it makes good clinical sense to refer him or her for a specialist assessment for autism. Table 3 indicates the parental concerns that should be red flags and prompt further investigation.[4]

There exist a number of screening tools, and reference to one of them is important when a diagnosis of autism spectrum disorder is being considered. Here is a brief description of the most widely used or readily available questionnaires or checklists to assist with screening for autism in very young children. Instruments that rely on parents’ reports must be evaluated carefully due to the possible sources of bias in either direction. Wording of items can be interpreted differently, memory can affect judgment, and some parents may have difficulty describing behaviors in English or have low literacy for reading questionnaires.

The Head to Toe checklist

The BCMA published a developmental checklist for parents in the Winter 1997 edition of Head to Toe,[13] which encouraged parents to discuss deviations in their child’s development with their family doctor or pediatrician. The word autism was not included in this checklist in order to avoid overdiagnosis and unnecessary heightened parental anxiety. This checklist has been endorsed by the Ministry for Children and Families and has been widely circulated in BC.[14]

The CHAT screening tool

The CHAT (Checklist for Autism in Toddlers—A Screening Tool) was developed in Britain as a brief screening device to be administered by GPs or health visitors to 18-month-olds. It was initially part of a study aimed to establish if detection of autism was possible at 18 months of age. It has nine questions for parents and five observations for the clinician. The absence of pretend play, gaze monitoring, and protodeclarative pointing predicted the children who received a diagnosis of autism at 3.5 years.[15] The authors concluded that consistent failure of these three items at 18 months carries an 83.3% risk of autism.[15,16] Although the CHAT, when used as a screen by a community health nurse, is highly specific (i.e., it screened out normal children), its sensitivity of picking up autism spectrum disorders was only 38%. Positive predictive value was 29%. When used as a two-stage measure, with the second administration 1 month after the first, by a team experienced with autism, the positive predictive value rose to more than 75%. Although it is promising that it was possible to identify many cases of autism at 18 months, the CHAT must be used, preferably by experienced professionals, as a screen only, and not for diagnostic purposes.[17]

The Pervasive Developmental Disorders Screening Test

The Pervasive Developmental Disorders Screening Test [18] developed by Bryna Siegel is another questionnaire for parents to identify ASD. This promising instrument is used between the ages of 18 months and 3 years, and includes 71 items that were reliably reported by parents and which correlated with later clinical diagnosis. The questions included were those rated "mostly true" by 50% or more of parents of children with autism. The test includes three levels, each of which takes about 5 minutes to complete. A cut-off is provided beyond which a child should be referred for further investigation of autism. This test reportedly detects potential cases of autism about 87% of the time, but it also has a high rate of false positives. The value of this test is in the staged approach, which allows for a brief screen in a primary-care clinic before referral to a developmental clinic.

Other screening measures exist, but discussion of these goes beyond the focus of this article.

Despite the usefulness of screening instruments, it should be noted that children with mental retardation and language disorders might score positive on them. In fact, there are risks involved in labeling all children with autistic-like behaviors as autistic,[19] since inappropriate, unnecessary, or costly treatments may be implemented as a result.

Ideally there should be a tiered approach to early identification of autism, which begins with developmental observation, or surveillance, for all young children. This starts with the birth of a child and continues as part of periodic health exams. The family physician, community health nurse, or other appropriately trained and qualified health-care provider is in a position to attend to the parents’ concerns, obtain the relevant developmental history and monitor milestones, observe the child carefully, then share his or her opinion and concerns with the parents. When there are questions about communication and social differences, this process is assisted by screening at an early age with a tool such as the CHAT or Pervasive Developmental Disorders Screening Test Level I, or comparison with DSM-IV-TR criteria. This should be followed by a review by a pediatrician or child psychiatrist and then full diagnostic evaluation by a multidisciplinary team. The child’s strengths and weaknesses must be profiled to assist in designing appropriate intervention. This approach has been widely recommended [3,4] and has been endorsed locally by the Autism Working Group within Children’s & Women’s Health Centre of BC. The C&W Clinical Guideline for the Diagnosis and Assessment of ASD/PDD is included in the article "Autistic disorders: What can a physician do?".

Comprehensive diagnostic evaluation is best provided by a multidisciplinary team skilled and experienced in the field of autism. The team may include a pediatrician, psychiatrist, speech language pathologist, psychologist, and possibly an occupational therapist or social worker. Team assessments are available at Sunny Hill Health Centre for Children (Autism Spectrum Resource Team) and Queen Alexandra in Victoria. Psychiatric assessments and other aspects of a multidisciplinary assessment (done in a serial fashion) are offered at BC’s Children’s Hospital. Child development centres throughout the province offer differing components of a multidisciplinary assessment, and the team assessment may be done in a serial fashion if the information is coordinated for sharing with the family.

In the absence of the availability of autism specialists or while the child is waiting for an assessment, intervention should not be postponed when a screening instrument indicates possible autism. Even if his or her developmental differences do not turn out to be autism, the child and family will benefit from early identification and intervention. The child’s progress and response to treatment often contributes valuable information to assist in clarifying and confirming the diagnosis.

Conclusion

Since autism is a spectrum disorder with a wide range in the number and severity of symptoms, it is not surprising that early identification can be difficult and imprecise. Children change as they grow and develop, particularly if intervention is effective, so diagnoses can change in either direction. In the experience of the team at Sunny Hill, fewer than half of the children who show some autistic-like behavior in early childhood were eventually diagnosed with autism at 4 to 5 years of age. Almost all, however, were found to have some form of developmental disability. Despite these problems, professionals should not abandon efforts to screen for autism in very young children. Provisional diagnoses such as "suspected autism," "pervasive developmental disorder," or "autism spectrum disorder" are useful to facilitate service delivery and leave room for change as the child grows.

Adapted from the Diagnostic and Statistical Manual of Mental Disorders IV-TR, American Psychiatric Association, 2000:75.

Adapted from the New York State Clinical Practice Guideline.

Adapted from Filipek et al., 1999.

Additional reading

Cohen DJ, Volkmar FR. Handbook of Autism and Pervasive Developmental Disorders. 2nd ed. New York: Wiley, 1997.

Lieberman AF, Wieder S, Fenichel E. DC 0–3 Casebook: A Guide to the Use of Zero to Three’s Diagnostic Classification of Mental Health and Developmental Disorders of Infancy and Early Childhood in Assessment and Treatment Planning. Washington, DC: Zero to Three, 1997.

Siegel B. World of the Autistic Child: Understanding and Treating Autistic Spectrum Disorders. New York: Oxford University Press, 1996.

Weider S (ed). Zero to Three: Diagnostic Classification of Mental Health and Developmental Disorders of Infancy and Early Childhood. Arlington, VA: Zero to Three/National Center for Clinical Infant Programs, 1994.

Terry M. Kope, MD, FRCPC, Linda C. Eaves, PhD, and Helena H. Ho, MD, FRCPC

Dr Kope is a clinical assistant professor of psychiatry at the University of British Columbia and is in private practice in Delta, BC. Dr Eaves is a psychologist on the Autism Spectrum Resource Team, Sunny Hill Health Centre for Children. Dr Ho is a clinical professor in the Department of Pediatrics, UBC, and team leader of the Autism Spectrum Resource Team at Sunny Hill Health Centre for Children.