Pervasive Developmental Disorders in Young Children

There is much concern about whether the prevalence of pervasive developmental disorders (PDDs), sometimes referred to as autistic spectrum disorders, is increasing.¹ No study is likely to answer that question because the last 20 years have seen dramatic changes in the diagnostic criteria, an increase in public awareness of autism and related disorders, and sweeping changes in the types and availability of therapeutic and educational services. However, problems with determining whether prevalence of the PDDs is increasing should not serve to distract from the question that can be answered—ie, what is the prevalence of PDDs now.

Determining the prevalence of a behaviorally defined disorder is more difficult than determining the prevalence of conditions for which biological markers exist. The symptoms of PDDs—deficits in social reciprocity and communication and restricted and repetitive behaviors—vary in severity and may appear quite different in children with different cognitive abilities. Autism, Asperger syndrome, and pervasive developmental disorder-not otherwise specified (PDD-NOS) have been proposed to occur along a spectrum within PDD.²

Autism, the most severe form of PDD, was estimated to have a general population prevalence of 4 to 5 per 10 000 prior to 1988, according to surveys of children and adolescents.³ Prevalence studies using current diagnostic formulations and tiered screening of regional populations of preschool children have reported prevalence rates for the specific diagnosis of autism of 21 to 31 per 10 000^{4 - 6} and of 57.9 per 10 000 for all PDDs combined.⁵ One small, regionally defined study of 7-year-olds reported rates for autism as high as 60 per 10 000 and for all PDDs reported 121 per 10 000.⁷ In a recent case finding study in Brick Township, New Jersey,^{8 - 9} the prevalence of autism was reported to be 40 per 10 000 and for all PDDs it was reported to be 67 per 10 000 with similar rates identified for children between the ages of 3 and 5 years and 6 and 10 years. The age of the sample and methods of case ascertainment may affect the prevalence rates reported in such studies.

In this issue of THE JOURNAL, Chakrabarti and Fombonne¹⁰ report a prevalence rate in the United Kingdom of 16.8 per 10 000 for autism and 63 per 10 000 for all PDDs in a regional sample of children younger than 5 years. The identification of developmental concerns was aided by health visitors, specialized nurses who assist in monitoring health and development when a child is approximately 7 months, 18 to 24 months, and 3 years of age. Children with delayed or atypical development were referred to a multidisciplinary team who used the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM IV)¹¹ criteria to establish the diagnosis of PDD. Children identified as having PDD on clinical grounds were further evaluated using the Autism Diagnostic Interview-Revised,¹² a semistructured interview valid for establishing the diagnosis of PDD and autism in children as young as 18 months. Chakrabarti and Fombonne provide further evidence that PDDs are more common in preschool children than was previously thought, even when case finding occurs only through careful clinical evaluation.

Just as newer studies are indicating higher prevalence rates for PDDs than were reported in earlier studies, they also indicate lower rates of mental retardation in children with PDDs. Only 26% of the children studied in the report by Chakrabarti and Fombonne were diagnosed with both PDDs and cognitive limitations. These were primarily children who met diagnostic criteria for autism. This finding agrees with the observations from other studies^{4 - 5 ,8} showing that 45% to 60% of young children with nonautistic PDDs have intellectual abilities in the normal range.

Chakrabarti and Fombonne reported that evaluations looking for an origin for the disorder were guided by clinical judgment. Ten percent of the children had medical conditions identified. This is not dissimilar from the average rate of 6% of patients with medical conditions that might be related etiologically across 23 epidemiologic studies of children and adolescents with autism that were examined by Fombonne in 1999.¹³ Most cases of PDD cannot be accounted for by another disorder that affects brain development. The low yield of diagnostic testing has resulted in the recommendation that medical evaluations in each child should be guided by the information obtained from a careful medical history and physical examination.^{2 ,14}

More than 75% of children with autism who have language delays and differences in social interaction and behavior are identified by their parents as being "different" by age 2 years.^{15 - 16} It is not surprising that Chakrabarti and Fombonne found that children with more symptoms are identified at an earlier age. The mean age of diagnosis for autism was 30.0 months while the mean age of diagnosis for PDD-NOS was 37 months. Children with Asperger syndrome have, by definition, normal early word acquisition and normal intelligence, and their mean age at diagnosis was 47.5 months. It is probable that even more children with Asperger syndrome will be identified once they are in a classroom situation, where their interactions with other children may be observed.

There is increasing evidence that behavioral and educational interventions with young children may significantly improve developmental and behavioral outcomes¹⁷ and that basic deficits in play and communication may be therapeutically modified.¹⁸ In the United States, the availability of developmental assessment and treatment for infants and toddlers with developmental delays is mandated by federal legislation through early intervention programs.¹⁹ Despite this system, there is often a lag of a year of more between when the time concerns are identified and when a diagnosis is made. In their study, Chakrabarti and Fombonne attribute 80% of the initial referrals of these young children for further evaluation to health visitors. The interval between the mean age at referral and the age at diagnosis was approximately 5 months. Developmental surveillance by busy practitioners in an office setting may not be as effective for early detection and referral as home visitation nurses trained in developmental surveillance. To increase professional awareness, the American Academy of Pediatrics and the American Academy of Neurology have recently published statements about assessment of young children with PDDs.^{2 ,14} However, it is uncertain whether efforts like this will be enough to bring the age of diagnosis closer to the age when symptoms are first recognized by the parents or other strategies will be required to improve developmental surveillance.

Chakrabarti and Fombonne have corroborated that PDDs are not rare among preschool children. Given the limitations of current knowledge, valuable data sets such as these should be used to increase understanding of the characteristics of young children with PDDs as well as to determine the current prevalence. Accurate prevalence data will assist in planning for services that affected children and their families will need. Parent groups have allied with the US Congress in a call for more research into the neurobiology, treatment, and epidemiology of these disorders.²⁰ While waiting for the answers that only careful research can give, it is incumbent on those who care for young children to identify children with atypical social and language development, refer them for assessment, and advocate for effective treatments.