CT FEAT: Families Helping Children Achieve Their Full Potential
  Print This Article    

Randomized Trial of Intensive Early Intervention for Children with Pervasive Developmental Disorder

Tristram Smith Washington State University
Annette D. Groen
Metropolitan State College of Denver
Jacqueline W. Wynn
University of California , Los Angeles

American Journal on Mental Retardation; v. 105, no. 4; (July 2000); p. 269-285; ISSN: 0895-8017

Young children with pervasive developmental disorder were randomly assigned to intensive treatment or parent training.  The intensive treatment group ( 7 with autism, 8 with pervasive developmental disorder not otherwise specified NOS) averaged 24.52 hours per week of individual treatment for one year, gradually reducing hours over the next 1 to 2 years.  The parent training group (7 with autism, 6 with pervasive developmental disorder NOS) received 3 to 9 months of parent training.  The groups appeared similar at intake on all measures; however, at follow-up the intensive treatment group outperformed the parent training group on measures of intelligence, visual-spatial skills, language, and academics, though not adaptive functioning or behavior problems.  Children with pervasive developmental disorder NOS may have gained more than those with autism.

After years of debating whether or not early intervention helps children with developmental delays (Weinberg, 1989), researchers have largely come to agree on a middle ground: Early intervention is beneficial for many children, but gains tend to be limited (e.g., Scarr & Arnett, 1987). For example, many researchers in the area of early intervention for children with developmental disabilities have shown that such intervention prevents declines in intellectual development and may reduce family stress (e.g., Guralnick, 1998), though children continue to display substantial delays. 

Nevertheless, there have been reports of larger improvements. Of particular interest, in peer-reviewed studies, seven independent groups of investigators have described dramatic gains with early intervention for children with autism (reviewed by Smith, 1999). In all studies, interventions were based on applied behavior analytic research and theory (Green, 1996) and were intensive (15 to 40 hours per week). Reported gains have included average increases of approximately 20 points in IQ (Harris, Handleman, Gordon, Kristoff, & Fuentes, 1991; Lovaas, 1987; Sheinkopf & Siegal, 1998) and other standardized test scores (Anderson, Avery, DiPietro, Edwards, & Christian, 1987; Birnbrauer & Leach, 1993; Hoyson, Jamison, & Strain, 1984; McEachin, Smith, & Lovaas, 1993), as well as less restrictive school placements (Fenske, Zalenski, Krantz, & McClannahan, 1985; Lovaas, 1987).

Such results may not only enhance the outlook for children with autism but also raise optimism about the extent to which children with other developmental disorders may benefit from early intervention (Guralnick, 1998). However, the validity of the results has been a topic of intense debate. A study by Lovaas and colleagues (Lovaas, 1987; McEachin, Smith, & Lovaas, 1993) has garnered particular attention. Lovaas evaluated three groups of children with autism who were under 4 years old at intake. The experimental group (n = 19) received intensive treatment, which consisted of 40 hours per week of one-to-one, in-home, applied behavior analytic intervention for 2 or more years. One control group (n = 19) received minimal treatment (10 hours per week or less); a second control group (n = 21) was treated at other agencies and had no contact with Lovaas's clinic. Though the three groups did not appear to differ at intake, the intensively treated children substantially outperformed the children in control groups at age 7. Their mean IQ was 83 compared to 52 and 58, respectively. Also, 9 of 19 received passing grades without special assistance in classes for typically developing children compared to only 1 of 40 in the control groups. Moreover, at a follow-up conducted when the children averaged 12 years of age, the intensively treated children maintained their gains and also functioned more satisfactorily than did minimally treated children on measures of adaptive behavior and personality (McEachin et al., 1993).

McEachin et al. (1993) identified a number of strengths of this study, including (a) experimental and control groups that did not differ on 19 of 20 intake variables, (b) intake and follow-up evaluations conducted by blind examiners independent of the study, (c) reliance on treatment approaches developed from extensive research on reducing maladaptive behaviors and enhancing skills in children with autism (cf. Newsom & Rincover, 1989), (d) use of a detailed treatment manual (Lovaas et al., 1981) and associated videotapes to standardize the interventions that children received, and (e) follow-ups to assess maintenance of treatment gains conducted many years after termination of treatment. However, others identified many possible flaws, notably the following (Gresham & MacMillan, 1997; Schopler, Short, & Mesibov, 1989): First, assignment to groups was based on whether or not therapists were available to provide intensive treatment rather than on a more arbitrary procedure, such as the use of a random numbers table. Thus, assignment could have been biased. Second, because children were referred to outside examiners, they received a variety of different tests rather than a uniform assessment protocol. Hence, assessment results may have been unreliable. Third, selection criteria such as IQ cut-offs may have been unduly restrictive, yielding a sample with an unusually favorable prognosis. Fourth, the large amount of treatment and the level of expertise required for proper implementation may have been too much for other professionals to duplicate, too stressful for most children and families to tolerate, and too costly for funding agencies to support. Lovaas and colleagues concurred with the first two of these criticisms, though doubting the importance of the second. They disputed the other criticisms but emphasized the need for replication to confirm the results (Lovaas, Smith, & McEachin, 1989; Smith & Lovaas, 1997; Smith, McEachin, & Lovaas, 1993).

Anderson et al. (1987), Birnbrauer and Leach (1993), and Sheinkopf and Siegal (1998) conducted partial replications of the study by Lovaas and colleagues. Children in these studies received fewer hours of treatment (18 to 25 hours per week vs. 40 hours) from less experienced personnel than in the Lovaas study. All studies showed substantial average increases in nonverbal IQ (22 to 29 points), but gains in other areas were smaller than those reported by Lovaas (1987).

The present study was designed to extend this literature. Children received early intervention based on the same treatment manual used by Lovaas (1987), implemented by personnel who met the qualifications specified in that study and were independent of Lovaas (1987). However, because of concerns about cost of service delivery and stress on children and families, intervention was made less intensive than that in the Lovaas study, as described later in the Treatment section. To address criticisms of previous research and increase methodological rigor, we conducted a fully randomized clinical trial with uniform, comprehensive assessment protocols for all participants. To evaluate treatment efficacy for a wider range of children, we studied not only children with autism but also children with pervasive developmental disorder not otherwise specified (NOS). Because such children are often viewed as having "mild autism" (Towbin, 1997), they were hypothesized to be appropriate candidates for the intervention we provided.



All referrals to the UCLA Young Autism Project between 1989 and 1992 who met the following criteria were enrolled in the study: (a) chronological age (CA) between 18 and 42 months at the time of referral, (b) residence within a one-hour drive of the research/treatment site (the UCLA Young Autism Project), (c) IQ ratio between 35 and 75, (d) diagnosis of autism or pervasive developmental disorder NOS, and (e) absence of major medical problems other than autism or mental retardation (e.g., cerebral palsy, blindness or deafness, known genetic disorders such as Down syndrome, or neurological conditions such as uncontrolled seizure disorders). Twenty-eight children met these criteria and participated in the study, including 14 diagnosed with autism and 14 diagnosed with pervasive developmental disorder NOS. There were no dropouts among this group of children. However, one child's family declined participation at intake, and 8 other children were excluded (4 because they did not have a diagnosis of autism or pervasive developmental disorder NOS; 2 because they scored below the IQ cutoff; and 2 because they were in foster care, without a permanent residence in which to provide the home-based services offered to children in this study.

Diagnosis for all 28 participants was made independently of the study by licensed psychologists at the California State Regional Centers (a state agency that coordinates services for individuals with developmental disabilities). Nineteen participants had also received a second, independent diagnosis prior to entry into the study (8 from the UCLA Neuropsychiatric Institute; 3 from White Memorial Hospital; 3 from former clinic supervisors at the UCLA Young Autism Project who had become licensed, doctoral psychologists and were blind to the children's previous diagnostic history and independent of the study; 1 from the University of Southern California Medical Center; 1 from Children's Hospital; and 1 from Cedars-Sinai Hospital). The second diagnosis was identical to the Regional Center diagnosis for all participants except one, who was diagnosed with autism at the UCLA Neuropsychiatric Institute but pervasive developmental disorder NOS at the Regional Center (and, hence, was classified as having pervasive developmental disorder NOS for the purposes of this study).

Table 1 summarizes background information on all participants in the two groups (intensive treatment and parent training, described in Treatment), as reported by each child's primary caregiver on the Family Background Questionnaire (Siegal & Elliott, 1988). The groups appeared similar on all variables. Participants had diverse ethnic and socioeconomic backgrounds, consistent with the general population in the Los Angeles area. In addition, they resembled other populations of children with pervasive developmental disorder in terms of sex ratio (Smith, 1997) and frequency of medical conditions (Rutter, Bailey, Bolton, & LeCouteur, 1994).

Table 1: Background Information by Group


                                                   Intensive treatment


Parent training








Intake CA (in months)a








Follow-up CA (in months)a













Boy/Girl ratio



































Single parent household (n)





Household incomec

$40-50,000 (<$10,000 to $75-100,000)

$40-50,000 (<$10,000 to



Years of schoolingc

































Medical conditionsd






Motor delayse






a Mean (SD).

b Autism/pervasive developmental disorder NOS (not otherwise specified).

c Median (range).

d Defined as any prenatal, perinatal, or neurological condition that resulted in medical treatment or any other medical condition that resulted in hospitalization. Medical conditions in the intensive group: one child had a skull fracture at 4 months of age; in the parent training group, one child had tubercular meningitis at 15 months of age.

e Defined as sitting independently after 8 months of age or walking independently after 16 months of age.


Children were assigned to intensive treatment or parent training based on the following matched-pair, random assignment procedure: Once intake assessments had been completed on 4 to 8 children, they were divided into two cohorts, those with a diagnosis of autism and those with a diagnosis of pervasive developmental disorder. Their first names and Bayley IQs were then given to an independent statistician, who had no other information about the children. The statistician paired the children in each cohort based on IQ (the two highest forming one pair, the next two forming another pair, etc.). Finally, using a random numbers table, he assigned one member of each pair to the intensive treatment group and the other to the parent tra

Back to Top



CT FEAT, P.O. Box 370352, West Hartford, CT 06137-0352 Phone (860) 571-3888
Copyright 2010 CT FEAT
CT FEAT    Contact Us