Doi:10.1016/j.rasd.2009.12.00

Research in Autism Spectrum Disorders 4 (2010) 526–538 J o u r n a l h o m e p a g e : h t t p : / / e e s . e l s e v i e r . c o m / R A S D / d e f a u l t . a s p An in-depth examination of optimal outcome children with a history ofautism spectrum disorders Elizabeth Kelley Letitia Naigles, Deborah Fein University of Connecticut, Department of Psychology, Storrs, CT 06269-1020, USA Previous research has suggested that some children with autism spectrum disorders (ASD) may improve to such an extent that they lose their diagnosis, yet little research has Received in revised form 25 November 2009 examined these ‘optimal outcome’ children in depth. We examined multiple aspects of functioning in a group of 13 optimal outcome (OO) children, matched on age, gender, andnon-verbal IQ to a group of typically developing children (N = 14) and a group of high- functioning children with ASD who still retained a diagnosis on the autism spectrum (N = 14). These children were tested on average about eight years after they had been diagnosed (OO = 93 months, HFA = 94 months). Unlike their high-functioning peers with ASD, the OO group’s adaptive and problem behavior scores fell within the average range.
They also showed average language and communication scores on all language measures.
The HFA group, however, continued to show pragmatic, linguistic, social, and behavioraldifficulties. The OO children tended to have been diagnosed at younger ages and weresignificantly more likely to have received intensive early intervention. Although the high-functioning children with ASD continued to show difficulties in the behavioral realm, theindividuals in the OO group were functioning within the average range on all measures.
Future research should address how this optimal outcome is achieved.
ß 2010 Elsevier Ltd. All rights reserved.
1. An in-depth examination of optimal outcomes in children with a history of autism spectrum disorders An extensive debate exists in the literature as to whether children with autism spectrum disorders (ASD) ever reach a point where they lose their diagnosis, so that they are no longer considered to be on the autism spectrum.Althoughoutcome for children with ASD, particularly for those with an IQ in the normal range of functioning, is thought to be betterthan it was 30 years ago (there is much disagreement in the field as to what thegeneral range of outcome is for these children and whether children can actually lose their diagnosis over the course ofdevelopment (Some authors have hypothesized that there are children withASD who are indeed recovered, or who appear to have lost the behavioral manifestations of a genetically based, biologicaldisorder (see ). However, others state that while these children may improve to such an extent that they are * Corresponding author at: Queen’s University, Department of Psychology, 62 Arch St., Kingston, ON K7L 3N6, Canada. Tel.: +1 613 533 2491; 1 As the evidence is quite mixed as to whether individuals with autistic disorder, Asperger syndrome, or pervasive developmental disorder—not otherwise specified show differential patterns of development, autism spectrum disorders (ASD) will be used throughout the paper to refer to all individualson the spectrum, unless otherwise specified.
1750-9467/$ – see front matter ß 2010 Elsevier Ltd. All rights reserved.
doi: E. Kelley et al. / Research in Autism Spectrum Disorders 4 (2010) 526–538 functioning well intellectually and academically, they may continue to experience the core deficits of this disorder, such associal and communicative problems ). To date, it has been difficult to address this debate empirically, as manystudies have simply used IQ or school placement as an outcome variable ). As points out, integrated school placement or IQ within the normal range is notsufficient to consider a child to have moved off the autism spectrum. The current study was designed to assess a group ofoptimal outcome children (OO) and compare them to both their typically developing (TD) peers and high-functioning peerswith ASD (HFA) on a wide range of variables in order to better delineate their profiles of functioning. We expected that theHFA group would continue to experience problems in communication, social adjustment and adaptive behavior, whereas theOO group would be functioning within the average range on these measures, but we considered it possible that someresidual cognitive or behavioral difficulties would be seen in the OO group.
Optimal outcome children were those who, as in previous studies, were mainstreamed into regular classrooms and had, according to their records, a full-scale IQ within the average range (i.e. greater than 70). However, we imposed additionalcriteria for optimal outcome: Unlike previous studies, the optimal outcome children in the current study were no longerreceiving extra help in the classroom, had lost their diagnosis according to the school system, and no longer met criteria for adiagnosis on the autism spectrum according to their ADOS-G scores. The HFA group also had a full-scale IQ within the averagerange, according to their records, but, unlike the OO group, retained their diagnosis in the school system and continued tomeet criteria for a diagnosis on the autism spectrum according to the ADOS-G.
1.1. Previous outcome studies in children on the autism spectrum Many of the outcome studies reported in the literature suffer from a restricted set of outcome variables, as pointed out by and Some of these studies report that the children have beenmainstreamed into regular classrooms but do not address whether the children are receiving continued support in terms ofan aide or other extra help (Since in many school systems there is a tendency to mainstream all but the mostimpaired children, we cannot determine with any certainty whether or not these children are truly functioning at the level oftheir typically developing peers Other researchers have argued that while IQ and school placement are of interest, outcome studies need to focus more on how the children are performing communicatively, socially and adaptively because these areas are commonly the mostresistant to treatment That is, there may be children who areperforming fairly well in the academic realm but have difficulty with navigating the social world, effectively communicatingwith others, and being flexible in their behaviors. This study is only the second that we know of (see also ) to address levels of autistic symptomatology, language and communication, adaptive behavior, and problembehaviors in a group of optimal outcome children. However, unlike those studied by Sallows and Graupner, the children inthe current study were tested well beyond the time they had finished treatment; thus, we investigate whether the children’simproved status has been maintained for a number of years.
The current study is not, strictly speaking, an outcome study as it did not follow a sample of children over a specific period of time. However, it is relevant to the outcome literature in that we examined two groups of children with a history of ASDwhose behavioral outcomes differ despite their being matched on non-verbal intelligence at the time of the study. Whilemost researchers seem to agree that early diagnosis and early intervention in children with ASD are critical little agreement exists as to what the best outcomes for children with ASD can be. Thisdisagreement is complicated by the fact that for a long time most outcome studies either tended to focus on very specifictreatment outcomes (e.g., learning of irregular past tense forms) over very short periods of time (see , for areview), or very general outcomes (e.g., ability to live independently in adulthood) ( Only a few studies have examined a wide variety of outcome variables over the course of childhood and these studies have reported mixed results. found that ASD children who were high- or low-functioning in preschoolgenerally remained so into the school years on a wide variety of measures, with some high-functioning children showingimprovement on standard scores of language and cognitive functioning while the low-functioning children tended to show adecline in these same areas. In that study, almost all of the low-functioning children stayed in the low-functioning groupfrom preschool to school-age, while the high-functioning preschool group split into two groups, one of which scored in thelow-functioning range at follow-up and the other (relatively small) had scores in the normal range on most cognitivevariables. The researchers did not investigate the treatments received by the children, however, nor did they report autisticsymptomatology or problem behaviors.
conducted a longitudinal study of a number of children with autism and found considerable variability in the trajectories of development. While overall the group means for intelligence did not change a great dealover time, many of the children showed either marked improvement or a decline in intelligence scores over the course ofthe study. However, the vast majority of the children in this study progressed very little in the language domain andcontinued to remain socially isolated and unable to relate to their peers. Interestingly, Sigman and Ruskin found that 17% ofthese children lost their diagnosis in late childhood or early adolescence. Finally, a few studies have examined outcomeafter several years of behavioral treatment. followed a group of children with ASD who had receivedhome-based behavior modification programs, and found that little change transpired in these children on a variety of E. Kelley et al. / Research in Autism Spectrum Disorders 4 (2010) 526–538 cognitive, language, and adaptive measures approximately two years after beginning treatment in preschool; gains weremade in adaptive skills, though. found that behavioral and sensory symptoms often improvedmore than social and communicative symptoms, and IQ improved more than adaptive functioning between the ages of twoand four.
In sum, the literature is mixed as to what the outcome of children on the autism spectrum is over time. One reason for the disparate findings is likely that many studies included both high- and low-functioning children. These two groups of childrenmay exhibit different developmental trajectories, and thus the effects of change may be occluded by the large variabilitywhich results when the two groups are combined. The current study comprised only initially high-functioning children andthus a clearer picture of outcome for this particular group of children was expected.
For many years the only group of studies that had yielded a large number of optimal outcome children was that of and Although Lovaas reported that 47% of the children who participated in anintensive 40-h-a-week program of behavioral therapy improved to such a degree that they were mainstreamed into regularclassrooms, his methodology has been criticized by some The current study does not directly addresstreatment comparisons as and did, but rather whether or not children truly canimprove to such an extent that they no longer appear to be on the autism spectrum.
More recently, a handful of studies have suggested that an optimal outcome for some high-functioning children with ASD is indeed a possibility. found that 48% of their sample reached optimal outcomes after anintensive program of behavioral therapy; that is, these children had been mainstreamed into regular classrooms and werefunctioning at a normal intelligence level. This study found, however, that general language ability and adaptive behavior didnot increase to the same extent as IQ. Moreover, the children in this study were still relatively young and had only recentlycompleted an intensive behavioral treatment program. The current study extends this work by examining older childrenwho had not been receiving treatment for several years.
demonstrated that some children with ASD reached a level of behavioral and cognitive functioning where they no longer met criteria for any ASD and were mainstreamed into a regular classroomwithout any extra help, though they continued to experience some subtle social and communicative difficulties. That is, theOO children in this study experienced continuing difficulties in semantic and pragmatic areas of language, such asunderstanding the meaning of mental state verbs, understanding theory of mind, engaging in a narrative, and being able toinduce categories properties using semantic information. It should be noted that 11 of the children in the current study werealso part of the Kelley et al. study, although the current study is not a longitudinal study because very different measureswere used in the different studies. Moreover, the Kelley et al. study did not measure autism symptomatology, adaptivefunctioning, or problem behaviors as the current study did.
We hypothesized that the children in the OO group would fall within the normal range of functioning on all measures, including those measures not used in defining them as a group, while the HFA group would continue to show difficulties withadaptive behavior, semantic and pragmatic language skills, as well as adjustment problems.
Three groups of children were examined in this study. After several children were eliminated for the reasons given below, the final sample included 13 children with a history of ASD who had reached an optimal outcome level (the OO group), 14typically developing children (the TD group), and 14 children on the autism spectrum who were of average intelligence (theHFA group).
The OO group was chosen a priori by the following criteria: 1. They had been mainstreamed into a regular classroom.
2. They had a full-scale IQ on their last school assessment which was greater than 70.
3. They were receiving no more than 1 h per week of service overall (e.g. 1 h of speech therapy, or 1 h of occupational therapy) and did not have an educational aide in the classroom.
4. They were considered by the school system to longer be on the autism spectrum.
5. They had previously been diagnosed on the autism spectrum by a clinician who specialized in ASD and they met criteria on the ADI-R based on their ‘ever’ scores.
6. They no longer met criteria for an ASD diagnosis on the ADOS-G.
The HFA group was chosen a priori by the following criteria: 1. If they had been mainstreamed into a regular classroom, they continued to receive extra help. Two of the HFA children were being home-schooled, two were half-time in special needs classrooms and half-time mainstreamed with a one-on-one aide, three were mainstreamed with a one-on-one aide, and seven were mainstreamed with a shared aide.
2. They had a full-scale IQ on their last school assessment which was greater than 70.
3. They retained their diagnosis of an ASD in the school system.
E. Kelley et al. / Research in Autism Spectrum Disorders 4 (2010) 526–538 Table 1Age, gender and non-verbal intelligence of the groups.
4. They had previously been diagnosed on the autism spectrum by a clinician who specialized in ASD and they met criteria on the ADI-R based on their ‘ever’ scores.
5. They currently met criteria for an ASD diagnosis on the ADOS-G.
The typically developing children had no history of academic, neurological, or psychological problems as reported by their parents. Children in all of the groups were from middle- to upper-middle class families that resided in suburban andrural areas of the northeastern US.
All of the parents from our previous study with these optimal outcome children () were contacted and asked to take part in the current study. Two of the optimal outcome children from the previous study were lost toattrition as the parents no longer wished to participate; however, two more were gained through the clinical files of thethird author. One of the children from the Kelley et al. study was not included because, although no longer consideredon the spectrum by the school system, he met the current criteria for an ASD on the ADOS-G. Five of the typicallydeveloping children were lost to attrition but five more were gained through the local school system. These testingsessions took place approximately three years after they had participated in the previous study Finally, the 14 children still on the autism spectrum were obtained through the clinical practice of the third author orwere participants who had previously engaged in research with our lab, and were known to have a full-scale IQ scorewithin the normal range.
A number of children were tested but not included in the current study: one typically developing child and one OO child, each of whom did not want to participate further after the first testing session; two OO children who were too young tocomplete the current protocol; one child thought to be OO, but who met criteria for an ASD diagnosis on the ADOS-G (seeabove); two HFA children who did not meet criteria on the ADOS-G or ADI-R yet still retained a diagnosis from the schoolsystem; and one HFA child who experienced difficulty during both testing sessions and was found to test with below averageintelligence. Information on the children’s age, gender, and non-verbal IQ scores can be found in . There were nosignificant differences between the groups on gender, age, or non-verbal intelligence as measured by the Matrix Reasoningsubtest of the Wechsler Intelligence Scale for Children-IV ( The Autism Diagnostic Interview-Revised (ADI-R) ) is a structured interview conducted with the parent(s) that is based in part on the DSM-IV’s and ICD-10’s diagnostic criteria for autistic disorder. In the currentstudy all parents were instead asked to describe their child’s behavior when their child’s behavior was the most autistic,whatever age that might be. This methodological adjustment was made because many of the optimal outcome children hadalready improved significantly by the age of four or five. Data from one child from the OO group was not available as theparents did not want to be interviewed. Her clinical files, however, showed that she had indeed been on the autism spectrumaccording to the DSM-IV symptoms and clinical reports.
2.2.2. The autism diagnostic observation schedule-generic (ADOS-G) is a standardized diagnostic instrument used to assess and diagnose children on the autism spectrum. Every child was given module 3 as they were all verbally fluent. Data from two children in the OO group wereunavailable due to experimenter error; however, these children met all of our other criteria for the OO group and did notmeet the criteria for an ASD on the ADI-R current functioning diagnostic algorithm. One child in the HFA group wasunwilling to complete the ADOS-G, yet he still clearly met criteria for an ASD on the ADI-R current functioning diagnosticalgorithm.
E. Kelley et al. / Research in Autism Spectrum Disorders 4 (2010) 526–538 The Vineland Adaptive Behavior Scales (is a frequently used parent interview which addresses three major aspects of the child’s adaptive functioning: Communication, Daily Living Skills, and Socialization. Twoparents in the typically developing group did not wish to complete this interview.
The Behavior Assessment System for Children (BASC) ) was designed to address children’s problem behaviors between the ages of 2 and 18. Four of the OO children, two of the TD children and four of the HFA childrendid not have complete BASC data as parents did not return the questionnaire through the mail.
To assess general language verbal understanding, the Similarities, Vocabulary, and Comprehension subscales of the Wechsler Intelligence Scale for Children-Fourth Edition (WISC; ) were used and the Verbal Composite Index(equivalent to verbal IQ) was calculated. Children were also tested on the Peabody Picture Vocabulary Test-Third Edition (PPVT;) to assess vocabulary knowledge. To assess pragmatic language ability, the Making Inferences andFigurative Language subtests from the Test of Language Competence (TLC; ) and the Interpreting Intentionsstandard score was calculated. To address a wider range of pragmatic language abilities, the Test of Pragmatic Language(TOPL: was also administered. Children were also administered the ComprehensiveAssessment of Spoken Language-Third Edition (CELF; ) to assess both semantic and syntacticaspects of language. All tests were administered in the standard procedure outlined in the manuals.
2.2.6. Differences between autism groups on history To assess possible reasons for differences in trajectories between the groups, a number of questions from the ADI-R were explored. We examined whether the OO or HFA groups were more likely to be taking medication and were more likely tohave a history of autism in the family. We also determined from the ADI-R the age at which the child was diagnosed.
Furthermore, we coded the amount and type of intervention the children had received as follows: a zero was coded if thechild had received no ABA therapy, a one was coded if the child had received an eclectic type of therapy (which may have hadsome ABA included), a two was coded if the child had received primarily ABA therapy but for fewer than 20 h per week, and athree was coded if the child had received more than 20 h per week of ABA therapy.
Informed consent was obtained from all parents and written assent was obtained from all children in the study before testing began. This research study was approved by the university’s Institutional Review Board.
The children were administered the tasks during one of two sessions in which the children were given the battery tasks in random order (as well as three experimental tasks that are not reported here). In order to make the testing proceed as quicklyas possible, children were videotaped and videotapes were scored after the fact. It should be noted that where missing dataoccurred it was never because the child scored at floor; rather, in most instances the testing sessions ran too long andchildren could not complete the final test or there was experimenter error or videotape malfunction. A demographic surveyincluding information about family structure, other children affected by ASD or learning or psychiatric difficulties, and SESwas gathered from all of the families.
To determine the differences between the three groups on the dependent measures, one-way ANOVAs were run in SPSS.
Alpha level was set at .01 as our sample size was relatively small. Moreover, effect sizes are reported for all ANOVAs to ensurethat differences between groups did not occur by chance. Scheffe’s post-hoc tests were conducted to determine specificgroup differences; the alpha for these tests was set at .01 to correct for multiple comparisons. Effect sizes are also reported forthe ANOVAs to demonstrate that effect sizes in most cases are quite large; we believe this justifies our use of a relativelyliberal multiple-comparison correction.
Given that our samples were determined in part by the OO group not meeting criteria on the ADOS-G and the HFA group still meeting criteria, it is not surprising that there were highly significant differences between the groups on theCommunication (F(1,22) = 62.85, p < .001) and Social Interaction (F(1,22) = 67,87, p < .001) subscales.
However, there were no significant differences on the ADI-R algorithm scores, which are based on earlier functioning (at the times when the child showed the highest level of symptoms in this study, as mentioned in Section The QualitativeAbnormalities in Reciprocal Social Interaction subscale showed no significant difference between the groups (OO = 20.5 (2.11), E. Kelley et al. / Research in Autism Spectrum Disorders 4 (2010) 526–538 F-test, significance, effect sizeand group comparison Scales daily living skills standard score HFA = 21.64 (3.23); F(1,25) = 1.10, n.s.), nor did the Qualitative Abnormalities in Communication subscale (OO = 18.58 (2.47),HFA = 17.86 (4.37); F(1,25) = .260, n.s.) or the Restricted, Repetitive, and Stereotyped Patterns of Behavior subscale (OO = 5.00(2.45), HFA = 5.50 (2.44); F(1,25) = .298, n.s.). On the Current Behavior Algorithm, however, which assesses currentfunctioning, the expected large differences between the groups were found and confirmed the differences in autisticdiagnostic status between the groups. On the Qualitative Abnormalities in Reciprocal Social Interaction subscale the OO grouphad an average of 2.00 (2.52) and the HFA mean was 8.36 (3.78) (F(1,25) = 26.07, p < .001; partial h2 = .51), on the QualitativeAbnormalities in Communication subscale the OO group had a mean of 1.08 (1.50) and the HFA mean was 6.36 (3.88)(F(1,25) = .21.15, p < .001; partial h2 = .46.) and on the Restricted, Repetitive, and Stereotyped Patterns of Behavior subscale theOO group had a mean of 1.08 (1.19) and the HFA mean was 3.07 (2.13) (F(1,25) = 8.838, p < .01; partial h2 = .26).
Unlike previous research with high-functioning children with ASD, the three groups scored within the average range on all subscales except Socialization. On the Socialization scale, as would be expected, the HFA group mean did fall in theimpaired range. It should be noted, however, that in both the OO and HFA groups there were individual children who fell inthe impaired range on all three subscales. While there were no significant differences between the groups on the DailyLiving Skills subscale, the OO group was not significantly different from the TD group on both the Communication andSocialization subscales, and the OO group was significantly higher on the Socialization subscale than the HFA group( The BASC t-scores can be seen in None of the t-score means for any of the groups fell within the clinical range. It should be noted, however, that the HFA group’s means were within the at-risk range on the Adaptability,Atypicality, Withdrawal, Social Skills, and Leadership measures. The OO group was just into the at-risk range on Attentionproblems. Again, however, it is important to note that individual children in both the OO and HFA groups fell within theclinical range on some of these subscales. Unlike previous research (e.g. , there were nosignificant differences between the groups on Hyperactivity, Aggression, Anxiety, Somatization, and Adaptability. On anumber of other subscales, the OO group was not significantly different from the HFA group (i.e., Conduct Problems,Depression, Atypicality, and Withdrawal); however, only the Atypicality subscale showed any signs of elevation from thenormal range.
The OO group fell within the normal range on all of the language and pragmatic language measures (see While the OO group performed significantly better than the HFA group on the PPVT, WISC, TOPL and TLC, no firm conclusions couldbe made about the CELF-3 composite measures as the group differences between the OO and HFA group were not significant.
However, as can be seen in , the OO group did perform at a significantly higher level than the HFA group on the WISC-IV Vocabulary and Comprehension subtests, as well as on the Concepts and Directions and Formulated Sentences subtests ofthe CELF-3. Moreover, while the HFA group had several individuals who fell well within the impaired range on individualsubtests, the majority of the OO group fell well within the normal range.
E. Kelley et al. / Research in Autism Spectrum Disorders 4 (2010) 526–538 Partial h2 = .30No significant group differences a Higher Scores indicate more problematic behavior; scores from 60–69 clinical at-risk range, >69 clinically significant range.
b Lower Scores indicate more problematic behavior; scores from 31–40 clinical at-risk range, <31 clinically significant range.
* p < .01.
** p < .001.
3.5. Differences between autism groups on history Demographic comparisons obtained from the ADI-R interview were used to investigate differences between the OO and HFA groups. The OO group was found to has been diagnosed approximately one year earlier than the HFA group (see but the difference between the groups was only approaching significance. Chi-square tests of significance were conducted todetermine if there were statistical differences between the OO group and the HFA group on measures of history of ASD withinthe family, whether the children were currently on medication, and an approximate measure of what type of interventionthey had received (see ). There were no significant differences between the groups on the measures of family historyof autism or current medication use. The categorical measure of intervention obtained from the ADI-R history showed thatthe OO group was significantly more likely to have received an intensive program of behavioral intervention than the HFAgroup.
The purpose of this study was to investigate how children who had been diagnosed with ASD as toddlers, but now in late childhood no longer carried an ASD diagnosis, compared with typically developing children and high-functioning childrenwho still carried the ASD diagnosis on early and current autism symptoms, adaptive functioning, problem behaviors, andlanguage and communication skills. Moreover, we examined some aspects of the OO and HFA children’s interventions in apreliminary attempt to ascertain why they developed different outcomes. We consider each of these in turn.
E. Kelley et al. / Research in Autism Spectrum Disorders 4 (2010) 526–538 Table 4Language and communication composite measures.
There were no differences between the OO and HFA groups on the number of overall autism symptoms in early childhood as measured by the ADI-R. This is consistent with the results of , who found that few symptoms or skillsobserved at age 2 predicted movement off the spectrum at age 4. Of course, there were significant differences between thegroups on their ADOS scores given that one way the groups were ascertained was by choosing children for the OO group thatno longer met diagnostic criteria on this measure. The HFA group also had higher scores than the OO group on all thesubscales of the ADI-R Current Diagnostic Algorithm, which is consistent with the fact that the former children are still on thespectrum.
The OO group scored within the normal range on all subscales of the Vineland, and did not differ significantly from the TD group on any subscale. In contrast, as expected, the HFA group scored significantly lower than the TD group on the VinelandCommunication and Socialization subscales. Interestingly, though, only on the Socialization subscale did the OO group scoresignificantly higher than the HFA group. There were no differences between the groups on the Daily Living Skills subscale. TheHFA group was within the normal range on the Communication subscale but the Socialization subscale was in the impairedrange, which is consistent with previous studies reported than socialization is generally lower than expected for mental agein children with ASD The OO group, however, appears to have overcome their difficulties on this measure,at least according to parent report.
used the BASC to assess the level of functioning of a group of individuals with Asperger syndrome.
Although they found no differences between the Asperger group and the typically developing group only on Aggression andSomatization, we found no significant differences between the groups on those subscales plus Hyperactivity, Anxiety andAdaptability, perhaps because of our more stringent alpha criterion. The OO group scored within the normal range on all of the E. Kelley et al. / Research in Autism Spectrum Disorders 4 (2010) 526–538 subscales of the BASC, although they fell in between the TD and HFA groups on a number of measures and were borderline at-riskon the Attention subscale. The borderline at-risk attention scores are consistent with the existence of a subgroup of optimaloutcome children who present with ADHD as they lose their autism symptoms The HFA group,in contrast, was in the at-risk range for their mean t-scores on the Adaptability, Atypicality, Withdrawal, Social Skills, andLeadership subscales, indicating that they continue to experience behavioral difficulties in these areas.
It should be noted, however, that the OO group was not significantly different from the HFA group on Conduct Problems, Depression, Atypicality, and Withdrawal. Moreover, on all of these subscales we were unable to determine (using our strictalpha criterion) whether or not the OO group was statistically distinguishable from the TD group. Because some of the OOchildren’s scores were elevated in the current study at 8–13 years of age, it is quite possible that problems with internalizingor externalizing behaviors might emerge as they enter adolescence. Furthermore, these data are based on a parent checklistrather than a parent clinical interview, direct psychiatric evaluation or other informant information, and these parents mightbe invested in a fully ‘‘normal’’ outcome for their children.
In summary, although the individuals with high-functioning ASD in this sample are continuing to display adjustment difficulties in middle childhood and early adolescence, for the most part the individuals with optimal outcome areexperiencing few adjustment problems, with the exception of attention difficulties.
The OO group did not perform significantly differently from the TD group on all the composite measures of language and pragmatic language given, and performed at a significantly higher level than the HFA group on all composite measures E. Kelley et al. / Research in Autism Spectrum Disorders 4 (2010) 526–538 4 Strattera, Concerta, Zoloft, and Adderall 7 Ritalin, 2 Adderall, 3 Prozac, and Risperidol except for the CELF. This is in contrast to the previous study, which included some of these OO children, which had found thatthe OO group experienced pragmatic difficulties (The current finding also contrasts with those of who found that language abilities were not improved to the same extent in their optimal outcomegroup as behavioral and cognitive measures. However, the OO children here are a few years older than those in the Sallowsand Graupner sample, which suggests that language abilities may continue to improve in optimal outcome children longafter treatment has been terminated.
It is worth noting here that the HFA group also scored within the normal range on all language measures, despite performing more poorly than the other two groups. Although their TOPL and TLC scores were significantly lower than thosein the TD and OO groups, and were the low points in their profile (an expected finding since they assess pragmatic languageabilities), their group mean did fall within the low normal range.
A slightly different pattern of findings emerged when individual subtests of the composite language tests were examined.
The OO group scored significantly higher than the HFA group on the WISC-IV Vocabulary and Comprehension, as well as theCELF-3 Concepts and Directions and Formulated Sentences. Moreover, while there were several individual children in theimpaired range on each of the language subtests in the HFA group, the OO group included no children in the impaired rangeon most subtests, and only one or two children in the impaired range on the CELF-3 Recalling Sentences and Listening toParagraphs subtests, and on the Comprehension subtest of the WISC-IV. This pattern of findings indicates that while manychildren in the HFA group may still be experiencing subtle language and communication impairments, the children with OOappear mostly unimpaired, with subtle difficulties only in remembering orally presented information, as well as in the moregeneral social knowledge that is assessed by the Comprehension subtest of the WISC.
4.5. Differences between autism groups on history There were no differences between the HFA and OO groups on having a history of ASD in the family or currently being on medication. Except for one HFA child on Risperidol, the medications were all for attention problems or depression. Althoughthere were no statistically significant differences between the groups on the age at diagnosis, this is probably due to the smallsample size of the current study as the group means were approximately one year apart and there was a trend towardsignificance (p = .095). The OO group was significantly more likely to have received intensive behavioral intervention, whichis likely to have contributed to their optimal outcome. Given the retrospective and categorical nature of the measure of earlyintervention, as well as the small sample size, however, this hypothesis must be interpreted with caution. It is also worthnoting that both the OO and HFA groups contained several children who were currently taking psychoactive medications ofvarious sorts (most for attention or depression/anxiety). Thus, while the OO group seemed to be functioning within theaverage range according to their parents on problem behaviors, it is clear that at least a subset of these youth experienceproblems with attention or depression and anxiety.
4.6. Comparisons with previous outcome studies This study is comparable to those that find there are individuals with ASD who go on to achieve relatively normal functioning (Unlike those other studies, however, thecurrent study cannot address what proportion of children has the potential to reach this optimal outcome. Moreover,because of the retrospective nature of this study we are unable to pinpoint precisely which early variables predicted thechildren’s membership in either the OO or HFA groups. There was an indication of treatment differences between the OO and E. Kelley et al. / Research in Autism Spectrum Disorders 4 (2010) 526–538 HFA groups but a more in-depth measure would need to be collected to give more weight to the findings, and of course aprospective study would be ideal for determining the effect of intervention on outcome.
A major strength of the current study is the wide variety of outcome measures given to the children. Several authors have criticized the field of outcome research for focusing mainly on IQ tests (Moreover, few studies have addressed both symptomatology and social and adaptive functioning as the current study did). By focusing on these measures we can show that although all children in this study showedintelligence within the average range, the HFA group continued to display residual autism symptoms, including problems insocial relationships, communication problems, atypical behaviors, and pragmatic language difficulties. In contrast, the OOgroup showed no group elevations on BASC scores, and their social and communication patterns were well within the normalrange. Granted, these results were from parent checklist responses and were filled out for the most part before adolescence.
We are continuing to follow these children into adolescence and in fact are observing elevated rates of anxiety anddepression in the OO children, which might be due to their advancing into adolescence and/or the use of more sensitivemeasures.
Another strength of this study is that it was undertaken several years after the OO children had stopped receiving treatment and was thus able to show that the children were continuing to display an optimal outcome. Moreover, it is thefirst study that we know of in which the outcome assessment was not conducted by the same researchers who conducted thetreatment; we had little vested interest in the outcome of this study.
This study was not able to address on any level what percentage of children diagnosed with an ASD may reach this optimal outcome level. Although the large majority of the children came from the clinical files of the third author, there is noway to ascertain what proportion of children with ASD these children represent.
The children were assigned to one of four categorical levels with regard to the type of treatment they had received as young children. This was a somewhat crude measure of intervention, though obtaining accurate and specific informationretrospectively is very difficult (). Some have argued that the type of intervention thechild receives is not as important as that s(he) receives it early and intensively (). Moreover, many haveasserted that not all interventions work for all children and there is a need for research on types of interventions other thanbehavioral interventions The current study was not able to address these issues, particularlybecause most of the children received at least some level of behavioral intervention (ABA).
Lack of power was also an issue in this study. While this was inevitable due to the relative scarcity of optimal outcome children, it does mean that the data must be interpreted with some caution. We did our best to ensure that the groupdifferences obtained were indeed group differences by only attributing group differences to measures that showed an alphaof .01 or lower. While this is not a complete solution, we believe that the consistent pattern of results across all measures(where the OO group means were higher than those of the HFA group means and closer to the TD group means) indicates thatthe children in the OO group are indeed performing quite well on most of the measures.
Mental retardation is present in more than half of children with ASD (). Relationships between variables found in this study may not be at all applicable to low-functioning children with ASD as these children likely follow differentdevelopmental pathways (). However, it has been suggested that low-functioning and high-functioningchildren with ASDs be studied separately as they should be conceptualized as substantially different disorders (It should be noted that the HFA group studied here were for the most part quite mildly affected. Had we used an HFAgroup with a higher level of autism symptoms, presumably we would have found greater differences between the OO andHFA groups.
The optimal outcome children in this study have, to a large extent, overcome their social, communicative, and behavioral difficulties and have become engaged in the social world. Although the OO children had lower group means than the TDgroup on a few of the standardized tests and behavioral measures, their group means were still well within the normal rangeon all measures. How they and their families and therapists accomplished this cannot be answered definitively through aretrospective study. The treatment data, however, suggest that the proportion of children receiving intensive behavioraltreatment was a significant factor: eight of the 13 OO children for whom we had treatment data (62%), but only two of the 14HFA children (13%) received more than 20 h per week of ABA treatment, and all of the OO children experienced some form ofABA intervention. We do not mean to suggest that ABA treatment alone can accomplish this outcome; clearly, there are childcharacteristics, including most importantly, potential for normal IQ and lack of an intractable language disorder, thatprobably are necessary as well (see ; and , for discussion of this issue).
Whether the difference between the OO and HFA children is a categorical difference or a matter of degree remains to be determined; the sample sizes in this study are too small and too variable to clarify this issue. However, it is apparent thatthese OO children have experienced something more dramatic than simply a reduction in symptoms across development(e.g. ). While both the OO and HFA groups demonstrated verbal ability within the normal range, theHFA group was found to experience continuing difficulty to a much greater extent than their OO peers, who were E. Kelley et al. / Research in Autism Spectrum Disorders 4 (2010) 526–538 indistinguishable from the TD group on many measures and within the average range on all of them. Further research isneeded to address more subtle aspects of social-cognitive functioning, including how the children interact with others.
Another issue that is clear from the current study is that there is a definite change in the symptom profile of individuals with high-functioning ASD across development. This study lends support to research that suggests that many children withHFA improve to a great extent in middle childhood, even if they do not attain optimal outcome status ). The fact that the HFA group means were within the normalrange on all of the language measures and several of the BASC subscales suggests that at least some high-functioningchildren with autism may improve much more in late childhood and early adolescence than was previously thought.
What does the existence of this optimal outcome group tell us about ASD and about development? It can be argued that optimal outcomes in these children suggest that the course of this disorder may be halted or even reversed, at least for high-functioning children with ASD. The current study does not provide an answer to how this change in course wasaccomplished; prospective studies are needed to address this issue. The existence of these optimal outcome children wouldalso seem to suggest that atypical development, and development itself, can be altered significantly by the environment.
While there are limiting factors (and mental retardation in children with ASD is probably the most common and obviouslimiting factor), for some children the trajectories of development can be altered quite dramatically (see fortheoretical speculations about the mechanisms by which behavioral treatment can accelerate development and reversesymptoms).
It is clear that at least some children with ASD are able to achieve an excellent outcome; further research must delineate the pathways by which this outcome is achieved. Perhaps one day optimal outcomes will be, at least for higher-functioningchildren, more common than the relatively rare occurrence that they are today. Until then it is important to gain as muchinformation about these children as possible, to determine which factors have placed them on such a positive developmentaltrajectory.
We would like to thank all of the families who gave so generously of their time to participate in this research and all of the undergraduate research assistants who helped collect and enter this data.
Beglinger, L., & Smith, T. (2005). Concurrent validity of social subtype and IQ after early intensive behavioral intervention in children with autism: A preliminary investigation. Journal of Autism and Developmental Disorders, 35, 295–303.
Bibby, P., Eikeseth, S., Martin, N. T., Mudford, O. C., & Reeves, D. (2002). Progress and outcomes for children with autism receiving parent-managed intensive interventions. Research in Developmental Disabilities, 23, 81–104.
Billstedt, E., Gillberg, C., & Gillberg, C. (2005). Autism after adolescence: Population-based 13- to 22-year follow-up study of 120 individuals diagnosed with autism in childhood. Journal of Autism and Developmental Disorders, 35, 351–360.
Charman, T., & Howlin, P. (2003). Research into early intervention for children with autism and related disorders: Methodological and design issues. Autism, 7, Dawson, G. (2008). Early behavioral intervention, brain plasticity, and the prevention of autism spectrum disorder. Development and Psychopathology, 20, 775– Dunn, L. M., & Dunn, L. M. (1997). Peabody picture vocabulary test (3rd ed.). Circle Pines, MN: American Guidance Services.
Eaves, L., & Ho, H. H. (1996). Brief report: Stability and change in cognitive and behavioral characteristics of autism through childhood. Journal of Autism and Developmental Disorders, 26, 557–569.
Eaves, L., & Ho, H. H. (2004). The very early identification of autism: Outcome to age 4(1/2)–5. Journal of Autism and Developmental Disorders, 34, 367–378.
Fein, D., Stevens, M., Dunn, M., Waterhouse, L., Allen, D., Rapin, I., et al. (1999). Subtypes of pervasive developmental disorder: Clinical characteristics. Child Fein, D., Dixon, P., Paul, J., & Levin, H. (2005). Brief report: Pervasive developmental disorder can resolve into ADHD: Case illustrations. Journal of Autism and Developmental Disorders, 35, 525–534.
Fombonne, E. (2005). The changing epidemiology of autism. Journal of Applied Research in Intellectual Disabilities, 18, 281–294.
Goldstein, H. (2002). Communication intervention for children with autism: A review of treatment efficiency. Journal of Autism and Developmental Disorders, 32, Helt, M., Kelley, E., Kinsbourne, M., Pandey, J., Boorstein, H., Herbert, M., et al. (2008). Can children with autism recover? If so, how?. Neuropsychology Reviews, 18, Howlin, P., Goode, S., Hutton, J., & Rutter, M. (2004). Adult outcome for children with autism. Journal of Child Psychology and Psychiatry, 45, 212–229.
Kasari, C. (2002). Assessing change in early intervention programs for children with autism. Journal of Autism and Developmental Disorders, 32, 447–461.
Kelley, E., Paul, J., Fein, D., & Naigles, L. (2006). Residual language deficits in optimal outcome children with a history of autism. Journal of Autism and Developmental Klinger, L. G., & Renner, P. (2000). Performance-based measures in autism: Implications for diagnosis, early detection, and identification of cognitive profiles.
Journal of Clinical Child Psychology, 29, 479–492.
Lindner, J. L., & Rosen, L. A. (2006). Decoding of emotion through facial expression, prosody, and verbal content in children and adolescents with Asperger’s syndrome. Journal of Autism and Developmental Disorders, 36, 769–777.
Liss, M., Harel, B., Fein, D., Allen, D., Dunn, M., Feinstein, C., et al. (2001). Predictors and correlates of adaptive functioning in children with developmental disorders. Journal of Autism and Developmental Disorders, 31, 219–230.
Lord, C., Rutter, M., & LeCouteur, A. (1995). Autism diagnostic interview (Revised (3rd ed., Short Form)). Chicago: University of Chicago Press.
Lord, C., Risi, S., Lambrecht, L., Cook, E. H., Leventhal, B. L., DiLavore, P. C., et al. (2000). The autism diagnostic observation schedule-generic: A standard measure of social and communicative deficits associated with the autism spectrum. Journal of Autism and Developmental Disorders, 30, 205–223.
Lovaas, I. O. (1987). Behavioral treatment and normal educational and intellectual functioning in young autistic children. Journal of Consulting and Clinical McEachin, J. J., Smith, T., & Lovaas, I. O. (1993). Long-term outcome for children with autism who received early intensive behavioral treatment. American Journal on Mental Retardation, 97, 359–372.
E. Kelley et al. / Research in Autism Spectrum Disorders 4 (2010) 526–538 Mundy, P. (1993). Normal vs. high-functioning status in children with autism. American Journal of Mental Retardation, 97, 381–384.
Mundy, P., & Crowson, M. (1997). Joint attention and early social communication: Implications for research on intervention with autism. Journal of Autism and Developmental Disorders, 27, 653–676.
Mundy, P., & Stella, J. (2000). Joint attention, social orienting, and non-verbal communication in autism. In A. M. Wetherby & B. M. Prizant (Eds.), Autism spectrum disorders: A transactional developmental perspective (pp. 55–77). Baltimore, MD: Brookes.
Phelps-Terasaki, D., & Phelps-Gunn, T. (1992). Test of pragmatic language. Austin, TX: Pro-Ed.
Reynolds, C. R., & Kamphaus, R. W. (1992). Behavior assessment system for children. Circle Pines, MN: Academic Guidance Services.
Sallows, G. O., & Graupner, T. D. (2005). Intensive behavioral treatment for children with autism: Four-year outcome and predictors. American Journal on Mental Schopler, E., Short, A., & Mesibov, G. (1989). Relation of behavioral treatment to normal functioning: Comment on Lovaas. Journal of Consulting and Clinical Semel, E., Wiig, E. H., & Secord, W. A. (1995). Clinical evaluation of language fundamentals (3rd ed.). San Antonio, TX: The Psychological Corporation.
Shonkoff, J. P., Hauser-Cram, P., Krauss, M. W., & Upshur, C. C. (1988). Early intervention efficacy research: What have we learned and where do we go from here? Topics in Early Childhood Special Education, 8, 81–93.
Sigman, M., & Ruskin, E. (1999). Continuity and change in the social competence of children with autism, down syndrome, and developmental delays. Monographs of the Society for Research in Child Development 64 (1, Serial No. 256).
Sparrow, S. S., Balla, D. A., & Cicchetti, D. V. (1984). Vineland adaptive behavior scales. Circle Pines, MN: American Guidance Service, Inc.
Starr, E., Szatmari, P., Bryson, S., & Zwaigenbaum, L. (2003). Stability and change among high-functioning children with pervasive developmental disorders: A 2- year outcome study. Journal of Autism and Developmental Disorders, 33, 15–22.
Stevens, M. C., Fein, D. A., Dunn, M., Allen, D., Waterhouse, L. H., Feinstein, C., et al. (2000). Subgroups of children with autism by cluster analysis: A longitudinal examination. Journal of the American Academy of Child and Adolescent Psychiatry, 39, 346–352.
Sutera, S., Pandey, J., Esser, E. L., Rosenthal, M. A., Wilson, L. B., Barton, M., et al. (2007). Predictors of optimal outcome in toddlers diagnosed with autism spectrum disorders. Journal of Autism and Developmental Disorders, 37, 98–107.
Szatmari, P., Bartolucci, G., Bremner, R., Bond, S., & Rich, S. (1989). A follow-up of high-functioning autistic children. Journal of Autism and Developmental Disorders, Tsatsanis, K. D., Foley, C., & Donehower, C. (2004). Contemporary outcome research and programming guidelines for Asperger syndrome and high-functioning autism. Topics in Language Disorders, 24, 249–259.
Wechsler, D. (2003). Wechsler intelligence scale for children (4th ed.). San Antonio, TX: The Psychological Corporation.
Wiig, E., & Secord, W. (1989). Test of language competence-expanded edition. San Antonio, TX: The Psychological Corporation.

Source: http://cll.uconn.edu/Papers%20pdf/Project2/Kelley,_Naigles,_&_Fein,_2010.pdf