Long term outcome for children with autism who received ...

AmerlcalljouNIal 011Mellta/ Retardattoll
199', Val. 97, No. ~, '59-'72
e 199' American Ao.,)CI.tJon on Mental Retard.tJol1

Long- Term Outcome for JohnJ. McEachln, T..istram SmJth, and O.lvar
Children With Autism Who Lovaas
Received Early Intensive Univ~rsity of California, Los Ang~l~s
Behavioral Treatment

After a very tntenstve behavtoral tnterventton, an e.xpeirtmental group oJr19
preschool-age chtldren with auttsm achteved less restrl(:;ttve school place~nents and
htgher IQs than dtd a control group of 19 stmtlar chtla'ren by age 7 (Lovaas,
1987) .The present study followed-up thts ftnding by a.I;sesstngsubjects at a mean
age of 11.5 years. Results showed that the experimental group preseroed tts gains
over the control group. 11Je9 experimental subjects whlo had achteved tl,e best
outcomes at age 7 received parttcularly extenstve eval~~attons tndtcattn~~ that 8 of
them were tndtsttngutshable from average chtldren on tests of tntelltgen,ce and
adapttve behavtor. Thus. behavioral treatment may pn)duce long-Iasttn,g and
stgntftcant gatns for many young chtldren wtth auttsm.

Infanttle auttsm is a condition to be extremely poor CLotter, 1978). For

marked by severe impairment in intellectual, example, in the longest prospective follow-

\~) social, and emotional functioning. Its onset up study witJi1a sound method,ological de-
OCClKSin infancy, and the prognosis appears sign, Rutter (1970) found that only 1 of 64

subjects with autism (fewer than 2%) could

I This study was supported by Grant No. MH- be considered free of clinical11'significant
11440from the National Institute of Mental Health. problems by adulthood, as evidenced by
The study was basedon a dissertation submitted holding a jo,b, living independently, and

to the University of California, Los Angeles, maintaining :!n active and age..appropriate

Department of Psychology' in partial fulfillment sodal life. 111eremaining subjt:cts showed

of the requirements for the doctoral degree. The numerous dysfunctions, such as marked
authors express their deep appreciation to the oddities in behavior, social isolation, and
many students at UCLA who served as therapists florid psychopathology .The majority of sub-
and helped to make this study possible. Special
jects required supervised living conditions.
i thanks to Bruce Baker and Duane Buhrmester 1 Professionals have attempted a wide
who helped in the design of this study. Requests
variety of interventions in an effort to help
i for reprints of this article, copies of the Clinical
Rating Scale,or additional information about this children witl:l autism. For mar:ly years, no

study should be sent to 0. Ivar Lovaas, 405 scientific evidence showed that any of these

Hilgard Ave., UCLA, Department of Psychology, interventioru; brightened the children 'slong-

Los Angeles, CA 90024-1563. tenn progno:5isCDeMyeret al., 1981).How-

McEachin, Smith, and lovaas 359

ever, since the 19605,one of these interven- Also, as 'Wasstated in the first follow-up
tions, behavioral treatment, has appeared
(Lovaas. 1987)."Certain residual deficits may
promising. Behavioral treatment has been remain in the normal-functioning group that
found to increaseadaptivebehaviors such as cannot be detected by teach,ersand parents
and can only be isolated on closer psycho-
language and social skills, while decreasing

disruptive behaviors such as aggression logical assessment,particularly asthese chil-
(DeMyer,Hingtgen,&Jackson,1981;Newsom dren gro~v oldern (p. 8). This possibility
& Rincover, 1989; Rutter, 1985). Further- points to tJheneed for amore I:letailedassess-
more, behavioral treatment has been con- ment and for continued foUow-ups of the
tinuously refined and improved asa result of group OVf!rtime.
ongoing researchefforts at a number of sites
11lepresent investigation contained two
(Lovaas & Smith, 1988).
parts: In dle first part we exaJ11inedwhether
Some recent evidence has indicated severalyearsafter the evaluation at age7, the
that behavioral treatment has developed to
the point that it can produce substantial experimental group in Lovaas's(1987) study
had maint:a.inedits treatment gains. Subjects
improvements in the overall functioning of in the experimental group and one of the
young children with autism (Simeonnson,
Olley, & Rosenthal, 1987). Lovaas (1987) control groups completed sta'l1dardizedtests
of intellectual and adaptive functioning. The
provided approximately 40 hours per week
of one-on-one behavioral treatment for a groups were then contrastedv..itheachother,
and their current performance was com-
period of2 yearsor more to an experimental pared to their performance cln previous as-
sessments,1lle second part of the investiga-
group of 19 children with autism who were tion focus:ed on those subjects who had
under 4 years of age.This intervention also
achieved tJhebest outcome at the end of first
included parent training and mainstreaming
into regular preschool environments. When
re-evaluated at a mean age of 7 years, sub- grade in the Lovaas (1987) st1L1d(yi.e.. the 9
jects in the experimental group had gained subjects w'ho were classified ~~nsormal func-
an average of 20 IQ points and had made tioning OUltof the 19 in tht! experimental
major advancesin educational achievement group). We examined the eJ[tent to which
Nine of the 19subjectscompleted first grade these best,-outcome subjects could be con-
in regular (nonspecial education) classes sidered fr(~eof autistic symptomatology .A
entirely on their own and had IQs that test battery was constructed to assess a
increased to the averagerange. By contrast, variety of possible defidts: for example,

two control groups totalling 40 children, also idiosyncraltic thought pattern:).mannerisms.
diagnosed as autistic and comparable to the and inter~;ts; lack of close relationships with

experimental group at intake, did not fare family and friends; difficulty iJ:lgetting along
nearly as well. Only one of the control with peoplle; relative weaknesses in certain

subjects (2.5%) attained nom1al levels of areas of cognitive functioning, such as ab-

intellectual and educational functioning. stract reasoning; not working up to ability in

Thesedatasuggestthat behavioral treat- school; flatness of affect; ab~;enceor pecu-
ment is effective. However, the durability of liarity in s(:nseof humor. Possible strengths

treatment gains is uncertain. In one prior to be identified included norrnal intellectual

major study, Lovaas,Koegel, Simmons, and functionin;g. good relationships with family
Long (1973) found that children with autism members. :lbility to function independently.
regressedfollowing the termination of treat- appropriate use of leisure time, and ad-
ment. Other studies have shown that chil-
dren with autism may display increased dif- equate so(:ialization with pel~rs.Numerous
ficulties when they enter adolescence
methodological precautions were taken to
(Kanner, 1971; Waterhouse & Fein, 1984). ensure ob}ectivity of the follo'w-up examina-
tion.

360 Autism and Early Intervention

Method treatment were comparable to dlildren with

autism seen elsewhere and (b) the minimal

Subjects and Background treatment pro'l'ided to the first control group

Characteristicsof the subjects and their did not alter intellectual functiolrling.

Statisticalanalysis of an extensive range
treatment have been described elsewhere of pretreatment measuresconfirI1rledthat the

(Lovaas, 1987) and will only be summarized experimental group and control ,group were
here. The initial treatment study contained comparableat intake and closely Irnatchedon
38 children who, at the time of intake, were such important variables as IQ and severity
very young (less than 40 months if mute, less of disturbancf:. The mean chronological age
~ than 46 months if echolalic) and had re- (CA) at diagnosis for subjects in the experi-
~

c~ived a diagnosis of autism from a licensed mental group was 32 months. Thl~irmean IQ
clinical psychologist or psychiatrist not in- was 53 (range'30 to 82; all IQs are given as
volved in the study. These 38 subjects were deviation scores). The mean CA of subjects
divided into an experimental group and a in the control group was 35 months; their
control group. The assignment to groups meanIQ was'i6 (range 30to 80). Most of the
was made on the basis of staff availability. At subjects were mute, all had gro:)Sdeficien-
the beginning of each academic quarter, des in recepti1felanguage, none:played with
treatment teams were formed. The clinic peers or sho\lfed age-appropriate toy play,
director and staff members then determined all were emotionally withdrawn, most had
whether any opening existed for intensive
treatment. If so, the next referral received severe tantrurns, and all showed extensive
would enter the experimental group; other- ritualistic and stereotyped (self-stimulatory)
wise, the subject entered the control group. behaviors. Thus, they appeared to be a
The experimental group contained 19 chil-
dren who received 40 or more hours per representative'sample of childrf:n with au-
week of one-to-one behavioral treatment for tism (Lovaas, Smith, & McEachiJrl,1989). A
2 or more years. The control group was more complel:e presentation of the intake
data was rep<J'rtedby Lovaas (19187).
comprised of 19 children who received a
much less intensive intervention (10 hours a The childlren in the experim,entalgroup

week or less of one-to-one behavioral treat- and control gJroupreceived their respective
ment in addition to a variety of treatments
treatments from trained studenlt therapists
who worked ilrlthe child's home. 'llie parents
also worked 'with their child, alrld they re-

provided by community agencies, such as ceived extensiveinstruction and :;upervision
parent training or special education classes). on appropriat4~treatment techniques. when-
The initial study also included a second ever possible. the children werf~integrated
control group, consisting of 21children with into regular preschools. The trf~atmentIo-
autism who were followed over time by a cused primal1ily on developing language,

nearby agency but who were never referred increasing SOI:ialbehavior, and promoting
for this stlldy. However, these 21 subjects
were not available for the present investiga- cooperative pJlaywith peers alonl~with inde-
pendent and 'appropriate toy pl;ly, Concur-
tion. On standardized measures of intelli- rently, substantial efforts were directed at
gence, the second control group did not decreasing excessive rituals, tarltrums, and
differ from either the experimental group or aggressive behavior. (For a m(]lre detailed
the first control group at intake, nor did it description of the intervention plrogram, see
differ from the first control group when the treatmentmanual (Lovaaset a:l.,19801and

evaluated again when the subjects were 7 instructional videotapes that supplement the
years old. These findings suggest that, as manual (Lova:~ & Leaf, 19811.)
measured by standardized tests,(a) the chil-
At the time of the presenIt follow-up
dren with autism who were referred to us for (1984-1985), Ibe mean CA of the experimen-

McEachin, Smith, and Lovaas 361

~

tal group children was 13 years (range ~, 9 to ~,ithintake IQ or outcome IQ. ConsequeJ

19 years). A!I c?ildren who had achileved a.lthoughthetendencyforthefirstreferra

normal functionIng by the ~geof7 year~h; ad enter the experimental group created a
ended treatment by that point. (Normaljunc- t(~ntialbias, the data indicate that this
ttontng was operationally defined asscoring unlikely;
within the normal range on standardized

intelligence tests and successfully complet- Procedure
ing first grade in a regular, nonspedal edu-
cation class entirely on one's own.) On the The assessment procedure inclu
other hand, some of the children who had a:;certaining school ]placementand admi
not achievednormal functioning at 7yearsof tf'ring three standardized tests. Informal
o:rl school placement was obtained fJ
age had, at the request of their parents, S\lbjects' parents, ~/ho classified then:
remained in treatment. The length of time b,eingin either a reB~lar or a special ed,
that experimental subjects had been O'lt of tilDn class (e.g., a l:lass for children ,
treatment ranged from 0 to 12years (mean = a,ltism or mental retardation, language
5), with the normal-functioning chili:lren lays, multihandicaps, or learning disal
having been out for 3 to 9 years (mean ,=5).

The mean age of S\lbjectsin the control tic~). The three standardized tests wert
group was 10 years (range 6 to 14). The fcillows:
length of time that these children had been
out of treatment ranged from 0 to 9 years 1. Intelligence rest.The Wechsler 111
(mean = 3). Thus, experimental subjects lil~enceScaleforChildren-Revised(Wecru
1~~74w) as administered when subjects~

tended to be older and had been ou.t of able to provide verbal responses. This

treatment longer than had control subjects. cluded all 9 best-outl:ome experimental s

This difference in age occurred becausf~the jects plus 8 of the ren1aining 10experimel

first referrals for the study were all assigned Sllbjectsand 6 of the 19 control subjoots.

to the experimental group due to the fact that Sllbjectswho were nI)t able to provide Vel

referrals cameslowly (7 in the f1fSt3.5years) responses, the Leitt~r International Per:

and therapists were available to treat all of fiance Scale(Leiter, 1959) and the Peabc

them. (As noted earlier, subjects were. as- PictureVocabulary Tt~t-Revised(Dunn,IS

signed to the experimental group if th,era- were administered. All of these tests h.

pists were available to treat them; otherwise, bt~en widely used for the assessment

they entered the control group.) intellectual functionlng in children with

Statistical analyses were conducted to ti~;m(Short & Marcus, 1986).

test whether a bias resulted from the ten- 2. 7be Vtne/and Adapttve Bebal

dency for the f1fSt refeITals to go into the S(;a/es(Sparrow, Balla; & Cicchetti, 19t

experimental group. For example, it is c:on- 111eVineland is a structured interview

ceivable that the first referrals could have ministered to parents assessing the ext

been higher functioning at intake or could to which their child exhibitS behaviors t

have had abetter prognosis than subseq'lent are !:leeded to cope effectively with

referrals. If so, the subject assignmentproce- everyday environment.

dure could have favored the experimental 3. The Pmona/tty Inventory for c,

group. To assessthis possibility, we corre- dJ'"en(Wirt, Lachar,Klinedinst, & Seat,19i

lated the order of referral with intake IQ and 111ismeasure is a 6c1O-itemtrue-false qu

with IQ at the first follow-up (age 7 years). tionnaire filled out by parents that asses

Pearson correlations were computed across the extent to which their children sh,
both groups and within each group. These va.riousforms of PSJ'chological disturbaJ

analyses indicated that the order in which (e.g., anxiety, depr~;sion, hyperactivity, a
Sllbjects were referred was not associated p~;ychoticbehavior).

362 A,utism and Early Interven1

1fiese three testswere intended to pro- each of them" the psychologist recruited a

vide a comprehensive evaluation of intellec- tester from th,esubject's hometown area as

tual, social, and emotional functioning. All of well asan age-matchedcontrol sllbject, and

the testshave been standardized on average data were collected as just described. In

populations. Hence, they provide an objec- addition, the child's examiner filled out a

tive basisfor comparing subjects to children clinical rating scale following a structured

without handicaps across the various areas interview that covered a list c,f standard

that they assess. topic includling friendships, fa.mily rela-

Data were obtained on all subjects ex- tions, and school and community activities.

I cept one girl in the control group, who was The interview was designed both for elicit-
known to be institutionalized and function- ing content al:ld for sampling inlterpersonal
ing very poorly. The 9 best-outcome subjects style. The ratil:lg scale consisted of 22 items,
(those who had been classified as normal each scored O (best clinical status) to 3

functioning at age 7) received particularly (marked devi:ince) points. 11le items were

extensive evaluations, as outlined later. Of designed to il1lcludelikely areasof difficulty

the 28 remaining subjects, 17were evaluated for children ~..ith autism of avel'age intelli-

by staff members in our treatment program, gence (e.g., compulsive or ritualiistic behav-

and 11 received evaluations from outside ior, empathy for and interest iI:l others, a

agencies such as schools or psychology senseof humor) aswell asareasI()fpotential

clinics. (In some cases,the outside agencies difficulty for the general child population

did not administer all of the measuresin this (e.g., depressledmood, anxiety, hyperactiv-
ity). (11le complete scale and a copy of
battery.)
Evaluation of Best-Outcome Subjec:ts. instructions for the clinical intenriew can be

To ensure objectivity in the evaluation of the obtained by ~vriting to the third author),

best-outcome subjects,we arrangedfor blind

administration and scoring of all tests for

thesesubjects asfollows. A psychologist not ReslJllts

associatedwith the study recruited advanced

graduate stlldents in clinical psychology to Experimental Versus Control Gl"OUp
administer the tests.The examinerswere not

familiar with the history of the children, and This firsIt section examines the overall

the psychologist told them simply tha~ the effects of tre~Ltmenthrough comparison of

testing was part of a research study on the follow-up' data from the 19 S1Jbjectws ho

assessment of children. The psychologist recetved the intensive (experimental) treat-

advised them that the nature of the study ment to the d~ltafrom those who I~eceivedthe

necessitatedproviding only certain standard minimal (control) treatment. Data were ob-

background information: age, school place- tained from alIIsubjects on school placement

ment and grade, and parent's name and and from all but one subject in the control

phone number. To increase the heterogene- group on IQ. On the Vineland, scores were

ity of the sample and to control for any obtajned for 18 of 19 experimental subjects

examiner bias, each examiner also tested and IS of IS' control subjects. 11le lowest

one or more S\.lbjectswho were matched in availability of follow-up scores was on the

ageto the experimental subjects and had no PersonalityInventory for d1ildrerl, with scores

history of behavioral disturbance. 'I11eexam- for IS experimental subjects and 12 control

iners were randomly assigned an approxi- subjects,
The subjects in the control group who
mately equal number of subjects for testing

in the experimental group and the compari- hadPersonalityInventory forChlldren scores

son group. Two experimental subjects were did not appf~arto differ from sllbjects who

not living in the local area. Therefore, for were missin~:these scores, as compared on

McEachin, Smith, and lovaas 363

t tests for differences in intake IQ, IQ at 7 rtreostpegcrotiuvpelya),t age 7 (m. ean

years old, or IQ in the present study. of the current evaJuatil:>n.
As noted earlier, 17 of the 29 subjects Table 1

who were not in the best-outcome group
were evaluated by Project staff members, 11
were evaluated by outside agencies, and 1
was not evaluated. To check whether Project
staff members were biased in their evalua-
tions or in their selection of which subjects
to evaluate, we used t tests to compare

subjects they evaluated to those evaluated Measure s'erimenlal Mean
by outside agencies on intake IQ, IQ at age M8i~n SD
7 years, and IQ in the present study. No
la 84.5 32.4
significant differences between subjects
evaluated by Projectstaff members and those Vinetand° 5.1 28.4
evaluated by outside agencies were found. 73.1 26.9
Communication 75.5 26.8
SchoolPlacement. In the experimental Daily Uving S,kills
group, 1 of the 9 Sllbjects from the best- Socialization 71.16 26.8
10.'6 8.2
outcome group who had attended a regular Adaptive Behavior
class at age 7 0. L.) was now in a special Composite 61.1~ 10.2
4.0 3.9
education class.However, 1 of the other 10 Maladaptive Behavior
PICb !)cales
Sllbjects had gone from a special education
classto a regtllar classand was enrolled in a Me;ln elevation
Scales > 70

"Vineland Adaptive Behavior Scale.
for Children.

junior college at the time of this follow-up. Adaptive and Ma'ladapttve
The remaining experimental subjects had On the Vineland, the m.ean overall
not changedtheir classification.Overall, then, posite SCore was 72 in the
the proportion of experimental subjects in group and 48 in the control group.

regular classesdid not change from the age this test is 100,
7 evaluation (9 of 19, or 47%). In the control
tion, Daily Living, and Sodalizatioft
group, none of the 19 children were in a
regular class, as had been true at the age 7
evaluation. The difference in classroomplace-
ment between the experimental group and
the control group was statistically significant,
r (1, N= 38) = 19.05,p< .05.

Intellectual Functioni ng.The testScores
for the experimental group and control group group consistently scored higher
on intellectual functioning, adaptive and

maladaptivebehaviors, and personality func- in the control group, 1(31)= 2.39,p< .05.
tioning aresummarized in Table 1.As can be mean score for the control group was in
seen in the table, the experimental group at clinically significant range whereas
the experimental group was not.
follow-up had a significantly higher mean IQ
than did the control group. This difference of clinically significant le'vels
was significant, 1(35) = 2.97,p < .01. Eleven behavior at ages6 to 9 years;
12to 13 years; and 10 or above, at
subjects (58%) in the experimental group and older.) Thus, the firldings indicate
obtained Full-Scale IQs of at least 8Ojonly 3

subjects (17%) in the control group did as
well. The scores were similar to those ob-

tained by the experimental group and con-

364

iors than did the control group. Children scores. Both the Vineland and Pler-

Pe~ona/ity Functiontng. Scoresfor the sonality Inventory for Children ~"ere com-
experimental group and control group did pleted by parents. In cases where the'se
not differ on overall scale elevation, with scores were not obtained, the p:lrents had
mean tscores of62 and 65,respectively. (On declined to participate.
this test, the mean t score for the general
population is approximately 50 [SD = 101.)T On the measuresthat providt~standard-
ized scores, the functioning of the best-
scores above 60 are considered indicative of
possible or mild deviance, whereas t scores outcome subjects was measured most pJre-
above 70 are viewed as suggesting a clini-
cisely by comparing the best-outcome gro'Jp
against the test norms. Therefore, this anaily-

cally significant problem, namely, one that sis is of primary interest. Data for t]he

may require professional attention. There nonclinical comparison group are mainly
was a significant interaction between the useful in confi~img that the assessme:nt
groups and the individual scaleson this test,
F(15, 390) = 2.36, p < .01. Results of the procedures were valid and in pl:Oviding a
contrast group for the one measu:rewithQlut
Tukey test indicated that the most reliable nonns, the Clinical Rating Scalt~.For the
difference between groups ocCllrred on the nonclinical comparison group, it ~I/illsuffice
Psychosisscale,on which the experimental to summarize the results as follov{s: On the

subjects had a mean of 78 and the control WISC-R this group had mean I(~s of 116
subjects had a mean of 104, F(1, 26) = 8.53, Verbal, 118Performance, and 119Pull-Scale.
p < .01. Sevensubjects in the experimental On the Vineland the group obtained mean

grou p scoredin the clinically preferred range standard scores of 102 Communic:ation, 100

(below 70), whereas no subjects in the con- Daily Living Skills, 102Socialization, and 101
trol group scored that low. Only one other Composite. The mean scale score on the
scale showed a significant difference, So-
matic Concerns, F(1, 26) = 4.60,p< .05. The Personality Inventory for Children was '(9.
1fius, the nonclinical comparison group dis-

control subjects tended to display a below played above-average or average functiQln-

averagelevel of somatic complaints (mean of ing across all areasthat were asst~sed.

45 as compared to 54 for the experimental The next sectlion is focust'd on the

subjects). functioning of the best-outcome group Ion
IQ, adaptive and maladaptive behavior, a'nd

Best-Outcome Versus Nonclinical personality measuresand contrasls the best-
Comparison Group outcome subjects with the comparison SlIlb-

jects on the Clinical Rating Scale.

A t test indicated no significant differ- Intellectual Functiontng. Table 2 pre-

encein agebetween thebest-outcorne group sentsthe IQ data for eachsubject in the best-

and the comparison group of children with- outcome group and the mean scQ,resfor the

out a history of clinically significant behav- group. This table shows that, asa 'whole, the

ioral distllrbance. Subjects in the best-out- 9 best-outcome subjects performc~dwell on

come group had a mean age of 12.42 years the WISC-R. Their IQs placed th,em in the

(range 10.0 to 16.25) versus 12.92 years high end of the normal range, :~bout t~'NO

(range 9.0 to 15.17)for the nonclinical com- thirds of an SD above the mean. Their Full-

parison group. Scores on the WISC-R and ScaleIQs ranged from 99 to 136.

clinical rating scale were obtained for all Subjects'score$were evenly ldistribul:ed

subjects; 1 experimental subject and 2 acrossa range from 80 to 125 on Verbal IQ

nonclinical comparison subjects were miss- and from 88 to 138Ion Performance IQ. The

ing Vineland scores, and 2 experimental subjects averaged31points higher on perfor-

subjects and 1 non clinical comparison sub- manceIQ than Veft!)alJQ.Two of them 0. L.

ject were missing Personality Inventory for and A" G.) had at least a 20-point difference

McEachin, Smith, and Lovaas :~65

Note. Infrm = Information, Simil = Similarities, Arith = Arithmetic, Vocab = Vocabulary , Compr = Compreheneion,
PicC = Picture Completion, PicA = Picture Arrangement, BlkD = Block Design, ObjA = Object Assembly , Cod
= Coding, VIO = Verbal la, Pia = Performance 10, and Full = Full-Scale la.

between Verbal and Perfonnance IQ. sonality Inventory for Children, asmeasured I
On each subtest of the WISC-R, the by the three validity scales (lie, FrequencyI
and Defensiveness).As can be seenfrom the
mean for the general population is 10 (SD ~ table, the subjew scored in the normal range
acrossall scales.They tended to score high-
3). It can be seen from Table 2 that the best- est on Intellectual-Screening, Psychosis,and
outcome subjects scored highest on Similari-
ties, Block Design, and Object Assembly. Frequency. Intellectual-Screening assesses
They scored lowest on Picture Arrangement slow intellectual development, and psycho-
and Arithmetic. Thus, the subjects consis- sis and Frequency assessunusual or strange

tently scored at or above average. behaviors. Only Intellectual-Screening was
Adaptive and Maladaptive BehaVior. above the normal range, and this scale is
affected by subjects' early history. For ex-
Table 3 presents the data for the best-out- ample, the scale contains statementssuch as
come group on the Vineland Adaptive Be- "My child first talked before he (she) was two
havior Scales. It can be seen that the best- yearsold," which would be false for the best-
outcome group scored about averageon the outcome subjew regardless of their current
Composite Scale and on the subscales for level of functioning.

Communication, Daily Living, and Socializa- As Table 4 indicates, 4 best-outcome
tion. However, Table 3 shows that some of
the best-outcome subjects had marginal subjew had a single scale elevated beyond

scores,includingJ. L., B. W., andM. M. Even Table 3
so, all of the best-outcome subjects had Scores on the Vlneland Adaptive Behavior Scale
for the Best-Outcome SubJ-ects
Composite scores within the normal range.
As can be seen in Table 3, on the ~aptive behavior Maladaptive
~ect cam DlS Sac -Comp behavior
Maladaptive Behavior Scale(PartsI and II),
the mean score for the best-outcome group R.S. 83 98 102 92 6
M.C. 119 93 86 98 16
indicated that, on average,thesesubjects did M.M. 119 79 105
not display clinically significant levels of L.B. 107 108 114 108 2
J.L. 77 103 112 88 4
maladaptive behavior. Three of them scored D.E. 80 13
in the clinically significant range versus one A.G. 93 61 94 98 15
B.W. 101 97 82 83 5
subject in the non clinical comparison group, B.R. 83 74 99 9
which had a mean of 7.7 on this scale. Mean 105 94
98 92 6.8
Pe~onaltty Functioning. The results of 99
the Personality Inventory for Children are
Note. Com = Communication, DLS = Daily Living Skills, Soc
summarized in Table 4. The best-outcome = Socialization, Comp = Adaptive Behavior Composite.

subjects obtained valid profiles on the Per-

Table 4 Inventory for Children for the Best-Outcorne Subjects
T Scores on the Personality

T score

the clinically significant range and a 5th 0. L.) treatment. In the present study we have

had nine scaleselevated, including the high- reported dataon these children at a mean age

est scores in the best-outcome group on of 13 years for subjects in the experimental

Intellectual-Screening, psychosis, and Fre- group and 10 years for those in the control

quency. Thus, this subject appeared to ac- group. 11le data were obtained from a com-

count for much of the elevation in scores on prehensive assessmentbattery.

these scales. By comparison, there were 3 11le main findings from the test battery

subjectsin the nonclinical comparison group were as follows: First, subjects in the experi-

with at least one scale elevated. mental group had maintained their level of
~, Clinical Rattng Scale.On this scale,8 of intellectual functioning between their previ-
&1
~~ the best-outcome subjects scored between 0 ous assessment at age 7 and the present

and 10, and the 9th 0. L.) scored 42. The evaluation at a mean age of 13, asmeasured
('i!i mean was 8.8, with a standard deviation of by standardizedintelligence tests.11leirmean

~ 12.9.The nonclinical comparison subjects all IQ was about 30 points higher than that of
1':~Ci scored between 0 and 5 (mean = 1.7, SD = control subjects. Second, experimental sub-

2.1). Because these SDs are unequal, we jects also displayed significantly higher lev-

!~~ used a nonparametric statistic, a Mann- els of functioning than did control subjects
1if;. Whitney Utest, revealing a significantdiffer- on measures of adaptive behavior and per-

~nce between groups, U= 19,p< .05. Thus, sonality. Third, in a particularly rigorous

me best-outcome subjects displayed more evaluation of the 9 subjects in the experi-

deviance than did the comparison subjects, mental group who had been classified as

but most of the deviance appeared to come best-outcome (normal-functioning) in the

from one subject, J. L. earlier study (Lovaas, 1987), the test results

consistently indicated that the subjects ex-

hibited average intelligence and average

Discussion levels of adaptive functioning. Some devi-

ancefrom averagewas found on the person-
" This study is a later and more extensive ality test and tl1e clinical ratings. However ,

follow-up of two groups of young subjects this deviance appeared to derive from the

with autism who were previously studied by extremescoresof one subject,J. L. (seeTable

Lovaas(1987): (a) an experimental group (n 2, 3, and 4). This subject also had been
;: 19) that had received very intensive behav- removed from nonspedal education classes
.oral treatment and (b) a control group (n = and placed in a class for children with

19) that had received minimal behavioral language delays, and he obtained relatively

McEachin, Smith, and Lovaas 367

low scores (about 80) on the Verbal section into a treatment study is beyond the scop I

of the intelligence test and the Communica- the presentreport (seeKazdin, 1980;Ken

tion section of the measure of adaptive & Norton-Ford, 1982;Spitz, 1986).Howe'

behavior. Thus, he no longer appearedto be we note that we incorporated alarge num

nonnal-functioning. However, the reroain- of methodological safeguards in boili

ing 8 subjects who had previously been original study (Lovaa:;,1987) and the pre5

classifiedasnormal-functioning demonstrated investigation:

average IQ, with intellectual perfonnance 1. The experimental ~:roup and

evenly distributed acrosssubtests,were able control group received equivalent assc

to hold their own in regular classes,did not ment batteriesat intal<eand were found tc

show signs of emotional disturbance, and very similar on a multitud{~ of import

demonstratedadequatedevelopmentofadap- variables. Moreover, the number of conI

tive and sodal skills within the normal range. group subjectswho w,erepredicted to achi.

In addition, Sllbjective clinical impressions nclnnal functioning, had they received int

of blind examiners did not discriminate them sive treatment, was approximately equal

from children with no history of behavioral the number of experimental subjects ~

disturbance. These 8 subjects (42% of the actually did achievenormal f\;lnctioning ~

experimental group) may be judged to have int,ensivetreatment (I.ovaas & Smith, 19E

made major and enduring gains and may be Thus, the subject assignment procedl

described as "nonnal-functioning." By con- yielded groups that ~vere comparable pi

trast, none of the control group subjects to treatment. This provided a,strong indi

achieved such a favorable outcome, consis- tion that the superior functioning of I

tent with the poor prognosis for children experimental group after treatment wa:

with autism reported by other investigators result of the treatment itself rather thai

(Preeman,Ritvo,Needleman,&Yokota,1985). biased procedure for assigning subjects

In order to evaluate this outcome, we the experimental group.

must pay close attention to whether or not 2.All subjectsrernainedin the grOUP!

our methodology was sound. The adequacy which they were assigned at intake. Onl:

of our methodology is crudal because the subjects dropped out, and they were I

outcome in the present stlldy represents a replaced. Therefore, the ori~~inalcornpc

major improvement over outcomes obtained tion of the groups W~j essentially preservc

in previous experimental studies on the 3. All subjects w'ere independently

treatment of children with autism (Rutter, agnosed asautistic by PhD or MD clinidal

1985). The only reports of comparable out- and there was high agreement on the di~

comes have come from uncontrolled case nosis between the independent clinidal

studies (e.g., Bettelheim, 1967), and subse- This provided evidence that subjects n

quent investigations have indicated that these criteria for a diagnosis of autism.

case studies grossly overestimated the out- 4. Prior to treatment, these subje

comes obtainable with the treatment that appeared to be comparable to those dia

was provided. Similarly, reports of major nosed as having autism in lother resear

gains in other populations, such as large IQ investigations. Evidence for this comes frc

increases in children from impoverished the second control group that was incorp

backgrounds, alsohavebeen basedon highly rated into the initial treatment study, 11

questionable evidence (Kamin, 1974;Spitz, group was evaluated, by another resear,

1986). Such reports have the potential to team (independent of ours), had similar I(

cause a great deal of harm by misleading at intal<Jebased on the sarolemeasures

consumers and professionals, intelligence that we used, yet showed simil

A detailed description of all the meth- outcome data to those reported by oth

odological safeguards that should be built investigators. Additional evidence can I

368 Autism and Early Interventi

~

derived from the similarity of our intake data have been maintained for an ext'~nded pe-

to datareportedby otherinvestigato~ (Lovaas riod of time.

et al., 1989).For example, although Schopler 9. A wide rangeof measures~rasadmin-

andhisassodates(Schopler,Short,& Mesibov, istered, avoiding overrelianceon intelligence

1989)suggestedthat our samplehad a higher tests,which have limitations if use~din iso:la-

mean IQ than did other samples of children tion (e.g., bias Iresultingfrom teaching to t:he

with autism, their own data do not appear to test, selecting :! test that would yield espe-

differ from ours (Lord & Schopler, 1989). dally favorable'results,failing to ~;sessothler

111us,there is evidence that our subjects aspectsof functioning such assoclialcompe-

were a typical group of preschool-age chil- tence or school perfonnance) (Spitz, 19~16j

dren with autism rather than a select group Zigler & Tricke~tt,1978).

ofhigh-level children with autismwho would 10. The u5e at follow-up of a nomlal

,,;!;0 have been expected to achieve nonnal func- comparisongroup, standardized t(~sting,alrld
"', ,
,.

0,..," tioning with little or no treatment. blind rating allowed for an obj(~ctive, dle-
jll;"C

5. The first control group, which re- tailed, and quantifiable assessmentof tre:a.t-

ceived up to 10 hou~ a week of one-to-one ment effectiveness. A particularl~, rigorous

behavioral treatment, did not differ at post- assessmentwa:sgiven to those subjects wlho

treatment from the second control group, showed the most improvement.

which received no treatment from us. Both Taken together, these safegtJardspro-

groups achieved substantially less favorable vide considerable assurance that the favor-

outcomes than did the experimental group. able outcome of the experiment:u subje,cts

Becauseall groups were similar at pretreat- can be attribul:ed to the treatment they Ire-

ment, this result confirms that our subjects ceived rather than to extraneous fa.ctorssuch

had problems that responded only to inten- as improvement that would hav(: occurred

sive treatment rather than problems such as regardless of treatment, biased procedures

being noncompliant or holding back (mask- for selecting s1ubjectsor assigning them to

ipg an underlying, essentially average intel- groups, or narrow or inappropri:lte assess-

lectual functioning that would respond to ment batteries.

smaller-scaleinterventions). Despite tlle numerous precal.ltions tllat

6. Subjects'families ranged from high to we have taken, several concernls may be

low socioeconomic status, and, on average, raised about tlle validity of the r(:sults. Per-

they did not differ from the general popula- haps the most imp<l>rtanits that the assi~;n-

tion (Lovaas,1987).111usa, lthough our treat- ment to the experitnental or control group

ment required extensive family participa- was made on the b~is of therapist availabil-

tion, a diverse group of families was ity rather than, a more arbitrary procedure

apparently able to meet this requirement. such as alternating:referrals (assigning l:he

7. The treatment has been described in first referral to the experimental group, I:he

detail (Lovaas et al., 1980; Lovaas & Leaf, second to the contrpl group, the Ihird to t:he

1981),and the effectivenessof many compo- experimental group, and so forth), Howev'er,

nents of the treatment has been demon- it seems unlikely that the assignment V{as

strated experimentally by a large number of biased in view of the pretreatment data we

ipvestigato~ over the past 30 years (cf. havepresentedon the similarity b,~tweenIthe

Newsom& RincoverI 1989).Hence, our treat- experimental and control grout:lS. On Ithe

ment may be replicable, a point that is other hand, we do not know asyet whether

discussedin greater detail later. there existsa pretreatment variable that does

8. The results of the present follow-up, predict outcome but was not among the 19

which extended several yea~ beyond dis- we chose, yet could have discrinlinated ibe-

chargefrom treatment for most subjects, are tween groups. In an earlier publication

an encouraging sign that treatment gains (Lovaaset al., 1989),we responded in some

McEachin, Smith, and Lovaas :369

detail to the concern about subject assign- averages are seldom interpreted this way.

ment as well as other possible problems However, as statisticians and methodolo-

associatedwith the original study. There are gistshave pointed out (e.g., Barlo,w& Hersen,

certain additional questions that may be 1984), there are many times when group

raised by this follow-up investigation: averages represent the performance of few

1. The experimental group was older or no subjectswithin the group. This wasone

than the control group at the time of this of those times, asis clearly shown by the data

follow-up evaluation.We explained this find- on individual subjects (Tables 2, 3, and 4).

ing earlier and noted that data analyses Deviance w~; found almost exclusively in

indicated that it was unlikely that this age one subject, not evenly distributed acrossall

difference reflected a bias in subject assign- subjects, and 'we have presented the results

ments. accordingly.

2. The follow-up assessmentsfor 17 of The most important void for researchto

the lower functioning subjects in this study fill at this time is replication by independent

were conducted by staff members from our investigators who employ sound method-

Project, who could have biased the test ologies. Given the objective asse'ssmenitn-

results. However, as noted previously, a stroments that we used and the detailed

check revealed no evidence of such a bias. description that we have provided of l:he

3.The Clinical RatingScale,based on an treatment (Lovaas et al., 1980),S\lch a repli-

interview with subjects who had been clas- cation should be possible. Hmwever, the

sified as normal-functioning in the original treatment is complex and to replicate it

studyI hasno norms or data on reliability and properly, an investigator probably needs to

validity. However, we regard the interview possess (a) a ~;trongfoundation in learoing

simply as an extra check on whether the theory research; (b) a detailed knowledge of

examiners detected residual signs of autism the treatment manual we used; (I:) a super-

or other behavior problems that were some- vised practicum of at least 6 monl:hsin one-

how overlooked in the three other Cwell- to-one work with clients who have develop-

standardized) measures in the study and mental delays, emphasizing discrimination

their 30 subscales. We do not regard the learning and building complex langua/~e;

interview as an instrument that by itself and (d) a commitment to provide 'iOhours of

yields conclusive results. No other interview one-to-one treatment to client per week, SO

that suited our purposes currently exists. In weeks per year, for at least 2 years.Our best-

future investigations, we plan to use an outcome subjects an required a minimum of

interview that Michael Rutter and his associ- 2 years of intensive treatment to achieve

ates are now developing for the purpose of average levels of functioning (another indi-

detecting of residual signs of autism in indi- cation that those subjects had pervasive

viduals with averageintelligence. disabilities and were not merely non-

4.As in most long-term follow-up stud- compliant).
ies, we had some missing data, However, A second void to fill concen1Sthe ma-

there is no evidence that the niissing data jority of children who did not benefit to the

would have changed the overall results. point of achieving normal functioning with

S. In our analysis of the best-outcome intensive treatment Perhaps an earlier start

group, we noted that the group averages in treatment would have been aJIthat was

deviated from "normal" on one subscale of needed to obtain favorable outcomes with

the Personality Inventory for Children and many of these children. More pessimistically,

on the Clinical Rating Scale. We then attrib- perhaps such childl1enrequire ne'w and dif-

uted this deviance to the extreme scores of ferent interventions that have "fet to be

one subject rather than to general problems discovered and implemented. In any case,it

within this grO\lp. We recognize that group is essential to develop more approprialte

370 Autism and Early Intervention

services for these children. 388-451. .
Finally, a rather speculativebut promis- Dunn, L. M. (1981). Peabody Pi(;tUI-eVocabu- ~

ing area for research is to detennine the lary Test-Revis,~dC. ircle River, MN: American Q
extent to which early intervention alters Guidance Service.
neurological structures in young children
with autism. Autism is almost certainly the Freeman, B.J., It.itvo, E. 1{., Needle(nan, R., .a:
result of deficits in such neurological struc- Yokota, A. (1'~8S). The stability olf cognitive
tures (Rutter & Schopler, 1987). However, and linguistic Jparametersin autisml:A 5-year
laboratory studies on animals have shown
that alterations in neurological structure are stUdy. Journa~' 01 the American A,r;ademyof
quite possible as a result of changes in the Child Psychiatry. 24, 290-311.
environment in the first yearsof life (Sirevaag
& Greenough, 1988),and there is reason to Huttenl0<:her, I'. a. (1984). Synapse elimina-
believe that alterations are also possible in tion and plasticity in developing h\Jlmancere-
young children. For example,children under bral cortex. At1oaericanjournal ofMentalDeft-
3 years of age overproduce neurons, den-
drites, axons, and synapses. Huttenlocher ciency, 88, 4~H96,
(1984) hypothesized that, with appropriate
stimulation from the environment, this over- KamJn, LJ. (19'74). Tbesciencea~rpoliticsof
production might allow infants and I,Q. New York: WjJey.
preschoolers to compensatefor neurological
anomalies much more completely than do Kanner, L. (197:0. Follow-up study of II autis-
older children. Caution is needed in gener- tic children originally reported in 1943.jour-
alizing from these findings on average chil- nal 01Autism ,and Childhood Schi:~opbrenia,
dren to early intervention with children with
autism, particularly becausethe exact nature 1, 119-145.
of the neurological anomalies of children Kazdin. A. (19811»)R. esearch design in clini(;al
with autism is unclear at present (e.g., Rutter
& Schopler I 1987).Nevertheless,the findings psych,%gy, New YO£k: Harper & Row.
suggestthat intensive earlyintervention could Kendalll, P. C., ~I:Norton-Ford, J. D. (1982).
compensate for neurological anomalies in
such children. Finding evidence for such Therapy outcome research methods. In P. c.
compensation wQuld help explain why the Kendall & J. N. Butcher (Eds.), H~ndbook of
treatment in this study was effective. More research methj>dsin clini(;ai psychology (pp,
generally, it might contribute to an under- 429-460). New' York: Wiley.
standing ofbrain-behavior relationsin young Leiter, R. G. (19~i9). Part I of the manual for the
children. 1948revision of the Leiter International Perfor-
mana~ Scale: ]~vidence of the reli~Lbility and
References validity of the Leiter tests. PsychologyService

Barlow, D.H.,&Hersen,M.(1984). Single case CenterJournaj~ 11, 1-72.
Lord, c., .a:SChopler, E. (1989). The role of age
experimental design: Strategiesfor studying
behavior change (2nd ed.). New York: Per- at assc~ssment:,levelopmentallevel, and testin
gamon Press. the stability oj: intelligence scores in young
Bettellielm, B. (1967). Theemptyforlress. New autistic chjJdreJ:lJ.ournal of Autism and Devel-
York: The Free Press.
DeMyer, M. K., Hlngtgen,J. N., &Jackson, R. opmental Disorders, 19, 483-499.
K. (1981). Infantile autism reviewed: A de- Lotter, V. (197:S). Follow-up studies. In M.
cade of research. Schizophrenia Bulletin, 7,
Rutter & E. Sdllopler (Eds.), Autis,n: A reap-

pra~ai 01 concepts and treatment. London:
Plenum Press.
Lovaas, 0. I. (1~'87). Behavioral treatment and
norm:u educatiional and intellectual function-
ing in young autistic children. joun2a1 ofCon-
suiting and cnni(;t:il Psychology,5!.;,3-9.

Lovaas, O.I.,Ac]~ennan,A. B.,Alex:lnder, D.,
Flrestone, P., PerkblS, J., .a: Young, D.

(1980). Teaclling developmental~v d~abled
childrr!n: The lne book. Austin, TX: Pro-Ed,

Lovaas, 0.1., K(Joegel,R. L, SlmIIlofas,J. Q., .a:
Long, J. S. (V~73). Some generalization and

follow-up me;asures on autistic children in
behavior therapy. journal of Applied Behavior

Analysis,6, 131-166.
Lovaas. 0. I., & Leaf, a. L. (1981). Five video

J McEachin, Smith, and Lovaas 371

I

r

tapes for teaching developmentally diS'abled Psychoeduc,ltional eva/uation oJts'choo/-aged
children wil'h /ow-incidence di~orders (pp.
children. Baltimore: University Park Press. lSS-l80). Orlando, FL: Grune & Stranon.
Lovaas, 0. I., a Smith, T. (1988). Intensive Simeonnson, R.J., Olley,J. G., 4 Rosenthal,
S. L (1987). Early intervention for children
behavioral treatment with young autistic cl1iI- with autism. In M.j. Guralnick & F. C. Bennett
dren. In B. B. Lahey & A. E. Kazdin reds.). (Eds.), The ~tfectivenessof ear/y interoention
Advances in clinical child psychology(Vol. 11, for at-n:sk a:nd handicapped cl,i/dren (pp.
pp. 285-324). New York: Plenum Press. 275--296).Orlando, FL: Academk Press.
Lovaas, 0. 1., Smith, T., a McEachln, J. J. Slrevaag, A. M. , 4 Greenough, W. T. (1988). A
(1989). Clarifying comments on the young multivariate statistical summary of synaptic
plasticity measures in rats expo!;ed to com-
autism study: Reply to Schopler, Short and plex, social arid individual environrnents. Brain
Mesibov. Journal of Consulting and Clinical
Research,441, 386-392.
Psychology, 57, 165-167. Sparrow. S. S., BaUa, D. A., a:.Clc(hettl. D. V.
McEachin, J. J. (1987). Outcome of autiS'tic
(1984). IntenliewEditionSuroeyFormManual.
children receiving intensiw behavioral treat- Circle Pines, lvlN: American Guidance Service.
ment: psychological status 3 to 12years later. Spin, H. H. (1~.86). The raising of;inte//igence.
Hillsdale, Nj: Erlbaum.
Unpublished doctoral dissertation, University Waterhouse, L., 4 Fein. D. (198~). Develop-
of California, Los Angeles. mental trenw; in cognitive skills J'occhildren
Newsom. C., a Rlncover, A. (1989). Autism. In diagnosed as :lutistic and schizophrenic. Child
E.J. Mash&R. A. Barkley(Eds.), Treatmentof
childhood disorrJers(pp. 286-346). New York: Deve/opment, 55, 236-248.
Guilford Press. WecMler, D. (1974). Manua/fortj!le Wechsler
Butter, M. (1970). Autistic children: Infancy to
adulthood. Seminars in Psychiatry, 2,435-450. Intelligence Scalefor Chi/dren-REwised.New
Butter, M. (1985). The treatment of autistic York: Psychological Corp.
children. Journal of Child Psychology & Psy- Win, R. D., La.char, D., K.linediOJit, J. K.. 4
Seat, P. D. (I~J77). Mu/tidimensional descrip-
chiatry, 26, 193-214. tions of chi/d persona/ity: A ma"lua/ for the
Butter, M., a Schopler, E. (1987). Autism and Personality 171~tory for ChildTe7l.Los Ange-
les: Western Psychological Service~.
pervasive developmental disorders: Concepts Zigler. E., a:.Trlckett, P. K. (1978). IQ, social
and diagnostic issues.Journal of AutiS'mand competence, and evaluation of E~arlychild-
hood intervention programs. Am,~rican Psy-
DevelopmentalDisorrJers, 17, 159-186. ch%gist,33, 789-798.
Schopler, E., Short, A., a Mesibov, G. (1989).
Received:5115191r;lrst decision:10116191a;cce~ed:
Relation of behavioral treatment to "normal 1/23/92.
functioning": Comment on Lovaas.Journal of
Consulting and Clinical Psychology, 57,
162-164.
Short, A., a Man:us, L. (1986). Psychoeduca-
tional evaluation of autistic children and ado-
lescents. In S. S. Strichart & P. Lazaros(Eds.),

372 Autism and Early Intervention