ERIC ED581234: First Step to Success: Applications to Preschoolers at Risk of Developing Autism Spectrum Disorders

2015·0.08 MB·English

by  ERIC

EducationandTraininginAutismandDevelopmentalDisabilities,2015,50(4),397–407
©DivisiononAutismandDevelopmentalDisabilities
First Step to Success: Applications to Preschoolers at Risk of
Developing Autism Spectrum Disorders
Andy J. Frey Jason W. Small, Edward G. Feil, and
UniversityofLouisville John R. Seeley
OregonResearchInstitute
Hill M. Walker Steven Forness
UniversityofOregon UniversityofCalifornia-LosAngeles
Abstract:PreschoolchildrenwithAutismSpectrumDisorder(ASD)maynotalwaysberecognizedassuchduring
their early years, but some of their behavioral problems may nonetheless prompt a referral for behavioral
intervention. Whether such an intervention brings any benefit has not been well studied. We identified a
subsampleof34preschoolchildrenatriskforautismspectrumdisorderfromalargerandomizedcontrolledtrial
(N(cid:2)126)oftheFirstSteptoSuccessprogram.ChildrenatriskofdevelopingASDdemonstratedsignificant
improvementsonsevenof11outcomemeasuresandonaresponderanalysesbasedonsymptomseverity.Process
andfidelitymeasuresalsosuggestedthatFirstStepwasbothfeasibleandsociallyacceptable.Implicationsfor
earlyinterventionforchildrenatriskofdevelopingASDarediscussed.
There are a variety of specific interventions dence-based approaches, and planned and
foryoungchildrenwithAutismSpectrumDis- systematic intervention.
order (ASD) designed to control disruptive There is considerable overlap between the
behaviors, improve school success, and in- core symptoms of ASD and other psychiatric
crease social interactions. They include such disorders, such as Attention Deficit Hyperac-
techniques as discrete trial training, compre- tivity Disorder, Oppositional Defiant Disor-
hensive behavioral treatment, joint-attention der, Conduct Disorder, Depression, and anx-
intervention,andself-management(Rogers& iety disorders (Simonoff et al., 2008; Van der
Vismara, 2008). Such techniques often serve Meeretal.,2012).Thesepsychiatricdiagnoses
as components of comprehensive interven- occur as comorbid conditions in preschool
tions, but to date there is a notable lack of children with ASD at a rate of 30% to 80%,
manualizedcomprehensiveschool-basedinter- which can not only complicate their early
ventionsthathavebeenevaluatedinrandom- identification, but also render them more re-
izedcontrolledtrials(Lopataetal.,2012;Szat- sistant to classroom instruction and less able
mari, Charman, & Constantino, 2012). For to manage daily school routines, particularly
very young children with ASD, Strain, Bar- those involving large groups (Hayashida, An-
ton, and Bovey (2014) suggest that such derson,Paparella,Freeman,&Forness,2010;
treatments include family involvement and
Kim,Freeman,Paparella,&Forness,2012).
participation, early delivery of services, evi-
Because of such significant comorbidity,
early attempts to intervene with young chil-
dren with ASD---particularly in school set-
Research supported by grants from NICHD tings---oftenoccurspriortoaformaldiagnosis
(#R01HD055334 and #R21HD43765) and ACYF (Maenneretal.,2013).Recentdevelopments
(#90YD0098).Wewouldalsoliketothankourpart-
indiagnosticpracticessuggestthatacceptable
nersintheJeffersonCountyPublicSchoolDistrict
diagnosticstabilityforASDcanbeachievedas
(Louisville, KY). Correspondence concerning this
earlyasage2usingstateofthearttechnology
article should be addressed to Andy Frey, Kent
SchoolofSocialWork,UniversityofLouisville,Lou- (Kleinman et al., 2008). Unfortunately, the
isville,KY40292.E-mail:[email protected] averageageofthosediagnosedwithASDis4.5
FirstSteptoSuccess / 397
years (Coonrod & Stone, 2004), suggesting a are at risk of being diagnosed with ASD. We
two-orthree-yeardelaybetweenpossibleand attempt to determine to what extent such a
actualdiagnosisinthiscriticaldevelopmental broad-based intervention might potentially
period. This delay can be even longer for have benefited children at risk of developing
some children with ASD depending on their ASD by retrospectively identifying and exam-
particular constellation of symptoms. Thus, ining a subsample of students at risk for this
utilizingclassroom–basedinterventionsispar- disorder from the large randomized con-
ticularly important for all teachers of young trolledtrialofpreschoolFirstStep(Feiletal.,
children with challenging behavior irrespec- 2015). We report outcomes not only on ge-
tive of whether they have been formally diag- neric measures of social functioning but also
nosedwithASDorotherrelateddisorders. onmeasuresspecifictoASDsymptomatology
AlthoughthefieldofASDhastraditionally and functional impairment. This subsample
examined the effectiveness of interventions approach aligns with calls to conduct transla-
designedspecificallyforchildrenwithanASD tional research that enhances access to evi-
diagnosis, Hoagwood et al. (2012) argue that dence-based treatments for young children
thetrendininterventionservicesisforinten- with ASD (Bregman, 2012; Freeman & Pap-
sivehome-orcommunity-basedinterventions parella, 2009; Odom, Cox, & Brock, 2013;
thatmaynotnecessarilybediagnosisspecific. Rothwell,2005).
Similarly, Forness (2011) has noted that be-
havioralapproachesinsocialinterventiontra-
Method
ditionally focus on broad-based interventions
for classroom management rather than on
Participants
more targeted intervention for specific diag-
noses. One such intervention, First Step to Participants for this multi-site study included
Success, focuses primarily on increasing pro- 34of126child-parent-teachertriadswhopar-
social academic behaviors in young children ticipated in the original efficacy study of pre-
with emotional or behavioral disorders and school First Step (Feil et al., 2015). All triads
hasbeenextensivelyvalidatedinrandomized were from Head Start and preschool pro-
controltrials(seeWalkeretal.,2014).Inthe grams located in three counties in Oregon,
multi-tiered framework for prevention, First one county in Kentucky, and one county in
Step is viewed as being in the second tier of Indiana. For each participating classroom,
interventionstobeusedwhenchildrendonot one child who exhibited elevated external-
respondtousualclassroominstructionorbe- izing behavior problems based on teacher-
havioral management (Dunlap & Fox, 2014). report was recruited and consented to par-
Apreschoolversionofthisprogramhasbeen ticipate in a randomized controlled trial
validated in a recent and large randomized examiningtheefficacyofthepreschoolver-
control trial (Feil et al., 2015). That such a sionoftheFirstSteptoSuccessintervention
broad-basedinterventionmightalsoproveef- program. The 34 children in this current
fective specifically for preschool children at study were subsequently screened and iden-
riskforASDisbasednotonlyonitsfocuson tified as being at risk for ASD, according to
enhancing pro-social behaviors but also its procedures described below. We first de-
emphasis on academic engagement and task scribethestudyproceduresandmanualized
completion. intervention methods for the original con-
Children at risk for developing ASD were trolled trial of 134 children, then describe
chosenforthisstudybecausethisisthefastest selection and outcome measures used for
growing category of special education; and the subsample of 34 preschoolers with ASD
suchchildren,especiallythosewithmildASD in the present study.
cases,maynotbeidentifiedassuch,although
theymayinfactbeidentifiedasneedingearly
ProceduresfortheOriginalTrial
intervention(NationalAutismCenter,2009).
Thus,thepurposeofthisarticleistoexamine Project staff obtained IRB approval for the
use of the preschool version of the First Step original study and then recruited and con-
toSuccessearlyinterventionforchildrenwho sentedparticipatingteachers.Weaskedteach-
398 / EducationandTraininginAutismandDevelopmentalDisabilities-December2015
erstoidentifychildrenintheirclassroomwho coachalsometweeklyfor6to8weekswiththe
exhibited externalizing behavior using an parent or caregivers of children randomized
adaptedversionoftheEarlyScreeningProject,a to the experimental condition on promoting
multi-gated behavioral screening tool (ESP; school success skills via reading, discussion,
Walker,Severson,&Feil,1995).AtESPstage role-plays, and demonstrations. Coaches also
one, teachers nominated and rank-ordered metweeklywithresearchstafftotroubleshoot
five children in the classroom who most casesandwerecloselymonitoredviafrequent
closely matched a detailed description of ex- fidelitychecksduringimplementation.
ternalizing behavior. Then, teachers com- Project staff distributed baseline question-
pleted three ESP screening stage-two rating nairepacketstoteachersandparentspriorto
scales – the Adaptive Behavior Index (ABI), training and randomization and distributed
Maladaptive Behavior Index (MBI), and Ag- post-testpacketsaftercompletionoftheinter-
gressive Behavior Scale (ABS) – for each of vention. Packets included both demographic
thechildrenidentifiedinstageone.Reliability questions and outcome measures described
alphas were .77, .81, and .79, respectively. below. The 126 children participating in the
Project staff scored these three scales, con- efficacy trial were primarily male (65%) and
verted raw scores to severity scores, and then had an average age of 4 years (SD (cid:2) 0.4).
rank-ordered the five children within each Basedonparentreport,childrenwereprimar-
classroomtoidentifythehighest-rankedchild ily either African American (31%) or Cauca-
mostinneedofrecruitmenttothestudy.Only sian (44%). Most participating teachers were
one child was recruited per class because the female (99%) and were either African Amer-
interventioncanonlybeadministeredtoone ican(18%)orCaucasian(72%).Onaverage,
student at a time. Note again that the ESP teachers had taught for 14 years (SD (cid:2) 9.2).
screener is for generic externalizing-type be- Teacher-reported education levels varied.
havior problems and is not necessarily tar- Twenty-two percent reported having a high
getedforchildrenwithaspecificdiagnosis. school diploma, 33% an associate’s degree,
After collecting baseline data from partici- 23%abachelor’sdegree,and22%amaster’s
pating parents and teachers, we randomized degreeorhigher.
children to a usual-care control condition or
an experimental condition. Of the 126 chil-
Intervention
dren included in the final sample, 61 were
randomized to the usual-care control condi- First Step to Success is a collaborative home
tion and 65 were randomized to the experi- and school intervention to help at risk chil-
mental condition. Teachers with a child ran- dreningettingofftoagoodstartinschool.It
domized to the usual-care control condition focusesonchildrenwhohavedifficultyadjust-
received a half-day training in general educa- ing to routine school demands. The teacher,
tion classroom management strategies and the child’s parents, and the First Step coach
the principles of positive behavior support work together as a triad team to teach the
(Sprague & Golly, 2013). Teachers with a target child generic school success skills such
child randomized to the experimental condi- asfollowingdirections,doingone’swork,and
tion received the same half-day training in getting along with peers. The three modular
classroommanagementandpositivebehavior components of First Step are screening (de-
support, plus a half-day training in the pre- scribed above), classroom, and home. The
school version of First Step, and one-on-one school component of the First Step intervention
consultation and support from a behavioral isbasedonagamethatutilizesagreencard,
coachwhoworkedwiththeteachertoimple- which the teacher shows to the target child
ment the program. Participating coaches to provide positive feedback for following
(eight per site) were employees of either Or- teacher expectations. The other side of the
egon Research Institute or the University of card is red, which is used to provide non-
Louisvillewithabachelor’sdegreeorhigher, verbalfeedbackwhenthechilddoesnotcom-
and attended a two-day training session dur- ply. One point can be awarded during every
ing which they received intensive training on 30-second interval if the child’s behavior is
First Step implementation. The behavioral appropriate.Whenthedailyperformancecri-
FirstSteptoSuccess / 399
terion of available points is earned, a brief talconditionandreceivedthePreschoolFirst
rewarding activity (selected by the child in Step intervention and 17 had been random-
collaboration with the teacher) involving the izedtotheusual-carecontrolcondition.
targetchildandpeersisprovidedthuspoten- Table 1 summarizes the analyses compar-
tiallyenhancingthetargetchild’ssocialstatus. ingthesampleidentifiedasbeingatriskfor
A daily note home also communicates results ASD (n (cid:2) 34) to the remaining, non-identi-
ofthegametotheparentswhothenprovide fied sample (n (cid:2) 92) on baseline outcome
positive reinforcement with an individual ac- measures. The two samples did not differ
tivityorrewardwhenthechildreturnshome. significantly on any of the teacher-reported
The First Step coach implements this inter- outcome measures with the exception of a
ventionforthefirst10daysthenturnsitover significantly lower SSiS Empathy score for
totheteacherfor10days,followedbya10-day the subsample. As expected, given that the
period in which the feedback card and rein- ASDrisksamplewasidentifiedusingparent-
forcements are gradually phased out. The reported data, the identified and non-iden-
homecomponentoftheFirstStepprogramcon- tifiedparticipantshadstatisticallysignificant
sistsofsixhome-visitationsessionsconducted differences on all parent-reported out-
bythebehavioralcoachwithaseriesofmanu- comes.
alized lessons for the parents focusing on
helpingthechildintermsofcommunication,
OutcomeMeasures
cooperation, limit setting, problem solving,
friendship-making, and confidence building. Theearlyscreeningprojectbehaviorratingscales
Parents, supervised by the First Step coach, (ESP). ESPstagetwogenericscales(Walker,
teach these school success skills at home and Severson,&Feil,1995)wereusedintheorig-
theteacheristrainedtorecognizeandpraise inal trial and are also reported for this study.
theirdisplayatschool.Thepreschoolversion They include the ABS, MBI, and ABI, de-
of First Step was modified to meet the needs scribed above. All three ESP measures are
of younger children and preschool teachers rated on a 5-point frequency scale ranging
who often have less formal training in class- fromNevertoFrequently.Higherscoresonthe
room management practices. Considerably ABS and MBI indicate higher levels of prob-
moredetailontheinterventionisavailablein lem behavior whereas higher scores on the
Feiletal.(2014). ABI indicate higher levels of social function-
ing.
Social skills improvement system rating scales
SelectionofASDSubsample
(SSiS). Forthisstudy,wealsoexaminedspe-
Weidentifiedthesampleof34childrenatrisk cific outcomes for ASD symptomatology as
fordevelopingASDforthisstudyusingparent well as measures related to ASD functional
report on the Early Childhood Inventory - 4 skills. We used the symptom subscale from
(ECI;Gadow&Sprafkin,2000).TheECIisa bothteachersandparentsaswellastheteacher-
screening tool based on diagnostic criteria reported and parent-reported communication,
specifiedintheDSM-IV(AmericanPsycholog- cooperation, and empathy subscales from the
ical Association, 1994) with excellent validity SSiS social skills rating system (Gresham &
andreliabilityforspecificpsychiatricordevel- Elliott, 2008). The SSiS autism subscale fo-
opmental diagnoses (Gadow & Sprafkin, cusesonASDsymptoms,andtheotherthree
2000). Informants rate items on a four-point focus on related forms of functional impair-
frequency scale ranging from never to very of- ment.Itemsareratedona4-pointfrequency
ten. The ECI normative sample was used to scale(Never,Seldom,Often,AlmostAlways).The
identify a subsample of children at risk for Autism symptom subscale includes 15
developingASD.Acutoffoftwoormorestan- itemsassessingimpairmentinsocialinterac-
dard deviations above the mean identified a tions and communication, as well as repetitive
subsample of 34 of the 126 children (27%) and stereotyped behavior patterns ((cid:2)(cid:2) .83
fromthelargerefficacytrialwhometcriteria for teacher report; (cid:2)(cid:2) .77 for parent re-
for being at risk for developing ASD. Seven- port), with higher scores indicating higher
teenhadbeenrandomizedtotheexperimen- levels of impairment. The communication
400 / EducationandTraininginAutismandDevelopmentalDisabilities-December2015
TABLE 1
BaselineMeansandStandardDeviationsofChildOutcomeMeasuresforSampleIdentifiedasatRiskof
ASDandtheRemainingNon-IdentifiedSample
IdentifiedSample Non-IdentifiedSample
(n(cid:2)34) (n(cid:2)92) TestStatistic
M(SD) M(SD) t p-value
Teacherreport
Symptoms
SSiSAutism 20.5(6.3) 19.3(6.4) (cid:3)0.96 .337
ESPMBI 29.0(6.0) 29.4(6.8) 0.28 .778
ESPABS 19.5(7.8) 20.5(6.0) 0.81 .421
Functioning
SSiSCommunication 10.0(3.2) 10.3(3.3) 0.35 .725
SSiSCooperation 7.4(2.6) 7.0(2.6) (cid:3)0.81 .422
SSiSEmpathy 5.7(3.9) 7.6(4.2) 2.19 .030
ESPABI 23.4(3.9) 22.4(5.1) (cid:3)1.07 .287
Parentreport
Symptom
SSiSAutism 21.7(4.3) 13.6(5.1) (cid:3)8.19 (cid:4).001
Functioning
SSiSCommunication 11.2(2.7) 14.3(2.9) 5.48 (cid:4).001
SSiSCooperation 7.6(2.5) 10.5(3.0) 4.99 (cid:4).001
SSiSEmpathy 8.6(3.6) 12.5(3.6) 5.28 (cid:4).001
subscale includes seven items assessing the an inter-rater reliability of .82. The Classroom
child’s verbal and non-verbal communica- monitoring form (CMF) recorded the target
tion skills ((cid:2)(cid:2) .76 for teacher report; (cid:2)(cid:2) child’scomplianceduringtheclassroomcom-
.69 for parent report). The six-item cooper- ponent such as daily points earned, whether
ationsubscaleexaminesthechild’sabilityto thedailycriterionwasmet,oriftheprogram
follow directions, participate appropriately, day was repeated because the child did not
pay attention, and complete tasks ((cid:2)(cid:2) .79 meet the daily reward criterion. Classroom
for teacher report; (cid:2)(cid:2) .83 for parent re-
dosageistheproportionofprogramdaysout
port). The six-item empathy scale assesses
of30completed.
the child’s ability to comfort, forgive, and
The Home monitoring form (HMF) was used
show kindness toward others ((cid:2)(cid:2) .92 for
tocomputedosage,parentfidelity,andparent
teacher report; (cid:2)(cid:2) .89 for parent report).
complianceforthehomecomponentofFirst
StepandiscompletedbytheFirstStepcoach
ProcessMeasures aftereachsession.TheAlliancesurveywascom-
pletedbythecoachandteacheratpostinter-
We collected First Step coach and teacher
ventiontoassessvariousaspectsoftheirwork-
implementation fidelity data related to the
ingrelationship.Coefficientalphawas.94for
school component and data to identify dos-
thecoachversionand.95fortheteacherver-
age, parent fidelity, and parent compliance.
sion. The Satisfaction survey was collected by
We also collected indicators of teacher and
coach alliance, as well as parent and teacher teachersandparentsatpostintervention,and
satisfaction data for participants randomized assessed the perception of training and sup-
to the intervention condition. The Implemen- portreceived,programusability,andprogram
tationfidelitychecklist(IFC)assessedtheadher- effectiveness, with alphas of .91 and .94, re-
enceandqualityofclassroomimplementation spectively. Further details on all of the above
and was collected once during the coach measures are available in Sumi et al. (2013)
phaseandtwiceduringtheteacherphasewith andWalkeretal.(2009).
FirstSteptoSuccess / 401
TABLE 2
BaselineEquivalenceofChildDemographicCharacteristicsandScreeningMeasuresforChildrenwithASD
Total(n(cid:2)34) Control(n(cid:2)17) Intervention(n(cid:2)17) TestStatistic p-value
Demographiccharacteristic
AgeM(SD) 4.1(0.4) 4.1(0.4) 4.1(0.3) (cid:3)0.45 .658
PercentFemale 8(23.5) 4(23.5) 4(23.5) 0.00 1.000
PercentAfricanAmerican 11(32.4) 3(17.6) 8(47.2) 3.35 .066
PercentCaucasian 19(55.9) 12(70.6) 7(41.2) 2.98 .084
PercentonIFSP 12(35.3) 6(35.3) 6(35.3) 0.00 1.000
Screeningmeasures
Percentranked1stonESP 31(91.2) 15(88.2) 16(94.1) 0.37 .545
ESPABSM(SD) 22.1(6.7) 23.2(7.0) 21.0(6.3) 0.98 .337
ESPABIM(SD) 21.4(4.1) 21.9(3.5) 20.9(4.8) 0.70 .490
ESPMBIM(SD) 30.8(6.3) 30.8(5.4) 30.8(7.2) (cid:3)0.03 .979
Note. Reportedteststatisticsaretforcontinuousand(cid:3)2fordichotomousmeasures.
IFSP(cid:2)IndividualizedFamilyServicePlan,ESP(cid:2)EarlyScreeningProject,ABS(cid:2)AggressiveBehaviorScale,
ABI(cid:2)AdaptiveBehaviorIndex,MBI(cid:2)MaladaptiveBehaviorIndex
StatisticalAnalysis post-intervention, we utilized normative data
for the SSiS Autism subscale. Specifically, we
For each outcome we estimated a linear re-
identified children from the ASD subsample
gression model using the full information
who were one standard deviation or more
maximum likelihood estimator in Mplus 6.0
above the normative mean at baseline and
(Muthe`n & Muthe`n, 2010), which utilizes all
specifiedourcutoffasmovementintothenor-
availabledatatocalculateunbiasedparameter
mative range (i.e, within 1 SD). We utilized
estimates and standard errors. For each
the cutoff and RCI to classify children into
model,weregressedthebaselinevalueofthe
one of four categories. Children were classi-
outcomeandadichotomouspredictorforin-
fiedashaving(a)respondediftheymovedinto
terventioncondition(1(cid:2)FirstStepinterven-
the normative range at post intervention and
tion, 0 (cid:2) wait-list control) on the post-inter- hadaRCIlessthan(cid:3)1.96,(b)improvedifthey
vention outcome. We report Hedges’ g as a
didnotmoveintothenormativerangebuthad
measure of effect size (What Works Clearing- anRCIlessthan(cid:3)1.96,(c)remainedunchanged
house,2011).Effectsizesof.2areconsidered
if they did not meet the RCI criterion (i.e.,
small and effect sizes of .5 and .8 are consid- RCI (cid:4) (cid:3)1.96), or (d) deteriorated if the RCI
ered medium and large effects, respectively. criterion worsened (i.e, RCI (cid:5) 1.96). We dis-
Duetoitssmallsamplesize,thisstudyhaslow
cussmethodsandissuesrelatedtoresponder
power. Although we examined 11 outcome
analysis in a recent paper for those wishing
measures we chose not to apply a correction
moredetail(Smalletal.,2015).
to adjust for multiple comparisons because it
wouldbeanoverlyconservativeapproachfor
analreadyunderpoweredstudy. Results
To identify clinically significant post-inter-
vention responses of study participants, we
PreliminaryAnalyses
conducted a responder analysis utilizing Ja-
cobson and Truax’s (1991) Reliable Change Wecomparedtheinterventionconditionand
Index (RCI). Jacobson and Truax define control condition on child, parent, and
changebasedonatwo-stepcriterionaddress- teacherbaselinedemographicsandonthe11
ing magnitude of change (i.e., the RCI) and outcome measures examined in this study to
changeinfunctioning(i.e.,movementacross evaluatetheequivalencyofchildreninthetwo
a specified cutoff). Since we did not collect conditionsatbaselineAscanbeseeninTable
our autism screening measure, the ECI, at 2,therewerenostatisticallysignificantdiffer-
402 / EducationandTraininginAutismandDevelopmentalDisabilities-December2015
ences between the intervention and control most children at baseline and post-interven-
conditions on these variables. There were, tion, missing item-level data occasionally pre-
however,trend-leveldifferencesontheracial/ cludedsubscalescoring.Atbaseline,twochil-
ethnic composition of the two groups. The dren (6%) were missing data for at least one
intervention condition had a higher percent- parent-reported outcome and another child
age of African American children (47% vs. was missing data for at least one teacher-re-
18%, respectively) and a smaller percentage portedoutcome.Thus,atbaseline,12%ofthe
of Caucasian children (41% vs. 71%, respec- sample was missing data for at least one out-
tively)ascomparedtothecontrolgroup. comemeasure.Atpost-intervention,complete
The intervention and control groups did scale-level data were available for all 33 chil-
not differ significantly on parent demo- dren for whom we collected questionnaire
graphic measures (not shown in Table 2) in- data.Wetestedtheassumptionthatdatawere
cluding percent living in a two-parent house- missingcompletelyatrandom(MCAR)using
hold (24% for both groups), number of Little’s MCAR test. The test was non-signifi-
childreninthehousehold(M[SD](cid:2)2.5[1.0] cant ((cid:3)2 (cid:2) 62.17, p (cid:2) .542) suggesting that
vs.2.6[1.1]),percentwithabachelor’sdegree dataweremissingcompletelyatrandom.
or higher (18% vs. 6%), or age of participat-
ingparent(M[SD](cid:2)31.1[7.6]vs.33.8[9.5]).
Fidelity,ProgramCompliance,Allianceand
Neither did the two groups differ on exam-
Satisfaction
ined teacher and classroom characteristics.
Across both groups, all teachers were female, In terms of process measures, coaches and
primarily Caucasian (88% and 77%, respec- teachers adhered to at least 95% of observed
tively), and had attained a bachelor’s degree coreprogramcomponents.Schoolimplemen-
or higher (53% and 59%, respectively). Al- tation quality (fidelity) was excellent for
though non-significant, teachers in control coaches (.94; range (cid:2) .88 to 1.00) and good
classrooms had more years teaching experi- for teachers (.82; range (cid:2) .41 to 1.00). As
ence (M[SD] (cid:2) 16.0[8.0]) as compared to for dosage, children received 96% (range (cid:2)
teachersininterventionclassrooms(M[SD](cid:2) 73% – 100%) of program days and families
11.1[8.2]). Additionally, there were no statis- received89%(range(cid:2)53%–100%)ofhome
tically significant differences between chil- sessions. On average, child compliance was
drenwithASDintheinterventionandcontrol excellent (.92) but parent compliance (.59)
conditionsonanyofthe11baselineparent- andqualityofimplementation(.62)wereonly
reported and teacher-reported outcome moderately satisfactory. Coaches (M[SD] (cid:2)
measures although there were two trend- 4.57 [.54] on a 5-point scale) and teachers
level differences on the teacher-reported (M[SD](cid:2)4.86[.32])ratedhighlytheirwork-
MBI and ABS. Children in the intervention ing alliance with one another. On average,
condition had non-significant lower MBI teachers and parents reported high levels of
scores(M[SD](cid:2)27.4[5.3]vs.30.9[6.4],p(cid:2) satisfaction with the First Step program
.094) and lower ABS scores (M[SD] (cid:2) (M[SD] (cid:2) 4.44[.42] and 4.34[.55], respec-
17.2[5.0] vs. 21.9[8.4], p (cid:2) .055) as com- tively,ona5-pointscale).
pared to children in the control condition.
PosttestDifferencesonOutcomeMeasures
AttritionandMissingData
Resultsfromthecovariate-adjustedregression
Of the 34 children in the ASD subsample, models are reported in Table 3. Children at
baseline questionnaire data were available riskfordevelopingASDwhoreceivedtheFirst
for 97% of teachers and 100% of parents. At Step intervention improved on all teacher-
post-intervention, we obtained questionnaire reported measures assessing symptoms (au-
data from 97% of teachers and parents. One tism, maladaptive, and aggressive behavior)
childwaslosttofollow-upbutwasretainedin andontwoofthefourmeasuresoffunction-
the sample in accordance with an intent-to- ingcomparedtochildreninthecontrolcon-
treat approach. Although parent-reported dition. Hedges’ g effect sizes ranged from
and teacher-reported data were obtained for (cid:3).65 to (cid:3)1.12 for teacher-reported reduc-
FirstSteptoSuccess / 403
TABLE 3
BaselineandPost-InterventionMeansandStandardDeviationforOutcomeMeasuresbyConditionand
RegressionResults
Control(n(cid:2)17) Intervention(n(cid:2)17) ConditionEffect EffectSize
Baseline Post-Intervention Baseline Post-Intervention
Test
Domain/Measure M(SD) M(SD) M M(SD) M(SD) M Statistic p-value Hedges’g
Adj Adj
Teacherreport
Symptoms
SSiSAutism 19.7(6.3) 16.9(6.3) 17.2 21.4(6.6) 13.8(8.0) 12.5 (cid:3)2.39 .020 (cid:3)0.65
ESPMBI 30.9(6.4) 28.4(7.0) 28.3 27.4(5.3) 20.4(5.6) 21.2 (cid:3)2.75 .006 (cid:3)1.12
ESPABS 21.9(8.4) 21.4(10.6) 22.0 17.2(5.0) 12.0(2.4) 14.5 (cid:3)2.87 .004 (cid:3)0.98
Functioning
SSiSCommunication 10.0(2.9) 12.1(3.3) 12.0 10.2(3.6) 14.1(3.7) 13.7 1.52 .128 0.48
SSiSCooperation 6.9(3.2) 9.1(2.6) 9.3 7.9(2.0) 12.2(2.5) 12.3 3.46 .001 1.18
SSiSEmpathy 6.5(4.0) 9.3(3.8) 8.5 5.4(3.7) 9.9(4.6) 9.8 1.14 .254 0.29
ESPABI 23.4(4.7) 25.8(6.0) 25.4 23.5(3.2) 30.6(4.7) 29.7 2.10 .036 0.80
Parentreport
Symptoms
SSiSAutism 21.4(4.3) 19.2(4.5) 18.9 21.8(4.6) 16.1(5.9) 14.8 (cid:3)2.49 .013 (cid:3)0.77
Functioning
SSiSCommunication 11.2(3.0) 12.6(3.0) 12.8 10.8(2.1) 13.5(3.1) 14.3 1.73 .084 0.52
SSiSCooperation 6.8(2.2) 8.3(2.9) 9.1 8.1(2.4) 10.7(3.3) 10.8 1.83 .068 0.52
SSiSEmpathy 8.1(3.5) 9.3(3.2) 9.9 9.2(3.7) 11.9(4.8) 12.2 2.12 .034 0.57
tions in symptoms and from .29 to 1.18 for children were above the normative range.
teacher-reportedimprovementsinfunctioning. Based on teacher report at post-intervention,
Parents of children in the intervention condi- four intervention children recovered (31%),
tionreportedsignificantlygreaterreductionson two children improved (15%), and seven chil-
the SSiS Autism subscale (Hedges’ g (cid:2) (cid:3).77) dren (54%) remained unchanged on the SSiS
and significant improvements on the SSiS Em- Autism subscale. For the comparison condi-
pathysubscale(Hedges’g(cid:2).57)ascompared tion, two children recovered (15%), two chil-
tothecontrolgroup.Althoughnon-significant, dren improved (15%), eight children (62%)
effect sizes were also in the medium range for remained unchanged, and one did not have
parent-reported improvement on the Commu- post-intervention teacher-reported data on
nication (Hedges’ g (cid:2) .52) and Cooperation the Autism subscale. Thus, based on teacher
(Hedges’g(cid:2).52)subscales. report,six(46%)childrenfromtheinterven-
tion condition recovered or improved as com-
paredtofour(31%)childreninthecontrol
ResponderAnalysis
condition ((cid:3)2 (cid:2) 0.43, p (cid:2) .513, OR[95%
To assess participant response, we examined C.I.](cid:2)1.7[0.4,8.7]).Basedonparentreport
movement into the normative range on the at post intervention, two intervention chil-
SSiSAutismsubscale,aswellasthemagnitude dren recovered (15%), three children im-
of change (i.e., the RCI). We examined re- proved (23%), and eight children remained
sponsesforchildrenwhowereabovetheSSiS unchanged (62%). For children in the con-
normative cutoff at baseline. Of the 34 chil- trol condition, one improved (7%), 12 re-
drenintheASDsample,13of17(76%)chil- mainedunchanged(86%),andonedidnot
dren in the intervention condition and 13 of have parent-reported post-intervention data
17 (76%) children in the control condition on the Autism subscale. Overall, based on
were 1(cid:6) SDs above the mean based on parent report, five intervention children
teacher report. Based on parent report, 13 (38%) recovered or improved as compared to
intervention (76%) and 14 control (82%) one(7%)comparisonchild((cid:3)2(cid:2)3.47,p(cid:2)
404 / EducationandTraininginAutismandDevelopmentalDisabilities-December2015
.063, OR[95% C.I.] (cid:2) 7.5[0.7,76.8]). No extensive learning opportunities to practice
children displayed deterioration on these newskills–bothofwhichhavebeenidentified
measures. asmoderatorsoftreatmenteffectiveness(Rog-
ers&Vismara,2008;Strainetal.,2014).
In review of studies on early detection for
Discussion
ASD, very few actually followed up to deter-
The First Step intervention was implemented mineifscreeningactuallyresultedinreferral
with similar levels of fidelity and program forservices(Daniels,Halladay,Shih,Elder,&
compliancefortheatriskforASDsubsample Dawson,2014).Asurveyofearlyintervention
and for the non-ASD First Step sample. Fur- service coordinators also reported that most
ther, child compliance/dosage, coach- and donotseeautismspecificscreeningactually
teacher-reported alliance, and teacher and completed in preschool settings (Pizur-
parent levels of satisfaction were all relatively Barnekow, Muusz, McKenna, O’Connor, &
similarbetweenthesetwosamples.Preschool Cutler, 2012). Thus,childrenatriskforASD
children at risk for developing ASD showed may not always be identified as such during
significant improvement on a broad range their early school years, as we have noted in
of teacher and parent outcome measures in theintroduction,butsomeoftheirbehavioral
this study, with almost all effect sizes in the problems and symptoms may nonetheless
medium-to-largecategory.Theseeffectsizes promptareferralforgenericbehavioralinter-
are comparable to those identified in the vention. While it may not be ideal, such chil-
non-ASD sample, suggesting the First Step dren could at least participate in helpful in-
intervention is similarly effective for both. terventions such as First Step without having
Additionally,althoughFirstStepwasnotde- to wait until their symptoms are confirmed
signed to target specific disorders such as diagnostically.
ASD,itnonethelessappearstoimproveboth Thisstudyhasseveralstrengths.Mostnota-
the symptoms of ASD and at least some of bly, the randomized design controls for
the impairment associated with this disor- threats to internal validity. Additionally, our
der. Note, however, that long-term out- measurement protocol consisted of relatively
comes of children with ASD suggest that reliableandvalidmeasuresthatarewell-estab-
symptoms may improve with age but that lishedandrespectedinbroad-basedinterven-
impairments tend to linger (Howlin, Moss, tion research for young children at risk for
Savage, & Rutter, 2013). ASD and other challenging behavior. There
Tobeclear,however,wearenotproposing was also substantial evidence that our sub-
theFirstStepinterventionaseitheracompre- sample of children at risk for ASD, though
hensive or specific treatment for young chil- small,wasrelativelywellmatchedbetweenin-
dren with ASD. Children who are known to terventionandcontrolconditionsonavariety
have ASD require a comprehensive interven- of baseline variables. Our outcome variables,
tiondesignedtoaddressallofthecoresymp- furthermore, included not only generic be-
toms of the disorder, such as language and havioral measures but also measures specific
restrictiveorrepetitivebehaviors.However,it toASD.
does address some deficits in social and re- Thisstudyhassomenoteworthylimitations
lated domains using recommended strategies as well. First, our interest in examining the
forASDinstruction.Forexample,itrelieson potential efficacy of the preschool version of
direct instruction of functional communica- FirstStepforchildrenatriskforASDdidnot
tion skills (attending to adults/peers), en- begin with an a priori hypothesis but was a
gageschildreninmeaningfulactivitiesinmul- relatively small subsample selected after the
tiple settings (classroom, and home), and is fact.Second,ourscreeningtargetedchildren
groundedinanecologicalapproachwithspe- with externalizing disorders. Thus, although
cificrolesandresponsibilitiesforpeers,teach- our sample is at risk for developing ASD, it
ers,andparents.TheFirstStepinterventionat maynotberepresentativeofallchildrenwith
the very least appears to provide a vehicle to ASD in that it is likely overrepresented by
begin services prior to formal diagnosis, and externalizers and underrepresented by chil-
offersastructuredenvironment(school)with drenwithinternalizingbehaviorsand/orlim-
FirstSteptoSuccess / 405
ited or no language. Third, as Table 1 sug- Early Intervention, 36, 151–170. Doi: 10.1177/
gests, our subsample at pre-testing was not 1053815114566090
viewed by teachers as significantly more im- Forness, S. R. (2011). Special education and the
paired on most ASD measures compared to newmentalhealth:AresponsetoNewDirection
ofCCBD?BehaviorDisorders,37,41–46.
the remaining sample, although parents did
Freeman, S. F., & Papparella, T. (2009). Autism
viewthemassignificantlydifferent.
interventionresearch:Fromthereviewstoimpli-
Rogers and Vismara (2008) suggest that
cationsforpractice.InternationalReviewofResearch
new readily accessible interventions for chil-
inMentalRetardation,38,195–239.
dren with ASD need to be developed. The Gadow,K.D.,&Sprafkin,J.(2000).EarlyChildhood
evidenceinthisarticlesuggestsapossiblenew Inventory-4 screening manual. Stony Brook, NY:
“tiered” model for interventions such as First CheckmatePlus.
Stepinwhichbroad-basedgroupsofchildren Gresham,F.M.,&Elliott,S.N.(2008).SocialSkills
with a variety of behavioral or developmental Improvement System: Rating Scales. Bloomington,
problemsarenotonlyprovidedwithageneric MN:PearsonAssessments.
Hayashida, K., Anderson, B., Paparella, T., Free-
tier-two intervention, but “non-responders”
man, S. F., & Forness, S. R. (2010). Comorbid
are then provided with additional screening
psychiatricdiagnosesinpreschoolerswithautism
designedtoselectthosewhomightbenefitfrom
spectrum disorder. Behavior Disorders, 35, 243–
various, tier-two interventions augmented with
254.
components of more specific interventions for
Hoagwood, K. E., Jensen, P. S., Acri, M. C., Olin,
certaindisordersthathavealreadybeendemon- S.S.,Lewandowski,R.E.,&Herman,R.J.(2012).
stratedtobeevidence-based(Weiszetal.,2013). Outcome domains in child mental health re-
search since 1996: Have they changed and why
doesitmatter?JournaloftheAmericanAcademyof
Child&AdolescentPsychiatry,51,1241–1257.
References Howlin,P.,Moss,P.,Savage,S.,&Rutter,M.(2013).
Social outcomes in mid- to later adulthood
Bregman,J.D.(2012),Evidence-basedtreatmentin amongindividualsdiagnosedwithautismandav-
autismspectrumdisorders.JournaloftheAmerican erage nonverbal IQ as children. Journal of the
AcademyofChild&AdolescentPsychiatry,51,1113– AmericanAcademyofChild&AdolescentPsychiatry,
1115. 52,572–581.
Coonrod, E. E., & Stone, W. L. (2004). Early con- Jacobson,N.S.,&Truax,P.(1991).Clinicalsignif-
cerns of parents of children with autistic and icance: A statistical approach to defining mean-
nonautisticdisorders.InfantsandYoungChildren, ingfulchangeinpsychotherapyresearch.Journal
17,258–268. ofConsultingandClinicalPsychology,59,12–19.
Daniels, A. M., Halladay, A. K., Shih, A., Elder, Kim,J.J.,Freeman,S.F.,Paparella,T.,&Forness,
L. M., & Dawson, G. (2014). Approaches to en- S. R. (2012). Five-year follow-up of preschoolers
hancing the early detection of autism spectrum with autism and comorbid psychiatric disorders.
disorders: A systematic review of the literature. BehavioralDisorders,38,57–70.
JournaloftheAmericanAcademyofChild&Adoles- Kleinman,J.M.,Ventola,P.E.,Pandey,J.,Verbalis,
centPsychiatry,53,141–152. A. D., Barton, M., Hodgson, S.,...Fein, D.
Dunlap, G., & Fox, M. L. (2014) Supportive inter- (2008).Diagnosticstabilityinveryyoungchildren
ventions for young children with social, emo- withautismspectrumdisorders.JournalofAutism
tional,andbehavioraldelaysanddisorders.InH. andDevelopmentalDisabilities,38,606–615.
M Walker & F. M. Gresham (Eds.), Handbook of Lopata, G., Thomeer, M. L., Volker, M. A., Lee,
evidence-based practices for students having emotional G. K., Smith, T. H., Rodgers, J. D. ...Toomey,
andbehavioraldisorders(pp.503–517).NewYork: J. A. (2012). Open-trial pilot study of a compre-
GuilfordPublishing,Inc. hensive school-based intervention for high-func-
Feil,E.G.,Frey,A.J.,&Golly,A.(2014).FirstStepto tioning autism spectrum disorders. Remedial and
Successforpreschoolchildren.InR.E.Slavin(Ed.). SpecialEducation,34,269–281.
Provenprogramsineducation:Classroommanagement Maenner,M.J.,Schieve,L.A.,Rice,C.E.,Cunniff,
andassessment(pp.105–109).NewYork:Corwin. C., Giarelli, E., Kirby, R. S....Durkin, M. S.
Feil,E.G.,Frey,A.J.,Walker,H.M.,Seeley,J.R., (2013). Frequency and pattern of documented
Golly,A.,&Small,J.W.(2015).Theefficacyofa diagnosticfeaturesandtheageofautismidenti-
home-school intervention for preschoolers with fication.JournaloftheAmericanAcademyofChild&
challenging behaviors: A randomized controlled AdolescentPsychiatry,52,401–413.
trialofPreschoolFirstSteptoSuccess.Journalof Muthe`n,L.K.,&Muthe`n,B.O.(2010).Mplususer’s
406 / EducationandTraininginAutismandDevelopmentalDisabilities-December2015