-
Q1Q2
Q3
Q4
12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758
5960
NEW RESEARCH
616263646566676869707172737475767778798081828384858687888990919293949596979899100101
Development of Three Web-Based ComputerizedVersions of the
Kiddie Schedule for Affective Disordersand Schizophrenia Child
Psychiatric DiagnosticInterview: Preliminary Validity DataLisa
Townsend, PhD, Kenneth Kobak, PhD, Catherine Kearney, MA, Michael
Milham, MD, PhD,Charissa Andreotti, PhD, Jasmine Escalera, PhD,
Lindsay Alexander, MPH, Mary Kay Gill, MSN,Boris Birmaher, MD,
Raeanne Sylvester, MSW, Dawn Rice, MS, Alison Deep, MCA,Joan
Kaufman, PhD
Objective: To present initial validity data on three web-based
computerized versions of the Kiddie Schedule for Affective
Disorders and Schizophrenia(KSADS-COMP).
Method: The sample for evaluating the validity of the
clinician-administered KSADS-COMP included 511 youths 6–18 years of
age who wereparticipants in the Child Mind Institute Healthy Brain
Network. The sample for evaluating the parent and youth
self-administered versions of theKSADS-COMP included 158 youths
11-17 years of age recruited from three academic institutions.
Results: Average administration time for completing the combined
parent and youth clinician-administered KSADS-COMP was less time
thanpreviously reported for completing the paper-and-pencil KSADS
with only one informant (91.9 � 50.1 minutes). Average
administration times for theyouth and parent self-administered
KSADS-COMP were 50.9 � 28.0 minutes and 63.2 � 38.3 minutes,
respectively, and youths and parents ratedtheir experience using
the web-based self-administered KSADS-COMP versions very
positively. Diagnoses generated with all three KSADS-COMPversions
demonstrated good convergent validity against established clinical
rating scales and dimensional diagnostic-specific ratings derived
from theKSADS-COMP. When parent and youth self-administered
KSADS-COMP data were integrated, good to excellent concordance was
also achievedbetween diagnoses derived using the self-administered
and clinician-administered KSADS-COMP versions (area under the
curve ¼ 0.89–1.00).Conclusion: The three versions of the KSADS-COMP
demonstrate promising psychometric properties, while offering
efficiency in administrationand scoring. The clinician-administered
KSADS-COMP shows utility not only for research, but also for
implementation in clinical practice, with self-report preinterview
ratings that streamline administration. The self-administered
KSADS-COMP versions have numerous potential research and
clinicalapplications, including in large-scale epidemiological
studies, in schools, in emergency departments, and in telehealth to
address the critical shortage ofchild and adolescent mental health
specialists.
Key words: child and adolescent psychiatric diagnoses,
computerized assessment, KSADS
J Am Acad Child Adolesc Psychiatry 2019;-(-):-–-.
T
Journal of tVolume - /
102103104105106
his article describes preliminary validity data forthree
updated, web-based computerized versionsof the Kiddie Schedule for
Affective Disorders and
107108109110111112113114115
Schizophrenia for school-age children (KSADS-COMP):1
aclinician-administered version, a self-administered youthversion,
and a self-administered parent version. The paper-and-pencil KSADS
was originally developed in 1978 as anextension of the adult
version of the Schedule of AffectiveDisorders and Schizophrenia
(SADS).2 The paper-and-pencilversion of the KSADS has been
translated into more than 30
he American Academy of Child & Adolescent PsychiatryNumber -
/ - 2019
FLA 5.6.0 DTD � JAAC2574_proof �
different languages and has undergone several revisions,
asreviewed elsewhere.3 The paper-and-pencil version of theKSADS has
been the diagnostic instrument used in multiplestudies sponsored by
the National Institutes of Health andthe pharmaceutical
industry,4-8 including clinical trials thatevaluated treatments for
attention-deficit/hyperactivity dis-order (ADHD),9,10 oppositional
defiant disorder (ODD),11
major depressive disorder (MDD),12,13 anxiety disorders,14
early-onset bipolar disorder,6 schizophrenia,5
posttraumaticstress disorder (PTSD),15 among many others. Many of
the
www.jaacap.org 1
14 June 2019 � 7:26 pm � ce
116
http://www.jaacap.org
-
Q7
TOWNSEND et al.
117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175
176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234
clinical trials that employed the KSADS have resulted inchanges
in pediatric drug labeling by the U.S. Food andDrug Administration
(FDA). It has also been used as avalidation instrument in
large-scale epidemiological studies ofpsychiatric disorders in
youths.16-18
Standardized interviews such as the KSADS are asso-ciated with
increased identification of suicidal ideation anddisorders
underreported in unstructured assessments.19,20
The paper-and-pencil KSADS has demonstrated superiordiagnostic
accuracy compared with traditional unstructureddiagnostic
assessments, particularly for complex, highly co-morbid cases in
outpatient, emergency department, andinpatient settings.21,22
Furthermore, prior work has shownthat computerized versions of
paper-and-pencil scales aresuperior to the paper-and-pencil
versions, with branchingand scoring errors minimized when
computerized assess-ment instruments are used.23-25
The paper-and-pencil version of the KSADS wasdesigned to assess
present and past symptoms according toDSM-IV criteria.1 It is a
semistructured diagnostic interviewwith probes that evaluate
specific symptoms using objectivecriteria regarding symptom
intensity and frequency. Theprobes for each symptom included in the
instrument aredesigned to be used flexibly, giving interviewers
ampleleeway for clarifying questions and probing further asneeded
to score individual items.
The paper-and-pencil KSADS consists of three primarycomponents:
1) an unstructured introductory interview, 2)a diagnostic screening
interview, and 3) supplements tofinalize the criteria required for
each diagnosis. The un-structured introductory interview gathers
demographic in-formation; family composition and history of
psychiatricillness data; a brief description of the presenting
problem;history of prior mental health treatment; and general
in-formation about the child’s interests and adaptive func-tioning
(hobbies, friendships, behavior and performance atschool), with new
questions added to the unstructuredintroductory interview of the
KSADS-COMP aboutbullying, sexual orientation, and gender
identification. Theintroductory interview is a critical component
of theKSADS because it helps to establish rapport, generate
hy-potheses about likely relevant diagnoses, and establish acontext
to elicit symptoms and evaluate the child’s func-tioning. The
diagnostic screening interview surveys two tofour symptoms of each
disorder assessed in the KSADS,with skip out criteria that
determine if the supplements forthose disorders should be
administered. The screen inter-view is designed to provide a good
diagnostic overview andwhen completed in its entirety before moving
to the sup-plements greatly facilitates differential diagnoses.
Diagnosticsupplements are then administered in the
chronological
2 www.jaacap.org
FLA 5.6.0 DTD � JAAC2574_proof �
order in which probable diagnoses emerged except when theonset
of one disorder (eg, a substance use disorder) mayhave influenced
the presentation of the other diagnosis (eg,mood disorder).
All three versions of the KSADS-COMP haveretained the three
primary components of the paper-and-pencil KSADS—the unstructured
introductory interview,the screen interview, and the diagnostic
supplements.However, four major changes were made in developingthe
three KSADS-COMP versions. First, the instrumentwas updated to
reflect DSM-5 diagnostic criteria; second,the instrument was
computerized, including automatedscoring algorithms and data
capture features; third, theKSADS-COMP was designed to generate
both categoricaldiagnoses and diagnosis-specific dimensional rating
scalesof current symptoms; and fourth, the scoring criteria
weremodified so that the response options for all currentsymptoms
are scored using the same standardized 5-pointrating scale. The
frequency of all current symptoms overthe past 2 weeks is now rated
on one common metric inall three versions of the KSADS-COMP (eg,
not at all,rarely, several days, more than half the days, and
nearlyevery day). The threshold for clinical significance
variesdepending on the item. For example, the threshold forfailure
to fulfill a major role obligation associated withsubstance use
(eg, missing school due to substance use) islower than the
threshold for depressive irritability, giventhat missing school
“rarely” or only once during a 2-weekperiod can signal a potential
substance misuse disorder,whereas the threshold for irritability in
the depressivedisorders section is “more than half the days.” The
paper-and-pencil version of the KSADS has unique scoringcriteria
for every item, making training and establishingreliability in
administration problematic.
Questions included in the KSADS-COMP were writtenat a sixth
grade Flesch-Kincaid level. Some of the probesincluded in the
KSADS-COMP were modified from thepaper-and-pencil version of the
KSADS; some were devel-oped by the investigative team; and as in
the development ofpast versions of the KSADS, others were developed
withinput from experts in the field (see Acknowledgments sec-tion
for list of experts who provided input on the devel-opment and/or
refinement of KSADS-COMP probes and/or scoring criteria).
The three versions of KSADS-COMP assess the sameset of diagnoses
contained in the DSM-5–updated versionof the paper-and-pencil
KSADS,26 including mood disor-ders (MDD, persistent depression,
mania, hypomania,cyclothymia, bipolar disorders, and disruptive
mood dys-regulation disorder), psychotic disorders
(schizoaffectivedisorders, schizophrenia, schizophreniform
disorder, brief
Journal of the American Academy of Child & Adolescent
PsychiatryVolume - / Number - / - 2019
14 June 2019 � 7:26 pm � ce
http://www.jaacap.org
-
Q8
DEVELOPMENT OF THE KSADS-COMP
235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293
294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352
psychotic disorder), anxiety disorders (panic
disorder,agoraphobia, separation anxiety disorder, simple
phobia,social anxiety disorder, selective mutism, generalized
anxietydisorder [GAD], obsessive-compulsive disorder),
neuro-developmental disorders (ADHD, autism spectrum
disorder,transient tic disorder, Tourette’s disorder, chronic motor
orvocal tic disorder), behavioral disorders (conduct disorder[CD],
ODD), eating and elimination disorders (enuresis,encopresis,
anorexia nervosa, bulimia, binge eating disorder),trauma- or
stressor-related disorders (PTSD, adjustmentdisorders), and alcohol
use and substance use disorders aswell as numerous other specified
diagnoses when full criteriafor these diagnoses are not met.
With the three versions of KSADS-COMP, a variety ofreports are
available to the clinicians in real time. TheSymptoms/Comments
Report provides a detailed listing ofeach symptom item administered
and responses of theyouth and caregiver to each item. All comments
writtenthroughout by the clinician, parent, or youth can also
beprinted using this report. This is useful for
summarizinginformation about how individuals describe their
symptoms(eg, “I feel like a volcano sometimes”), capturing
informa-tion about specific events (eg, reports of adverse
childhoodexperiences), and details about clinically significant
behav-iors (eg, suspensions). The Diagnosis Report provides
cur-rent and past diagnoses; their associated ICD-10 codes; a
listof all threshold level symptoms; and information aboutwhether
the diagnosis is current, past, or in partial remis-sion. The
Diagnostic Report also provides a comprehensivelist of all
suicidality items and a rating according to theColumbia
Classification Algorithm of Suicide Assessment(C-CASA),27 as
recommended for FDA clinical trials.Additional unique features of
the clinician-administered andself-administered versions of the
KSADS-COMP aredescribed in “Methods.”
This article describes two studies. The first studyexamined the
validity of the clinician-administered versionof the KSADS-COMP,
and the second study examined thevalidity of the parent and youth
self-administered versionsof the KSADS-COMP.
METHOD: STUDY ONE: VALIDATION OFCLINICIAN-ADMINISTERED
KSADS-COMPProceduresThe sample for the initial validation study of
the clinician-administered KSADS-COMP comprised participants
fromthe Child Mind Institute Healthy Brain Network (HBN)initiative,
which includes the clinician-administeredKSADS-COMP and a number of
other relevant clinicalassessments in its standard assessment
battery.28 Subjectsrecruited for the HBN initiative before October
2018 who
Journal of the American Academy of Child & Adolescent
PsychiatryVolume - / Number - / - 2019
FLA 5.6.0 DTD � JAAC2574_proof �
had the clinician-administered KSADS-COMP with bothinformants
and all the relevant ratings scales were includedin this report.
HBN is a large-scale data collection effort(target N ¼ 10,000)
focused on the generation of an openresource for studying child and
adolescent mental health(see HBN website at
http://fcon_1000.projects.nitrc.org/indi/cmi_healthy_brain_network/index.html
for details). Adata sharing agreement was signed allowing for the
sharingof de-identified data for the purposes of examining
thevalidity of the KSADS-COMP, and the use of the de-identified
data included in this article was approved bythe Johns Hopkins
Institutional Review Board.
ParticipantsA total of 511 English-speaking youth and parent
dyadsfrom the HBN initiative were assessed using the
clinician-administered KSADS-COMP interview by two doctoral-level
clinicians. According to the clinician-administeredKSADS-COMP, 75
youths had no psychiatric diagnosis,with an overrepresentation of
children with psychopathol-ogy contained in the community-based HBN
cohort givenincentives for recruitment, include free psychiatric
andlearning assessments and referrals for services when
clinicallyindicated. The average age for the sample was 11.8
years(SD 2.7), and youths ranged in age from 6 to 18 years at
thetime of the interview. There were 307 boys (60%). Of thesample,
62% identified as white (N ¼ 317), and 19.5%identified as Hispanic.
Only 63% of caregivers reportedincome data that were scored
categorically based on incomeearnings below and above $90,000, with
53% of thosereporting income data above this threshold.
MeasuresClinician-Administered KSADS-COMP. The
clinician-administered KSADS-COMP was administered to parentand
youth participants by the same clinician. The parentinterview was
completed first if the youth was a preado-lescent; the order was
reversed if the youth was an adoles-cent. With the
clinician-administered KSADS-COMP, aswith the paper-and-pencil
version of the KSADS, final di-agnoses were based on consensus
ratings integrating infor-mation derived from the parent and youth
interviews. Ingeneral, greater weight is given to the youth’s
reports ofinternalizing symptoms and the caregiver’s report of
exter-nalizing symptoms, although latitude in clinical judgment
isallowed.
In addition to having the three primary components ofthe
paper-and-pencil KSADS discussed in the introduction(the
unstructured introductory interview, the screen interview,and the
diagnostic supplements), the clinician-administeredKSADS-COMP
includes computerized youth and parent
www.jaacap.org 3
14 June 2019 � 7:26 pm � ce
http://fcon_1000.projects.nitrc.org/indi/cmi_healthy_brain_network/index.htmlhttp://fcon_1000.projects.nitrc.org/indi/cmi_healthy_brain_network/index.htmlhttp://www.jaacap.org
-
web3C=FPO
TOWNSEND et al.
353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411
412413414415416417418419420421422423424425426427428429430431432433434
preinterview self-report ratings of the screen interview items
tostreamline the administration of the clinical interview.Figure 1
depicts a screenshot of the clinician-administeredKSADS-COMP
interface. The figure shows the screen thatappears when the
clinician is administering the parent inter-view. The screenshot
shows the youth’s interview response inthe upper left corner, the
parent’s preinterview response in theupper right corner, and the
clinician’s response options in thecenter of the screen. All
symptoms in the KSADS-COMP areinitially surveyed for severity over
the past 2 weeks. If athreshold level response is provided, the
KSADS-COMPinterview progresses to inquire about the next symptom;
ifa subthreshold response is given, the interviewer is promptedto
inquire about the lifetime occurrence of the symptom, withthe
presence of past symptoms rated dichotomously.Threshold criteria
are presented below the response options,allowing the clinician to
determine what responses will beabove and below threshold for that
symptom. In addition, ashighlighted by the red arrow on the screen,
there is also aComments dropdown option associated with each item
thatallows clinicians to write notes throughout the interview.
FIGURE 1 Kiddie Schedule for Affective Disorders and
SchizophClinician-Administered Interview Interface
Note: This screenshot of the KSADS-COMP shows the
clinician-administered parent inpreinterview self-report ratings in
the upper right corner. The availability of these datacomments
section, which can be expanded if the clinician wishes to make
notes in respand the KSADS-COMP.
4 www.jaacap.org
FLA 5.6.0 DTD � JAAC2574_proof �
The instructions for the clinician-administeredKSADS-COMP are
similar to the instructions for thepaper-and-pencil KSADS. The
KSADS-COMP is likewisea semistructured instrument and designed to
be adminis-tered in a conversational style. Whereas fewer
sampleprobes are included in the computer version, clinicians
aretold they do not need to recite the probes verbatim, thatthey
are free to make stylistic changes and incorporatelanguage
generated by the parent or youth when con-ducting the interview,
and that they need only ask as manyquestions as is necessary to
score each item. In addition,information learned in the
unstructured introductoryinterview can be used to further probe
individual items.
Current threshold level and past “ever” responses willtrigger
the supplement for a given disorder to appear at thebottom of the
dashboard. As in the paper-and-pencil versionof the KSADS, the
supplements include the necessaryfollow-up questions to determine
if diagnostic criteria forthe disorder are met, if more than one
episode of the dis-order was experienced, and if the current
disorder is inpartial remission.
renia Computerized Version (KSADS-COMP) Screenshot—
terview screen with teen interview responses in the upper left
corner and parenthelp to streamline the diagnostic interview. The
red arrow calls attention to theonse to this item. See the text for
a more complete description of this screenshot
Journal of the American Academy of Child & Adolescent
PsychiatryVolume - / Number - / - 2019
14 June 2019 � 7:26 pm � ce
435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470
http://www.jaacap.org
-
Q9
web3C=FPO
DEVELOPMENT OF THE KSADS-COMP
471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529
530531532533534535536537538539540541542543544545546547548549550551552553554555556557
Figure 2 shows a screenshot of the diagnostic dashboardof the
clinician-administered KSADS-COMP and high-lights several
additional features of the user interface. Thediagnostic interview
dashboard appears once the unstruc-tured introductory interview of
the KSADS-COMP hasbeen completed. The screen modules are shown in
the toptwo thirds of the figure. Screening modules that have
notbeen administered appear in green; completed modulesappear in
gray. Thus, clinicians can determine at a glancethe screen modules
that have and have not been completed.The bottom third of the
figure depicts the supplementmodules that should be completed
because threshold levelresponses were given in the screen
interview, alleviating theneed for clinicians to track which
supplements should beadministered after the screen interview is
completed. Allmodules of the KSADS-COMP do not need to be
adminis-tered; there is a “choose as you go” option for clinicians.
Forexample, if a clinician has prior diagnostic information for
ayouth and wishes to assess for only a specific disorder or if
thepreinterview ratings completed by the parent and youthsuggest
the likely presence of just one disorder, that onemodule can be
selected independently of the other diagnosticmodules. This feature
greatly enhances the efficiency andversatility of the
clinician-administered KSADS-COMP forapplication in a variety of
clinical settings.
FIGURE 2 Clinician-Administered Kiddie Schedule for Affective
DCOMP) Dashboard—Screen Modules and Activated Supplements
Note: This screenshot shows the dashboard of the
clinician-administered KSADS-COMPcompleted. All the screen
interview modules are depicted on the top two thirds, andcomplete
description of this screenshot.
Journal of the American Academy of Child & Adolescent
PsychiatryVolume - / Number - / - 2019
FLA 5.6.0 DTD � JAAC2574_proof �
Validation Measures. The measures used to validate thediagnoses
generated with the clinician-administeredKSADS-COMP are identical
(eg, Child Behavior Check-list, Screen for Child Anxiety and
Related Emotional Dis-orders) or comparable (eg, Child Depression
Inventory vs.Mood and Feelings Questionnaire) to the measures used
tovalidate the paper-and-pencil KSADS.1
The Mood and Feelings Questionnaire (MFQ) (longform) is a
33-item (child report) or 34-item (parent report)well-validated
scale that measures depressive symptoms inyouths 6 to 17 years of
age, with individual items rated on a0 (“Not true”) to 2 (“True”)
point rating scale.29-33 MFQtotal scores were used in the current
report to validateclinician-administered KSADS-COMP depressive
disorderdiagnoses.
The Screen for Child Anxiety Related EmotionalDisorders (SCARED)
is a 41-item instrument that mea-sures anxiety disorder symptoms in
children and adoles-cents via youth and parent report using a point
rating scale(0–2).34-36 The parent and youth SCARED Total Scoreand
GAD scale scores were used in the current report foranalytic
purposes.
The Child Behavior Checklist (CBCL) is one of themost widely
used instruments for measuring behavioral andemotional
psychopathology in youths.37-39 The Attention
isorders and Schizophrenia Computerized Version (KSADS-
interview. The dashboard appears once the unstructured
introductory interview isa sample of activated supplements are
depicted below. See the text for a more
www.jaacap.org 5
14 June 2019 � 7:26 pm � ce
558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587588
http://www.jaacap.org
-
TOWNSEND et al.
589590591592593594595596597598599600601602603604605606607608609610611612613614615616617618619620621622623624625626627628629630631632633634635636637638639640641642643644645646647
648649650651652653654655656657658659660661662663664665666667668669670671672673674675676677678679680681682683684685686687688689690691692693694695696697698699700701702703704705
Problem, Rule Breaking, and Aggressive Behavior stan-dardized
scale scores were used in the current report tovalidate behavioral
diagnoses generated with the clinician-administered KSADS-COMP.
Data AnalysesUnivariate descriptive statistics were calculated
to characterizestudy participants and to evaluate the frequencies
of DSM-5current and lifetime diagnoses generated with the
KSADS-COMP. Current diagnoses included current and
partiallyremitted episodes of disorders. For analytic purposes,
fourdiagnostic categories were generated for youths who metcriteria
for the following current disorders: any depressivedisorders, any
anxiety disorders, ADHD, and ODD or CD.To assess convergent
validity of the categorical diagnosesgenerated with the
clinician-administered KSADS-COMP,Wilcoxon rank sum tests were
used. Youths who met currentcriteria for a particular category of
disorder were comparedwith youths with no lifetime history of a
disorder in thatcategory on the measures assessing symptoms of that
disorder.Nonparametric statistics were used given that the
outcomemeasures were non–normally distributed. Spearman
rankcorrelation coefficients were also calculated to evaluate
theassociations between the standardized symptom measures(MFQ,
SCARED, and CBCL subscales) and the KSADS-COMP dimensional rating
scales associated with these dis-orders; Table S1 (available
online) lists items included in eachKSADS-COMP diagnostic-specific
dimensional rating scaleexamined in this report.
RESULTS: STUDY ONE: VALIDATION OFCLINICIAN-ADMINISTERED
KSADS-COMPClinician-Administered KSADS-COMP Current andLifetime
DiagnosesRates of current diagnoses are depicted in Table 1.
Lifetimerates of diagnoses were as follows. A total of 66
(13%)youths met criteria for a lifetime depressive disorder;
di-agnoses included MDD (n ¼ 45), persistent depressivedisorder (n
¼ 6), and other specified depressive disorder(n ¼ 17). A total of
213 (42%) youths met criteria for alifetime anxiety disorder,
including panic disorder (n ¼ 4),other specified panic disorder (n
¼ 8), agoraphobia (n ¼18), separation anxiety (n ¼ 52), other
specified separationanxiety disorder (n ¼ 15), social anxiety (n ¼
81), specificphobia (n ¼ 80), GAD (n ¼ 92), other specified GAD (n
¼6), obsessive-compulsive disorder (n ¼ 36), and otherspecified
obsessive-compulsive disorder (n ¼ 1). A total of339 (66%) youths
met criteria for a lifetime ADHD diag-nosis. Of these, 278 youths
met full criteria for currentADHD, 11 met criteria for ADHD in
partial remission, 24
6 www.jaacap.org
FLA 5.6.0 DTD � JAAC2574_proof �
met criteria for other specified ADHD, and 26 met criteriafor
past ADHD. A total of 99 (19%) youths met criteria fora lifetime
ODD, and 21 youths (4%) met lifetime criteriafor CD (16 childhood
onset and 5 adolescent onset). Fre-quencies for the other disorders
assessed with the KSADS-COMP were much lower than the
above-describeddepressive, anxiety, and behavior disorders and are
notpresented here.
Convergent Validity DataTable 1 presents the means and standard
deviations foryouth and parent reports on the MFQ, SCARED, andCBCL
subscales for youths with a current diagnosis of aparticular
category of disorder and youths with no lifetimehistory of a
disorder in that category. Youths with a currentdiagnosis differed
significantly from youths without a his-tory of that given disorder
on all standardized scales (p <.0001, all analyses).
Correlations Between Standardized Clinical RatingScales and
Clinician-Administered KSADS-COMPDimensional Scales and
Associations Between Clinician-Administered KSADS-COMP Dimensional
Scales andDiagnostic Group AssignmentAs noted previously, Table S1
(available online) lists itemsincluded in each of the KSADS-COMP
diagnostic-specificdimensional rating scales examined in this
report. Thedimensional scales for these analyses were derived from
theconsensus ratings that integrated parent and youth reports.The
KSADS-COMP 3-item depression scale derived fromthe consensus
ratings correlated significantly with theyouth (rS ¼ .25, p <
.001) and parent (rS ¼ .40, p < .001)MFQ scores; the KSADS-COMP
1-item GAD consensusratings scale correlated significantly with the
youth (rS ¼.33, p < .001) and parent (rS ¼ .43, p < .001)
totalSCARED scores and the youth (rS ¼ .37, p < .001) andparent
(rS ¼ .45, p < .001) SCARED GAD subscalescores; the KSADS-COMP
4-item consensus ADHD scalecorrelated significantly with the CBCL
Attention Problemsubscale (rS ¼ .59, p < .001); and the
KSADS-COMP 2-item ODD consensus ratings scale correlated
significantlywith the CBCL rule breaking (rS ¼ .56, p < .001)
andaggressive behavior (rS ¼ .61, p < .001) subscales.Table S2
(available online) provides means and standarddeviations for the
clinician-administered KSADS-COMPdimensional scales for each
diagnostic group. For eachdisorder, youths with a positive current
diagnosis scoredsignificantly higher on the corresponding
dimensionalscale than youths who did not meet criteria for that
diag-nosis, suggesting the clinical utility of the screen
itemsincluded in the scales.
Journal of the American Academy of Child & Adolescent
PsychiatryVolume - / Number - / - 2019
14 June 2019 � 7:26 pm � ce
706
http://www.jaacap.org
-
Q10
Q11
TABLE 1 Criterion Validity Data for Clinician-Administered
Kiddie Schedule for Affective Disorders and
SchizophreniaComputerized Version Current Diagnostic Groups on
Standard Clinical Measures (N ¼ 511)
DSM-5 Current Diagnoses Standardized MeasureDiagnosis
Positive,
Mean (SD)
DiagnosisNegative,Mean (SD) Z and p
Depressive disorders (n [ 26) MFQ-C 30.69 (18.25) 12.61 (9.64) Z
[ L5.32; p < .0001MFQ-P 23.27 (12.48) 8.43 (8.18) Z [ L5.89; p
< .0001
Anxiety disorders (n [ 158) SCARED-C total score 31.71 (19.08)
20.63 (15.06) Z [ L6.15; p < .0001SCARED-P total score 23.16
(13.17) 10.30 (8.48) Z [ L10.46; p < .0001
ADHD (n [ 313) CBCL attention problems 66.53 (9.73) 56.07 (7.31)
Z [ L12.12; p < .0001ODD/CD (n [ 78) CBCL rule breaking 63.81
(7.56) 54.61 (6.00) Z [ L9.36; p < .0001
CBCL aggressive behavior 68.41 (8.93) 55.66 (7.14) Z [ L10.55; p
< .0001
Note: Youths with positive current diagnoses scored greater than
youths without positive diagnoses on each of the standard clinical
rating scales,providing convergent validity of the diagnoses
generated with the clinician-administered KSADS-COMP. Wilcoxon rank
sum tests were used toevaluate the differences between KSADS-COMP
positive and negative diagnostic groups on the standardized
measures. Boldface indicates significantresults. ADHD ¼
attention-deficit/hyperactivity disorder (cutoff score ¼ 65); C ¼
Child; CBCL ¼ Child Behavior Checklist; KSADS-COMP ¼ KiddieSchedule
for Affective Disorders and Schizophrenia Computerized Version; MFQ
¼ Mood and Feelings Questionnaire (cutoff score ¼ 27); ODD/CD
¼oppositional defiant disorder/conduct disorder (cutoff score ¼
65); P ¼ Parent; SCARED ¼ Screen for Child Anxiety Related
Emotional Disorders(cutoff score ¼ 25).
DEVELOPMENT OF THE KSADS-COMP
707708709710711712713714715716717718719720721722723724725726727728729730731732733734735736737738739740741742743744745746747748749750751752753754755756757758759760761762763764765
766767768769770771772773774775776777778779780781782783784785786787788789790791792793794795796797798799800801802803804805806807808809810811812813814815816817818819820821822823824
METHOD: STUDY TWO: VALIDATION OFPARENT AND YOUTH
SELF-ADMINISTEREDKSADS-COMPSParticipantsA total of 158 youth and
parent dyads were recruitedfrom three university and clinical sites
to validate the self-administered KSADS-COMPs: Kennedy Krieger
Insti-tute (KKI) and other Johns Hopkins child and
adolescentpsychiatry mental health programs (n ¼ 39),
WesternPsychiatric Institute and Clinic (WPIC) at University
ofPittsburgh Medical Center (n ¼ 71), and the Child MindInstitute
(CMI) (n ¼ 48), with all the youths from CMIwho participated in the
validation of the self-administeredKSADS-COMP also participants in
study one, the vali-dation of the clinician-administered
KSADS-COMP.A subset of 106 youths who completed the
self-administered KSADS-COMP also completed a secondresearch visit
to complete the clinician-administeredKSADS-COMP (see “Procedures”
below regarding se-lection criteria for completing the second
assessment).Inclusion criteria across the sites were 1) 11 to 17
years ofage, 2) parent available and willing to participate in
theresearch, and 3) fluent in English. At the KKI and JohnsHopkins
sites, all participants were required to bereceiving mental health
services for study participation; atWPIC, normal controls were
recruited from the offspringof healthy controls participating in
the Bipolar OffspringStudy (BIOS) (n ¼ 30; grant MH060952; PI:
B.B.), andyouths with psychopathology were recruited from theBIOS
study and WPIC outpatient clinics; and at CMI,subjects were
recruited from a pool of youths consenting
Journal of the American Academy of Child & Adolescent
PsychiatryVolume - / Number - / - 2019
FLA 5.6.0 DTD � JAAC2574_proof �
for the HBN initiative who also agreed to participate inthe
KSADS-COMP study.
Subject CharacteristicsA total of 158 adolescents (n ¼ 82 [52%]
boys, n ¼ 76[48%] girls) and their parents completed the
self-administered KSADS-COMPs. The mean age of youthswas 13.8 years
(SD 1.7). The sample was 54% white (n ¼86); 31% African American (n
¼ 49); 8% Hispanic orLatino (n ¼ 13); 1% each for American Indian
or AlaskanNative (n ¼ 1), Asian (n ¼ 1), and Native Hawaiian
orother Pacific Islander (n ¼ 2); and 3% other (n ¼ 5)
(notepercentages equal more than 100%, as subjects couldchoose more
than one racial identity). Of youths, 92% wereliving with a
biological parent. The subset of 106 youthswho also completed the
clinician-administered KSADS-COMP had an average age of 13.7 years
(1.7); exactly 50%of the subsample (n ¼ 53) was male, and 65% (n ¼
69)were white.
ProceduresGiven the well-documented tendency for informants
toattenuate symptom reports on retest,40 to avoid systematicbias in
results examining the concordance between the self-administered and
clinician-administered KSADS-COMP,at the first study visit half the
subjects were randomlyassigned to complete the adolescent and
parent self-administered KSADS-COMP, and half the subjects
wererandomly assigned to complete the
clinician-administeredKSADS-COMP. Only subjects who met criteria
forMDD, a bipolar diagnosis, ADHD, ODD or CD, PTSD,
www.jaacap.org 7
14 June 2019 � 7:26 pm � ce
http://www.jaacap.org
-
TOWNSEND et al.
825826827828829830831832833834835836837838839840841842843844845846847848849850851852853854855856857858859860861862863864865866867868869870871872873874875876877878879880881882883
884885886887888889890891892893894895896897898899900901902903904905906907908909910911912913914915916917918919920921922923924925926927928929930931932933934935936937938939940941942
a substance use disorder, or no lifetime diagnoses duringthe
first assessment were invited for a second study visit tocomplete
the alternate (eg, self-report vs. clinician-administered) version
of the KSADS-COMP. Among theyouths included in the reassessment
sample, 53% (56 of106) completed the self-administered version of
theKSADS-COMP at the initial assessment. The standardizedclinical
assessment measures were completed at theconclusion of the first
study visit, after completion ofeither the self-administered or
clinician-administeredKSADS-COMP. Parents and youths were each
compen-sated $50 for completion of the first KSADS-COMP
as-sessments, and each received an additional $75 if invitedback to
complete the second interview. Compensation fortravel was also
provided.
When completed second, the clinician-administeredKSADS-COMP
assessments were conducted blinded toinitial self-administered
KSADS-COMP diagnoses and allstandardized clinical assessment
measures completed duringthe first visit. The second KSADS was
completed within 3weeks of the initial assessment for 98% of the
cases (meanduration between KSADS-COMP assessments: 9.78 days;SD
6.33; range, 1–39 days). All study procedures receivedapproval by
the Institutional Review Boards at each of theparticipating
sites.
MeasuresSelf-Administered KSADS-COMP. The
self-administeredversions of the KSADS-COMP are designed for
youthsage 11 and older. Similar to the
clinician-administeredKSADS-COMP, the self-administered versions of
theinstrument contain the same three primary componentsof the
paper-and-pencil KSADS, the introductory inter-view, the screening
interview that evaluates key symptomsfrom each of the disorders
covered in the KSADS, andsupplements that are administered for each
diagnosis withabove threshold scores on the screening items of
thatdiagnosis, in order to thoroughly evaluate the
disordersaccording to DSM-5 criteria. The youth and parent
self-report versions are administered to each informant
sepa-rately. The self-administered KSADS-COMP can becompleted
on-site or remotely, but was completed on-sitefor the current
investigation.
The self-administered KSADS-COMP was designed toemulate the
probing done by a trained clinician. As such, itcontains the same
probes, response options, and scoring andbranching logic as the
clinician-administered KSADS-COMP. For example, if a child endorsed
long-standingdifficulties with inattention and ADHD symptoms
andendorsed difficulties with concentration when completingthe
depression supplement, the child would be presented
8 www.jaacap.org
FLA 5.6.0 DTD � JAAC2574_proof �
with a question that asks whether the concentration
diffi-culties got worse with the onset of the depressed mood.
Asanother example of how the instrument was designed toemulate the
probing done by a trained clinician, if the childendorsed a history
of bullying and paranoid thoughts thatothers are out to get them, a
question would be asked todetermine if the child feels it is just
those who have beenbullying them that are out to get them or if the
paranoidideation is more pervasive.
The parent version of the KSADS-COMP is just text-based.
Questions in the youth self-report version areadministered with
prerecorded video clips to facilitateadministration, and youths can
choose a male (K.K.) orfemale (J.K.) interviewer; Figure S1
(available online)shows a screenshot of the youth report
self-administeredKSADS-COMP. Youths also have the option to turnoff
the video clips and simply read the probes. Parentsand youths have
the ability to add comments at any timeto clarify their answers by
either typing in or, with tablets,writing in with a stylus.
There is a suicide and homicide alert system that con-tacts the
clinician via text or e-mail if a respondent reportssuicidal or
homicidal ideation when completing the self-administered KSADS-COMP
and an option to omit theseitems if the interview is being
administered without aclinician on-site. Reports for the
self-administered KSADS-COMP are likewise available to the
clinician immediately,listing the diagnoses for which the youth met
criteria; thesymptoms endorsed, including homicidality and
suicidalityitems; C-CASA ratings; and all notes written in the
com-ments sections.
Clinician-Administered KSADS-COMP. As described un-der “Method”
for study one, the clinician-administeredKSADS-COMP is a
computerized diagnostic interviewderived from the paper-and-pencil
KSADS. All interviewersfor study two were licensed clinicians with
extensive expe-rience administering the paper-and-pencil KSADS
whoreceived a didactic training session by one of the authors(J.K.)
on the administration of the KSADS-COMP and hadthe opportunity to
experiment with the computer programbefore the initiation of the
investigation. To establishinterrater reliability across sites, the
eight assessors (eg, twoat CMI, four at UPMC, two at KKI) scored
all the screenitems on two mock patient interviews that were
videorecorded, with parent and youth preinterview ratingsavailable
to assessors during the administration. On the firstinterview, all
eight raters scored 94% (154 of 163) of theitems identically, and
the eight raters agreed if the itemswere at or above the clinical
threshold for 97.5% (159 of163) of the items. In rating the second
interview, all eight
Journal of the American Academy of Child & Adolescent
PsychiatryVolume - / Number - / - 2019
14 June 2019 � 7:26 pm � ce
http://www.jaacap.org
-
Q12
Q13
DEVELOPMENT OF THE KSADS-COMP
94394494594694794894995095195295395495595695795895996096196296396496596696796896997097197297397497597697797897998098198298398498598698798898999099199299399499599699799899910001001
10021003100410051006100710081009101010111012101310141015101610171018101910201021102210231024102510261027102810291030103110321033103410351036103710381039104010411042104310441045104610471048104910501051105210531054105510561057105810591060
raters scored 96% (154 of 160) of the items identically, andthe
eight raters agreed if the items were at or above theclinical
threshold for 98% (157 of 160) of the items.Although diagnostic
concordance was not determined, asdiagnoses are computer generated
based on scores of indi-vidual KSADS-COMP items, diagnostic
concordancewould thus be expected to be comparably high.
System Usability Scale. User satisfaction with the
technicalaspects of the self-report versions of the KSADS-COMPwas
assessed using the System Usability Scale (SUS).41,42
The SUS is a well-validated scale for assessing usabilityacross
diverse types of user interfaces (eg, tablet, desktop,interactive
voice response, cell phone), with good internalconsistency
reliability (coefficient a ¼ .91).43 The SUScontains 10 items
covering different aspects of the user’sexperience with the
technology (eg, “I thought the com-puter interview was easy to
use”; “The features of thecomputer interview were too complex”; “I
would take acomputer interview designed like this again”). Each
item israted on a 5-point scale, with anchor descriptions
providedfor the endpoints (1 ¼ strongly disagree, 5 ¼
stronglyagree). A global rating of user-friendliness is also
obtained.
User Satisfaction Questionnaire. The User
SatisfactionQuestionnaire has been used in prior studies to
assesssatisfaction with computer-administered versions of
mentalhealth assessments (see scale items listed in Table 2).44
Users rated their experience on a 4-point scale (stronglyagree,
agree, disagree, strongly disagree) and were asked ifthey would be
willing to be interviewed with the self-reportKSADS-COMP again and
whether they preferred to beasked these types of questions by
computer or clinician or ifthey had no preference.
Patient Health Questionnaire. The 9-item Patient
HealthQuestionnaire (PHQ-9) is a scale designed to measure
keysymptoms of depression. Initially developed to screen
fordepressive disorder among adults in primary care,45 the scalehas
also demonstrated good psychometric properties amongadolescents.46
Response options range from 0 (not at all) to3 (nearly every
day).
Brief Child Mania Rating Scale Parent and ChildReport. The
10-item Brief Child Mania Rating Scale(BCMS) was used in the
current report; the BCMSdemonstrates similar psychometric
properties and per-formance as the long version of the Child Mania
RatingScale.47 Response options range from 0 (never/rare) to 3(very
often). The original 21-item Child Mania RatingScale scale was
designed to measure symptoms of bipolarspectrum illness. It has
good psychometric properties andreliably distinguishes between
symptoms of bipolar
Journal of the American Academy of Child & Adolescent
PsychiatryVolume - / Number - / - 2019
FLA 5.6.0 DTD � JAAC2574_proof �
disorder, characteristics of ADHD, and ratings of
normalcontrols.48
GAD Scale. The 7-item GAD-7 youth report instrument isdesigned
to measure key symptoms of GAD. Responseoptions range from 0 (not
at all) to 3 (nearly every day).49
Strengths and Weaknesses of ADHD Symptoms andNormal Behavior
Scale (SWAN). The Strengths andWeaknesses of ADHD Symptoms and
Normal BehaviorScale (SWAN) was designed to measure
parent-reportedsymptoms of ADHD (18 items) and ODD (12
items).50
Items are positively worded in order to measure youths’strengths
as well as weaknesses. For example, parents areasked, “Compared to
other children, how does your childdo the following: .” and “give
close attention to detail andavoid careless mistakes.” Response
options for the scaleallow for scoring of strengths and weaknesses
and rangefrom �3 (far above average) to 3 (far below
average).Primary Care PTSD Screen. The Primary Care PTSDscreen is a
4-item instrument with dichotomized responseoptions (yes/no) of
core PTSD DSM-5 symptoms that wasdeveloped for adults, but has been
shown to be an effectivescreening tool in adolescents as well.51,52
One point isassigned for each “yes” answer.
Data AnalysesUnivariate statistics were used to examine
demographic char-acteristics, examine responses to the user
satisfaction scales,and describe the frequencies of selected
diagnoses. Bivariatestatistics examined associations between
selected diagnosticgroups attained on the parent and youth
self-report interviewsand the standardized assessment instruments
and KSADS-COMP dimensional rating scales. Given that most of
theoutcome measures were non–normally distributed, nonpara-metric
statistics were used to examine differences betweendiagnostic
groups and associations between scales. Percentagreement, Cohen’s
k,53 and Gwet’s first-order agreementcoefficient (AC1) statistics54
were calculated to examineconcordance between parent and youth
self-report andclinician-generated diagnoses for selected current
psychiatricdisorders, with both Cohen’s k and Gwet’s AC1
statisticscalculated, as Gwet’s AC1 is less affected by prevalence
andmarginal probability than Cohen’s k.53-55
Given the expected and observed high rates of informantvariance,
multinomial logistic regression analyses were con-ducted to derive
weights for integrating parent and youth datafrom the
self-administered KSADS-COMP to predict di-agnoses derived from the
clinician-administered KSADS-COMP, with the items selected for
entry in the regressionmodels from the self-administered KSADS-COMP
data
www.jaacap.org 9
14 June 2019 � 7:26 pm � ce
http://www.jaacap.org
-
TABLE 2 Satisfaction Ratings With Self-Administered
KSADS-COMP
Item
Youth Responses Parents Responses
Mean (SD)Percent Agree orStrongly Agree Mean (SD)
Percent Agree or StronglyAgree
1. I was comfortable answeringquestions on the computer
3.3 (0.7) 91% 3.6 (0.5) 99%
2. The questions were clearlystated and understandable
3.2 (0.9) 85% 3.5 (0.6) 94%
3. The computer did a good jobasking me about my feelings
3.2 (0.7) 90% 3.4 (0.6) 96%
4. I felt less embarrassedanswering these questions onthe
computer than I wouldhave with a clinician
2.8 (1.0) 71% 2.6 (1.0) 54%
5. I found the computer interviewto be a helpful process to
gothrough
3.2 (0.7) 89% 3.4 (0.6) 96%
Note: Overall, both parents and youths felt comfortable
answering the questions via computer, found the questions clearly
stated, and found theinterview a helpful process. Satisfaction
rating scale: 1 ¼ strongly disagree, 2 ¼ disagree, 3 ¼ agree, 4 ¼
strongly agree. KSADS-COMP ¼ KiddieSchedule for Affective Disorders
and Schizophrenia Computerized Version.
TOWNSEND et al.
10611062106310641065106610671068106910701071107210731074107510761077107810791080108110821083108410851086108710881089109010911092109310941095109610971098109911001101110211031104110511061107110811091110111111121113111411151116111711181119
11201121112211231124112511261127112811291130113111321133113411351136113711381139114011411142114311441145114611471148114911501151115211531154115511561157115811591160116111621163116411651166116711681169117011711172117311741175117611771178
generated by the clinical experience of the
investigators.Receiver operating characteristic curve analyses were
thenconducted to determine the accuracy of the multinomiallogistic
models generated using the self-administeredKSADS-COMP data in
predicting clinician-deriveddiagnoses.
RESULTS: STUDY TWO: VALIDATION OFPARENT AND YOUTH
SELF-ADMINISTEREDVERSIONS OF KSADS-COMPAdministration Time
Self-Administered and Clinician-Administered KSADS-COMPThe mean
(SD) interview duration times for the parent andyouth self-report
KSADS-COMP were 63.15 (38.3) mi-nutes and 50.92 (28.0) minutes,
respectively. The self-administered KSADS-COMP was completed by 81%
ofthe parents and 90% of the youths within 90 minutes. Theparent
and youth portions of the clinician-administeredKSADS-COMP had mean
(SD) duration times of 50.3(29.9) minutes and 41.5 (28.5) minutes,
respectively, withthe combined parent and youth
clinician-administeredKSADS-COMP completed in less than 1 hour for
31%of the dyads, less than 90 minutes for 59.4% of the dyads,less
than 2 hours for 75.5% of the dyads, and less than 3hours for 95.3%
of the dyads.
User SatisfactionBoth youths and parents expressed high
satisfaction with thetechnical features of the self-administered
KSADS-COMP;
10 www.jaacap.org
FLA 5.6.0 DTD � JAAC2574_proof �
on the SUS, the mean parent rating was 90.0 (corre-sponding to
“best possible”), and the mean youth rating was81.7 (between “good”
and “excellent”). Similarly, the globalrating of “user
friendliness” of the technology was high: 5.5(SD 1.2) for
adolescents and 5.8 (SD 0.7) for parents (7-point scale: 1 ¼ worst
possible, 2 ¼ awful, 3 ¼ poor,4 ¼ OK, 5 ¼ good, 6 ¼ excellent, 7 ¼
best imaginable).Ratings of parent and youth satisfaction with the
KSADS-COMP are presented in Table 2. Overall, both parentsand
youths felt comfortable answering the questions viacomputer, found
the questions clearly stated, and found theinterview a helpful
process. Among the youths, 85% statedthey were willing to be
interviewed by computer again, andwhen asked if they would prefer
to be asked these types ofquestions by computer or clinician after
completing the self-administered KSADS-COMP, 54% said computer,
11%said clinician, and 35% had no preference. Among theparents, 99%
(n ¼ 132) said they would be willing to beinterviewed again by
computer. In terms of interviewpreference, 28% of the parents
stated they preferred thecomputer, 22% stated they preferred a
clinician, and 50%had no preference.
Convergent Validity DataTable 3 presents the means and standard
deviations foryouth and parent report on the PHQ-9, BCMS,
GAD-7,SWAN (ADHD and ODD), and PTSD measures foryouths who did and
did not meet criteria for the corre-sponding current diagnoses
generated by youth and parent
Journal of the American Academy of Child & Adolescent
PsychiatryVolume - / Number - / - 2019
14 June 2019 � 7:26 pm � ce
http://www.jaacap.org
-
DEVELOPMENT OF THE KSADS-COMP
11791180118111821183118411851186118711881189119011911192119311941195119611971198119912001201120212031204120512061207120812091210121112121213121412151216121712181219122012211222122312241225122612271228122912301231123212331234123512361237
1238123912401241124212431244124512461247124812491250125112521253125412551256125712581259
report on the self-administered KSADS-COMP. Wilcoxonsigned rank
tests indicated that for all these disorders,youths who generated a
positive diagnosis by parent reportscored significantly higher on
the corresponding symptommeasure than youths who were not rated by
their parents ashaving that diagnosis. The same pattern emerged for
youth-generated diagnoses except for youth report of ODD.
Associations Between Self-Report KSADS-COMPDimensional Scales
and Diagnostic Group Assignmentand Correlations Between KSADS-COMP
DimensionalScales With Standardized MeasuresTable S3 (available
online) provides means and standarddeviations for the KSADS-COMP
dimensional scales bydiagnostic group assignment. For each
disorder, youths with apositive current diagnosis scored
significantly higher on thecorresponding dimensional scale than
youths who did notmeet lifetime criteria for that diagnosis. All
comparisons werestatistically significant for parent-rated and
youth-rated di-agnoses. As depicted in Table 4, all KSADS-COMP
youth-
TABLE 3 Scores on Standardized Clinical Measures by Current
DSelf-Administered KSADS-COMP
Never, Mean (SD)Major depressive disorder(PHQ-9)Youth diagnosis
3.59 (5.35) (n [ 96)Parent diagnosis 3.80 (5.45) (n [ 88)
Bipolar 1 or 2 (BCMS)Youth diagnosis 3.81 (4.39) (n [ 113)Parent
diagnosis 2.35 (3.48) (n [ 124)
Anxiety disorder (GAD-7)Youth diagnosis 1.82 (2.91) (n [
82)Parent diagnosis 2.83 (4.89) (n [ 77)
ADHD (SWANeparent report)Youth diagnosis 9.59 (11.96) (n [
87)Parent diagnosis 3.75 (6.05) (n [ 72)
ODD (SWANeparent report)Youth diagnosis 8.24 (8.97) (n [
102)Parent diagnosis 3.77 (5.54) (n [ 74)
PTSD (PTSD-PC)Youth diagnosis .58 (1.03) (n [ 106)Parent
diagnosis .81 (1.25) (n [ 101)
Note: Youths with positive diagnoses scored greater than youths
without poconvergent validity of the diagnoses generated with the
self-administered Kexception to this pattern of findings. Wilcoxon
rank sum tests evaluated diffstandardized measures (PHQ-9, BCMS,
GAD-7, SWAN, and PTSD scale.). Theself-report on the KSADS-COMP.
The “Parent diagnosis” row represents dihyperactivity disorder;
BCMS ¼ Brief Child Mania Rating Scale; GAD-7 ¼ 7-Administered
Kiddie Schedule for Affective Disorders and Schizophrenia
ComPrimary Care Screen for Posttraumatic Stress Disorder, DSM-5
version; PHQWeakness of ADHD Symptoms and Normal Behavior–parent
report.
Journal of the American Academy of Child & Adolescent
PsychiatryVolume - / Number - / - 2019
FLA 5.6.0 DTD � JAAC2574_proof �
generated and parent-generated dimensional rating scales
werealso significantly positively associated with their
correspond-ing same informant standardized measure. Table S4
(availableonline) presents correlations between
clinician-administeredand youth and parent self-administered
KSADS-COMPdiagnostic specific dimensional rating scales, which
likewiseshowed significant correlations.
Diagnostic Concordance Among InformantsTable 5 depicts the
concordance between informants. Asindicated in the “Percent
Negative Agreement” column in thetable, agreement between
informants was greatest when adiagnosis was not present.
Concordance was lower in ratingthe presence of each diagnosis, with
the highest concordancebetween informants found in diagnosing ADHD.
Gwet’sAC1 concordance ratings between diagnoses generated usingthe
parent and youth self-administered KSADS-COMPranged from 0.76 to
0.89; Gwet’s AC1 concordance ratingsbetween diagnoses generated
using the clinician and youthself-administered KSADS-COMP ranged
from 0.80 to 0.91,
iagnostic Groups Determined by Youth and Parent Report on
Current, Mean (SD) Z p
7.65 (6.06) (n [ 20) L3.904 .016.38 (5.03) (n [ 21) L2.770
.01
8.35 (6.38) (n [ 20) L3.56 .016.20 (3.77) (n [ 15) L3.61 .01
12.40 (5.44) (n [ 10) L4.951 .016.13 (4.85) (n [ 15) L3.394
.01
14.69 (9.58) (n [ 29) L2.929 .0121.67 (9.65) (n [ 48) L8.33
.001
11.59 (9.09) (n [ 17) L1.50 .1316.29 (8.26) (n [ 49) L7.61
.001
3.08 (1.26) (n [ 13) L5.44 .0012.40 (1.34) (n [ 5) L2.77 .01
sitive diagnoses on each of the standard clinical rating scales,
providingSADS-COMP. Youths who generated ODD diagnoses was the
onlyerences between KSADS-COMP negative and positive groups on
the“Youth diagnosis” row represents diagnostic groups generated by
youthagnostic groups generated by parents. ADHD ¼
attention-deficit/item Generalized Anxiety Disorder Scale;
KSADS-COMP ¼ Self-puterized Version; ODD ¼ oppositional defiance
disorder; PC-PTSD-5 ¼-9 ¼ 9-item Patient Health Questionnaire; SWAN
¼ Strengths and
www.jaacap.org 11
14 June 2019 � 7:26 pm � ce
1260126112621263126412651266126712681269127012711272127312741275127612771278127912801281128212831284128512861287128812891290129112921293129412951296
http://www.jaacap.org
-
TABLE 4 Spearman Rank Correlations Between KSADS-COMP
Self-Report Dimensional Scales and StandardizedSymptom Measures
KSADS Depression Scale(3 items) PHQ-9 (9 items)Youth self-report
.56**
KSADS Bipolar Scale (3 items) BCMS (10 items)Youth self-report
.55**Parent report .61**
KSADS GAD Scale (1 item) GAD-7 (7 items)Youth self-report
.51**
KSADS ADHD Scale (4 items) SWAN (18 items)Parent report
.76**
KSADS ODD Scale (2 items) SWAN-ODD (12-items)Parent diagnosis
.73**
KSADS PTSD Scale (3 items) PTSD-PC Total (4 items)Youth
self-report .56**
Note: The brief self-administered KSADS-COMP diagnostic
specificrating scales correlated significantly with all standard
clinical ratingscales examined. ADHD ¼
attention-deficit/hyperactivity disorder;BCMS ¼ Brief Child Mania
Rating Scale; GAD ¼ generalized anxietydisorder; GAD-7 ¼ 7-item
Generalized Anxiety Disorder Scale; KSADS ¼Kiddie Schedule for
Affective Disorders and Schizophrenia; KSADS-COMP ¼ Kiddie Schedule
for Affective Disorders and SchizophreniaComputerized Version;
PC-PTSD-5 ¼ Primary Care Screen for Post-traumatic Stress Disorder,
DSM-5 version; PHQ-9 ¼ 9-item PatientHealth Questionnaire; SWAN ¼
Strengths and Weakness of ADHDSymptoms and Normal Behavior–parent
report; SWAN-ODD ¼Strengths and Weaknesses of ADHD symptoms and
Normal Behavior,Oppositional Defiant Symptoms–parent report.**p
< .01.
TOWNSEND et al.
12971298129913001301130213031304130513061307130813091310131113121313131413151316131713181319132013211322132313241325132613271328132913301331133213331334133513361337133813391340134113421343134413451346134713481349135013511352135313541355
13561357135813591360136113621363136413651366136713681369137013711372137313741375137613771378137913801381138213831384138513861387138813891390139113921393139413951396139713981399140014011402140314041405140614071408140914101411141214131414
and Gwet’s AC1 concordance ratings between diagnosesgenerated
using the clinician and parent self-administeredKSADS-COMP ranged
from 0.86 to 0.94. The k valueswere consistently lower for all
comparisons.
Predicting Clinician-Derived Diagnoses Using Youth andParent
Self-Administered KSADS-COMP DataThe parent and youth
self-administered KSADS-COMPitems used to predict
clinician-administered KSADS-COMPdiagnoses are depicted in Table 6,
together with the results ofthe receiver operating characteristic
curve analyses conductedto determine the accuracy of the prediction
models. Overallgood to excellent concordance was achieved between
di-agnoses derived using the self-administered and
clinician-administered KSADS-COMP when parent and
youthself-administered KSADS-COMP data were integrated.56
DISCUSSIONResults from this initial validity study of the
clinician-administered and self-administered versions of the
KSADS-COMP are promising. As evidence of convergent validity,
12 www.jaacap.org
FLA 5.6.0 DTD � JAAC2574_proof �
youths with current KSADS-COMP–generated mood, anxi-ety, ADHD,
and ODD and CD diagnoses scored signifi-cantly higher on the
relevant standardized measure for theparticular diagnosis than
youths without that diagnosis. Thevalidity of the brief dimensional
measures constructed fromKSADS-COMP screen items was also supported
by signifi-cant differences on the scales between
diagnosis-positive anddiagnosis-negative groups. There were also
significant corre-lations between the brief KSADS-COMP dimensional
ratingscales and established standardized clinical rating scales,
whichwere higher when the informants on the measures were thesame
on the KSADS-COMP and standardized rating scales(Table 4).
Longitudinal data collection will be required todetermine if the
brief KSADS-COMP diagnostic specificdimensional rating scales are
useful in tracking treatmentresponse and symptoms over time.
There are currently no plans to compare the paper-and-pencil
version of the KSADS with the clinician-administeredKSADS-COMP.
Existing literature suggests that puttingpaper-and-pencil versions
of clinician-administered diagnosticinterviews on an electronic
platform improves reliability andvalidity substantially by reducing
missing data and eliminatinghuman error in branching and choosing
appropriate interviewquestions.23 In addition, clinician tallying
when using paper-and-pencil versions of structured interviews has
been foundto contribute to significant errors, and research has
shown thatcomputerized versions of structured diagnostic
instrumentsexceed the psychometric performance of their
paper-and-pencil counterparts.24,25
There are three primary limitations to the currentinvestigation:
the restriction of interrater reliability assess-ments to the items
in the screen interview, restriction of thediagnoses present in the
validation samples and the numberof youths who met criteria for
each of the diagnosesexamined, and the somewhat extended period of
time be-tween self-administered and clinician-administered
KSADS-COMP assessments. Despite these limitations, the move toan
electronic format and the other modifications made tothe KSADS
offer many advantages over the paper-and-pencil version of the
instrument.
For example, the clinician-administered KSADS-COMP addresses
several limitations that have been notedpreviously with the
paper-and-pencil version of the KSADS.One such limitation is that
the interview can be excessivelytime-consuming.3,57 The mean
administration time for thecombined parent and youth
clinician-administered KSADS-COMP was 91.9 minutes, which is less
time than haspreviously been reported for completing the
paper-and-pencil version of the KSADS with only one
informant.58
Administration time is reduced and the assessmentstreamlined by
the youth and parent self-administered
Journal of the American Academy of Child & Adolescent
PsychiatryVolume - / Number - / - 2019
14 June 2019 � 7:26 pm � ce
http://www.jaacap.org
-
TABLE 5 Youth, Parent, and Clinician Concordance in Current
Diagnoses (N ¼ 106)Parent and Youth Concordance
Current DiagnosisPercent
Agreement Cohen’s kGwet’sAC1
ParentDiagnosisFrequency
YouthDiagnosisFrequency
PercentPositive
Agreement
PercentNegative
AgreementMDD 82 .31 .76 22 21 43 89Bipolar spectrum 83 .16 .80
11 20 20 94Social anxiety 91 .27 .89 12 7 43 93GAD 86 .22 .84 16 11
36 91OCD 63 .14 .38 28 52 29 85ADHD 71 .32 .52 50 30 67 73ODD 66
.14 .47 50 17 59 67CD 81 .17 .76 18 18 28 89PTSD 89 .06 .87 5 13 8
97No diagnoses 73 .32 .56 41 35 NA NA
Clinician and Youth Concordance
Current DiagnosisPercent
Agreement Cohen’s kGwet’sAC1
ClinicianFrequency
YouthReport
Frequency
PercentPositive
Agreement
PercentNegative
AgreementMDD 87 .23 .84 8 21 25 95Bipolar spectrum 84 .19 .80 7
20 19 96Social anxiety 96 .65 .96 7 7 80 97GAD 84 .18 .80 14 11 33
89OCD 66 .15 .47 10 52 18 95ADHD 72 .40 .41 46 30 79 69ODD 75 .25
.62 31 17 67 76CD 85 .12 .82 8 18 17 94PTSD 92 .52 .91 6 13 42 99No
diagnoses 78 .49 .62 33 32 NA NA
Clinician and Parent Concordance
Current DiagnosisPercent
Agreement Cohen’s kGwet’sAC1
ClinicianReport
Frequency
ParentReport
Frequency
PercentPositive
Agreement
PercentNegative
AgreementMDD 90 .42 .87 8 22 38 97Bipolar spectrum 94 .54 .94 7
11 57 97Social anxiety 94 .54 .94 7 12 57 97GAD 89 .47 .86 14 16 58
93OCD 82 .25 .77 10 28 26 94ADHD 90 .79 .78 46 50 91 89ODD 76 .48
.57 31 50 57 89CD 89 .34 .86 8 18 33 96PTSD 93 .19 .93 6 5 33 95No
diagnoses 88 .72 .78 33 34 NA NA
Note: Consistent with prior research, considerable variability
was noted across informants. Gwet’s AC1 is considered the most
reliable concordancestatistic when the prevalence and marginal
probability of diagnosis are low. AC1 ¼ first-order agreement
coefficient; Q16ADHD ¼ attention-deficit/hyperactivity disorder;
Bipolar spectrum ¼ bipolar 1, bipolar 2, and other specified
bipolar disorder; CD ¼ conduct disorder; GAD ¼ generalizedanxiety
disorder; OCD ¼ obsessive-compulsive disorder; ODD ¼ oppositional
defiant disorder; PTSD ¼ posttraumatic stress disorder.
DEVELOPMENT OF THE KSADS-COMP
14151416141714181419142014211422142314241425142614271428142914301431143214331434143514361437143814391440144114421443144414451446144714481449145014511452145314541455145614571458145914601461146214631464146514661467146814691470147114721473
14741475147614771478147914801481148214831484148514861487148814891490149114921493149414951496149714981499150015011502150315041505150615071508150915101511151215131514151515161517151815191520152115221523152415251526152715281529153015311532
preinterview screen items, the automated branching andscoring,
and the computer tracking of the supplements to becompleted.
Journal of the American Academy of Child & Adolescent
PsychiatryVolume - / Number - / - 2019
FLA 5.6.0 DTD � JAAC2574_proof �
Another limitation of the paper-and-pencil KSADS wasthe need for
extensive clinician training to establish inter-rater reliability
given that each symptom on the paper-and-
www.jaacap.org 13
14 June 2019 � 7:26 pm � ce
http://www.jaacap.org
-
TABLE 6 Predicting Clinician Current Diagnoses Using Youth and
Parent Self-Administered KSADS-COMP Data (N ¼ 106)Depression
Bipolar Disorder ADHD ODDYouth KSADS Depression Scale; youth
reportof suicidality; parent and youth adaptivefunctioning measures
(eg, drop in grades,extracurricular activities,
friendships);antidepressant medication
Youth report decreasedneed for sleep; youthreport elation;
familyhistory of bipolar
disorder; mood stabilizeror atypical antipsychotic;inpatient
hospitalization
Parent KSADS ADHDScale; age of ADHD
onset; ADHDmedication; GAD
diagnosis
Parent KSADS ODD Scale;parent report ofsuspensions
anddetentions; GAD
diagnosis; criterion Atrauma history
AUC [ 0.877 (p < .001) AUC [ 1.00 (p < .001) AUC [ 0.977
(p < .001) AUC [ 0.913 (p < .001)Sensitivity [ 0.94
Sensitivity [ 1.00 Sensitivity [ 0.92 Sensitivity [ 0.92Specificity
[ 0.67 Specificity [ 1.00 Specificity [ 0.91 Specificity [ 0.91
Note: Multinomial logistic regression analyses were conducted to
derive weights for integrating parent and youth data from the
self-administeredKSADS-COMP to predict diagnoses derived from the
clinician-administered KSADS-COMP, with the items selected for
entry in the regression modelsgenerated by the clinical experience
of the investigators. Overall good to excellent concordance was
achieved between diagnoses derived using theself-administered and
clinician-administered KSADS-COMP when parent and youth
self-administered KSADS-COMP data were integrated. AUC ¼area under
the curve; ADHD¼attention-deficit/hyperactivity disorder; GAD ¼
generalized anxiety disorder; KSADS ¼ Kiddie Schedule for
AffectiveDisorders and Schizophrenia; KSADS-COMP ¼ Kiddie Schedule
for Affective Disorders and Schizophrenia Computerized Version; ODD
¼ opposi-tional defiant disorder.
TOWNSEND et al.
15331534153515361537153815391540154115421543154415451546154715481549155015511552155315541555155615571558155915601561156215631564156515661567156815691570157115721573157415751576157715781579158015811582158315841585158615871588158915901591
15921593159415951596159715981599160016011602160316041605160616071608160916101611161216131614161516161617161816191620162116221623162416251626162716281629163016311632163316341635163616371638163916401641164216431644164516461647164816491650
pencil version of the KSADS was assessed using uniquerating
criteria. The uniform rating scale used to assess allcurrent
symptoms in the KSADS-COMP (eg, not at all,rarely, several days,
more than half the days, and nearlyevery day) and the automated
branching features of theKSADS-COMP reduce the need for such
training, whilestill allowing clinicians the flexibility to use
their clinicaljudgment in probing and rating symptoms. The
uniformrating scale for assessing current symptoms likely also
ac-counts for the excellent interrater reliability in scoring
items.
There are multiple additional features that render
theclinician-administered KSADS-COMP more feasible than
itspaper-and-pencil predecessor for routine clinical practice.
Forexample, the “choose as you go” modular format of theKSADS-COMP
allows clinicians to select a subset of modulesof interest rather
than completing the entire interview. Theself-administered
preinterview screen items of the clinician-administered KSADS-COMP
can inform module selection,and the unstructured introductory
interview provides anexcellent initial assessment of adaptive
functioning that lendsgreater confidence to the selection of the
subset of modules tobe administered as well as providing other
relevant informa-tion needed for clinical reports (eg, family,
school, treatmenthistory). The availability of diagnostic reports
in real timefurther addresses efficiency concerns and allows
clinicians toprovide meaningful feedback to children and families
in atimely fashion.
When comparing youth-generated, parent-generated,
andclinician-generated diagnoses derived with the self-administered
and clinician-administered KSADS-COMPs,consistent with research
findings in the field, there was a lack
14 www.jaacap.org
FLA 5.6.0 DTD � JAAC2574_proof �
of strong concordance between informants, with the concor-dance
observed in this investigation comparable to or betterthan that
observed in prior studies.59-62 Prior studies havereported
parent-child k values in diagnosing depressive dis-orders of 0.09
or less61,62 compared with the k value of 0.31 indiagnosing MDD
observed in the current study. Parent-childconcordance in rating
the other major diagnoses was essen-tially comparable to that
observed in prior investigations.61,62
Diagnostic concordance between the parent and clinician
werehigher than between the parent and child across all
diagnoses,with youth and clinician concordance highest for social
anx-iety and PTSD. Overall, agreement between informants onthe
self-administered and clinician-administered KSADS-COMP was highest
when a diagnosis was not present.
Ultimately, in clinical practice in making treatmentdecisions,
cross-informant variance needs to be reconciledand, to date, relies
on clinical judgment to do this. Thereceiver operating
characteristic curve analyses reported inthis article provide proof
of concept that parent and youthdata from the self-administered
KSADS-COMP can beintegrated and used to derive diagnoses with good
toexcellent concordance with clinician-derived diagnoses.However,
further refinement, replication, and validation ofthe models used
to integrate parent and youth self-reportdata to generate diagnoses
similar to clinician-derived di-agnoses are required in larger
scale representative samples.At the present time, in treatment
settings, such as busyemergency departments, the diagnostic
information attainedwith the self-report KSADS-COMP can best be
used toexpedite evaluations and help clinicians finalize
diagnosticimpressions.
Journal of the American Academy of Child & Adolescent
PsychiatryVolume - / Number - / - 2019
14 June 2019 � 7:26 pm � ce
http://www.jaacap.org
-
Q5
Q17
Q6
DEVELOPMENT OF THE KSADS-COMP
16511652165316541655165616571658165916601661166216631664166516661667166816691670167116721673167416751676167716781679168016811682168316841685168616871688168916901691169216931694169516961697169816991700170117021703170417051706170717081709
17101711171217131714171517161717171817191720172117221723172417251726172717281729173017311732173317341735173617371738173917401741174217431744174517461747174817491750175117521753175417551756175717581759176017611762
Whereas there are validated internet-based mental healthscreens
for adolescents,63 unvalidated diagnostic internet-basedassessment
tools available for purchase,64 preliminary workthat has been
conducted on the development and validation ofthe internet and
voice Diagnostic Interview Schedule forChildren for DSM-IV,65,66
and more extensive workcompleted on the Development and Well-Being
Assessment(DAWBA) instruments,67-69 there are many features that
areunique to the three KSADS-COMP instruments that enhancetheir
utility. To the best of our knowledge, the three versions ofthe
KSADS-COMP are the only computer-administered childand adolescent
psychiatric diagnostic interviews that use in-formation attained in
the introductory interview to guideprobing of symptoms (eg,
information about bullying to guidequestions generated when probing
about paranoid ideation)and the only assessment tools to include a
screen interview thatprovides a comprehensive diagnostic overview
to facilitatedifferential diagnoses before surveying the full range
of symp-toms associated with the different diagnoses. The
clinician-administered KSADS-COMP is also the only
computerizeddiagnostic interview that includes a parent and youth
self-report preassessment to streamline interviewing and the
onlytool to give the clinician access to the preinterview
responsesand the responses of the other informant (eg, teen)
whenconducting the interview (Figure 1). The youth
self-reportKSADS-COMP is also the only psychiatric diagnostic
instru-ment with video clips to facilitate administration. To
date,Spanish, Dutch, and Danish translations of the KSADS-COMP
instruments have been produced, with automatedmethods developed to
create future translations.
Beyond determining categorical psychiatric diagnoses,there is
growing interest in the field since the initiation of theNational
Institute ofMental Health Research Domain Criteria(RDoC) program in
using dimensional assessments that mapmore clearly onto distinct
neural circuits rather than hetero-geneous categorical diagnoses.70
With the diagnostic-specificdimensional scales included within the
KSADS-COMP andplans to create KSADS-COMP transdiagnostic rating
scalesand refine the algorithms to integrate parent and youth
self-report data to derive categorical diagnoses that more
closelyapproximate clinician diagnoses, the KSADS-COMP mayhelp
serve as a bridge between DSM and RDoC diagnosticperspectives.
The RDoC, however, is a research framework withthe goal of
generating the necessary database to helpderive a new psychiatric
nomenclature informed by
Journal of the American Academy of Child & Adolescent
PsychiatryVolume - / Number - / - 2019
FLA 5.6.0 DTD � JAAC2574_proof �
neuroscience, genetics, and psychology.70 In the
interim,clinicians are required to generate DSM or ICD-10
di-agnoses for assessment, treatment, and billing purposes.The
clinician-administered KSADS-COMP shows utilitynot only for
research but also for implementation inclinical settings, with the
self-report preinterview ratings,choose-as-you-go module options,
and automated scoringto streamline assessments and shorten
administrationtime. The self-administered versions of
KSADS-COMPhave numerous potential research and clinical
applica-tions, including use in large-scale epidemiological
studies,in schools and busy emergency departments, and in
tel-ehealth to address the critical shortage of child andadolescent
mental health specialists in many areas of theUnited States.
1
Accepted May 13, 2019.
Drs. Townsend, Kearney, and Kaufman are with the Kennedy Krieger
Institute,Baltimore, MD. Drs. Townsend and Kearney are also with
Johns HopkinsSchool of Medicine, Baltimore, MD. Dr. Kobak and Ms.
Deep are with Centerfor Telepsychology, Madison, WI. Drs. Milham,
Andreotti, Escalera and Ms.Alexander are with Child Mind Institute,
New York. Dr. Milham is also withNathan Kline Institute,
Orangeburg, NY. Dr. Birmaher and Mss. Gill, Sylvester,and Rice are
with Western Psychiatric Institute and Clinic, University of
Pitts-burgh School of Medicine, PA.
KSADS-COMP Access Information: To test a demo or obtain access
to theKSADS-COMP diagnostic instruments go to ���.This work was
funded by the National Institutes of Health/National Institute
onDrug Abuse (NIH/NIDA; grant R44 MH094092 to Drs. Kaufman and
Kobak) andthe National Institute of Mental Health (NIMH; grant
MH060952 toDr. Birmaher), with additional support provided by the
Zanvyl and IsabelleKrieger Fund (Dr. Kaufman).
Dhananjay Vaidya, PhD, of Johns Hopkins School of Medicine,
served as thestatistical expert for this research.
The authors extend appreciation to the consultants who
contributed to thisinstrument, including Deanna M. Barch, PhD, of
Washington University; MarcoGrados, MD, of Johns Hopkins School of
Medicine; Daniel Hoover, PhD, ofKennedy Krieger Institute; Ellen
Leibenluft, MD, of National Institute of MentalHealth; Danny Pine,
MD, of National Institute of Mental Health; Kenneth Sher,PhD, of
University of Missouri; and Susan F. Tapert, PhD, of University
ofCalifornia San Diego.
Disclosure: Dr. Kobak may in the future receive royalties from
KSADSCOMP,LLC. Dr. Milham has received grant funding from NIH. Dr.
Birmaher hasreceived grant funding from NIH. Dr. Kaufman has
received grant funding fromNIH, has served as a consultant for
Pfizer and Otsuka Pharmaceuticals, and mayin the future receive
royalties from KSADSCOMP, LLC. Drs. Townsend,Andreotti, and
Escalera and Mss. Kearney, Alexander, Gill, Sylvester, Rice,
andDeep report no biomedical financial interests or potential
conflicts of interest.
Correspondence to Joan Kaufman, PhD, Center for Child and Family
TraumaticStress, Kennedy Krieger Institute, John Hopkins School of
Medicine, 1741Ashland Avenue, Room 434, Baltimore, MD 21205;
e-mail: [email protected]
0890-8567/$36.00/ª2019 Published by Elsevier Inc. on behalf of
the AmericanAcademy of Child and Adolescent Psychiatry.
https://doi.org/10.1016/j.jaac.2019.05.009
17631764
REFERENCES
17651766
1. Kaufman J, Birmaher B, Brent D, et al. Schedule for Affective
Disorders andSchizophrenia for School-Age Children–Present and
Lifetime Version (K-SADS-
PL): initial reliability and validity data. J Am Acad Child
Adolesc Psychiatry. 1997;36:980-988.
www.jaacap.org 15
4 June 2019 � 7:26 pm � ce
17671768
mailto:[email protected]:[email protected]://doi.org/10.1016/j.jaac.2019.05.009http://www.jaacap.org
-
Q14
TOWNSEND et al.
17691770177117721773177417751776177717781779178017811782178317841785178617871788178917901791179217931794179517961797179817991800180118021803180418051806180718081809181018111812181318141815181618171818181918201821182218231824182518261827
1828182918301831183218331834183518361837183818391840184118421843184418451846184718481849185018511852185318541855185618571858185918601861186218631864186518661867186818691870187118721873187418751876187718781879188018811882188318841885
2. Endicott J, Spitzer RL. A diagnostic interview: the schedule
for affective disorders andschizophrenia. Arch Gen Psychiatry.
1978;35:837-844.
3. Ambrosini PJ. Historical development and present status of
the schedule for affectivedisorders and schizophrenia for
school-age children (K-SADS). J Am Acad Child AdolescPsychiatry.
2000;39:49-58.
4. Robb AS, Cueva JE, Sporn J, Yang R, Vanderburg DG. Sertraline
treatment of childrenand adolescents with posttraumatic stress
disorder: a double-blind, placebo-controlledtrial. J Child Adolesc
Psychopharmacol. 2010;20:463-471.
5. Findling RL, Robb A, Nyilas M, et al. A multiple-center,
randomized, double-blind,placebo-controlled study of oral
aripiprazole for treatment of adolescents with schizo-phrenia. Am J
Psychiatry. 2008;165:1432-1441.
6. Findling RL, Pathak S, Earley WR, Liu S, DelBello MP.
Efficacy and safety of extended-release quetiapine fumarate in
youth with bipolar depression: an 8 week,
double-blind,placebo-controlled trial. J Child Adolesc
Psychopharmacol. 2014;24:325-335.
7. Findling RL, Nyilas M, Forbes RA, et al. Acute treatment of
pediatric bipolar I disorder,manic or mixed episode, with
aripiprazole: a randomized, double-blind, placebo-controlled study.
J Clin Psychiatry. 2009;70:1441-1451.
8. Biederman J, Krishnan S, Zhang Y, McGough JJ, Findling RL.
Efficacy and tolerabilityof lisdexamfetamine dimesylate (NRP-104)
in children with attention-deficit/hyperactivity disorder: a phase
III, multicenter, randomized, double-blind,
forced-dose,parallel-group study. Clin Ther. 2007;29:450-463.
9. Schulz E, Fleischhaker C, Hennighausen K, et al. A
double-blind, randomized, placebo/active controlled crossover
evaluation of the efficacy and safety of Ritalin LA in childrenwith
attention-deficit/hyperactivity disorder in a laboratory classroom
setting. J ChildAdolesc Psychopharmacol. 2010;20:377-385.
10. Santisteban JA, Stein MA, Bergmame L, Gruber R. Effect of
extended-release dexme-thylphenidate and mixed amphetamine salts on
sleep: a double-blind, randomized,crossover study in youth with
attention-deficit hyperactivity disorder. CNS Drugs.
2014;28:825-833.
11. Dell’Agnello G, Maschietto D, Bravaccio C, et al.
Atomoxetine hydrochloride in thetreatment of children and
adolescents with attention-deficit/hyperactivity disorder
andcomorbid oppositional defiant disorder: a placebo-controlled
Italian study. Eur Neuro-psychopharmacol. 2009;19:822-834.
12. Emslie GJ, Mayes T, Porta G, et al. Treatment of Resistant
Depression in Adolescents(TORDIA): week 24 outcomes. Am J
Psychiatry. 2010;167:782-791.
13. Brent D, Emslie G, Clarke G, et al. Switching to another
SSRI or to venlafaxine with orwithout cognitive behavioral therapy
for adolescents with SSRI-resistant depression: theTORDIA
randomized controlled trial. JAMA. 2008;299:901-913.
14. Fluvoxamine for the treatment of anxiety disorders in
children and adolescents. TheResearch Unit on Pediatric
Psychopharmacology Anxiety Study Group. N Engl J
Med.2001;344:1279-1285.
15. Runyon MK, Deblinger E, Steer RA. Group cognitive behavioral
treatment for parentsand children at-risk for physical abuse: an
initial study. Child Fam Behav Ther. 2010;32:196-218.
16. Kessler RC, Avenevoli S, Green J, et al. National
comorbidity survey replicationadolescent supplement (NCS-A): III.
Concordance of DSM-IV/CIDI diagnoses withclinical reassessments. J
Am Acad Child Adolesc Psychiatry. 2009;48:386-399.
17. Findling RL, Youngstrom EA, Fristad MA, et al.
Characteristics of children with elevatedsymptoms of mania: the
Longitudinal Assessment of Manic Symptoms (LAMS) study.J Clin
Psychiatry. 2010;71:1664-1672.
18. Axelson D, Birmaher B, Strober M, et al. Phenomenology of
children and adolescentswith bipolar spectrum disorders. Arch Gen
Psychiatry. 2006;63:1139-1148.
19. Lauth B, Levy SR, Juliusdottir G, Ferrari P, Petursson H.
Implementing the semi-structured interview Kiddie-SADS-PL into an
in-patient adolescent clinical setting:impact on frequency of
diagnoses. Child Adolesc Psychiatry Ment Health. 2008;2:14.
20. Sorensen MJ, Nissen JB, Mors O, Thomsen PH. Age and gender
differences indepressive symptomatology and comorbidity: an
incident sample of psychiatricallyadmitted children. J Affect
Disord. 2005;84:85-91.
21. Jensen-Doss A, Youngstrom EA, Youngstrom JK, Feeny NC,
Findling RL. Predictorsand moderators of agreement between clinical
and research diagnoses for children andadolescents. J Consult Clin
Psychol. 2014;82:1151-1162.
22. Miller PR. Inpatient diagnostic assessments: 2. Interrater
reliability and outcomes ofstructured vs. unstructured interviews.
Psychiatry Res. 2001;105:265-271.
23. Kobak KA, Greist JH, Jefferson JW, Katzelnick DJ.
Computer-administered clinicalrating scales. A review.
Psychopharmacology (Berl). 1996;127:291-301.
24. Kobak KA, Taylor LH, Dottl SL, et al. A
computer-administered telephone interview toidentify mental
disorders. JAMA. 1997;278:905-910.
25. Brodey BB, First M, Linthicum J, Haman K, Sasiela JW, Ayer
D. Validation of theNetSCID: an automated web-based adaptive
version of the SCID. Compr Psychiatry.2016;66:67-70.
26. Kaufman J, Birmaher B, Axelson D, Pereplitchikova F, Brent
D, Ryan N. The KSADS-PL DSM-5. Baltimore, MD: Kennedy Krieger
Institute; 2016.
16 www.jaacap.org
FLA 5.6.0 DTD � JAAC2574_proof �
27. Posner K, Oquendo MA, Gould M, Stanley B, Davies M. Columbia
Classification Al-gorithm of Suicide Assessment (C-CASA):
classification of suicidal events in the FDA’spediatric suicidal
risk analysis of antidepressants. Am J Psychiatry.
2007;164:1035-1043.
28. Alexander LM, Escalera J, Ai L, et al. An open resource for
transdiagnostic research inpediatric mental health and learning
disorders. Sci Data. 2017;4:170181.
29. Angold A, Costello EJ, Messer S, Pickles A, Winder F, Silver
D. Development of a shortquestionnaire for use in epidemiological
studies of depression in children and adolescents.Int J Methods
Psychiatr Res. 1995;5:237-249.
30. Thapar A, McGuffin P. Validity of the shortened Mood and
Feelings Questionnaire in acommunity sample of children and
adolescents: a preliminary research note. PsychiatryRes.
1998;81:259-268.
31. Kuo ES, Stoep AV, Stewart DG. Using the short Mood and
Feelings Questionnaire todetect depression in detained adolescents.
Assessment. 2005;12:374-383.
32. Daviss WB, Birmaher B, Melhem NA, Axelson DA, Michaels SM,
Brent DA. Criterionvalidity of the Mood and Feelings Questionnaire
for depressive episodes in clinic andnon-clinic subjects. J Child
Psychol Psychiatry. 2006;47:927-934.
33. Kent L, Vostanis P, Feehan C. Detection of major and minor
depression in children andadolescents: evaluation of the Mood and
Feelings Questionnaire. J Child Psychol Psy-chiatry.
1997;38:565-573.
34. Birmaher B, Khetarpal S, Brent D, et al. The Screen for
Child Anxiety Related EmotionalDisorders (SCARED): scale
construction and psychometric characteristics. J Am AcadChild
Adolesc Psychiatry. 1997;36:545-553.
35. Birmaher B, Brent DA, Chiappetta L, Bridge J, Monga S,
Baugher M. Psychometricproperties of the Screen for Child Anxiety
Related Emotional Disorders (SCARED): areplication study. J Am Acad
Child Adolesc Psychiatry. 1999;38:1230-1236.
36. Muris P, Merckelbach H, Ollendick T, King N, Bogie N. Three
traditional and threenew childhood anxiety questionnaires: their
reliability and validity in a normal adolescentsample. Behav Res
Ther. 2002;40:753-772.
37. Hughes CW. Child and Adolescent Measures for Diagnosis and
Screening. Washington,DC: American Psychiatric Association;
2008.
38. Achenbach TM, Rescorla LA. Manual for ASEBA School-Age Forms
and Profiles.Burlington, VT: University of Vermont, Research Center
for Children, Youth, &Families; 2001.
39. Nakamura BJ, Ebesutani C, Bernstein A, Chorpita BF. A
psychometric analysis of theChild Behavior Checklist DSM-Oriented
Scales. J Psychopathol Behav Assess. 2009;31:178-189.
40. Lucas CP, Fisher P, Piacentini J, et al. Features of
interviews questions associated withattenuation of symptom reports.
J Abnorm Child Psychol. 1999;27:429-437.
41. Bangor A, Kortum P, Miller J. Determining what individual
SUS scores mean: adding anadjective rating scale. J Usability Stud.
2009;4:114-123.
42. Brooke J. SUS: a ‘quick and dirty’ usability scale. In:
Jordan PW, Thomas B,Weerdmeester BA, McClelland IL, eds. Usability
Evaluation in Industry. London: Taylor& Francis;
1996:189-194.
43. Bangor A, Kortum P, Miller J. The system usability scale
(SUS): an empirical evaluation.International Journal of
Human-Computer Interaction. 2008;24:574-592.
44. Kobak KA, Reynolds WM, Griest JH. Computerized and clinician
assessment ofdepression and anxiety: respondent evaluation a