RESEARCH ARTICLE

A Simplified Diagnostic Observational Assessment of Autism

Spectrum Disorder in Early Childhood
David Grodberg, Paige Siper, Jesslyn Jamison, Joseph D. Buxbaum, and Alexander Kolevzon

Subspecialty physicians who have expertise in the diagnosis of autism spectrum disorder typically do not have the
resources to administer comprehensive diagnostic observational assessments for patients suspected of ASD. The autism
mental status exam (AMSE) is a free and brief eight-item observation tool that addresses this practice gap. The AMSE,
designed by Child and Adolescent Psychiatrists, Developmental Behavioral Pediatricians and Pediatric Neurologists
structures the observation and documentation of signs and symptoms of ASD and yields a score. Excellent sensitivity
and specificity was demonstrated in a population of high-risk adults. This protocol now investigates the AMSE’s test per-
formance in a population of 45 young children age 18 months to 5 years with suspected ASD or social and communica-
tion concerns who are evaluated at an autism research center. Each subject received a developmental evaluation,
including the AMSE, performed by a Child and Adolescent Psychiatrist, that was followed by independent standardized
assessment using the Autism Diagnostic Observation Schedule and the Autism Diagnostic Interview-Revised. A Best Esti-
mate Diagnosis protocol used DSM-5 criteria to ascertain a diagnosis of ASD or non-ASD. Receiver operating characteris-
tic curve analysis was used to determine the AMSE cut point with the highest sensitivity and specificity. Findings
indicate an optimized sensitivity of 94% and a specificity of 100% for this high prevalence group. Because of its high
classification accuracy in this sample of children the AMSE holds promise as a tool that can support both diagnostic
decision making and standardize point of care observational assessment of ASD in high risk children. Autism Res
2016, 9: 443–449. V C 2015 International Society for Autism Research, Wiley Periodicals, Inc.

Keywords: Autism diagnosis; Autism mental status exam; Observational assessment

Introduction resources or training to administer such comprehensive

assessment tools. The ADI-R can take several hours to
Diagnostic assessment for children suspected of autism administer and is rarely used in clinical practice and
spectrum disorder has significantly evolved over the the ADOS-2 requires rigorous reliability training, time,
past decade, especially with the development of stand- and expense. This relative lack of standardized ASD
ardized assessment tools such as the Autism Diagnostic assessment in real world settings may contribute to sig-
Observation Schedule, second edition (ADOS-2) [Lord nificant variability in diagnostic practice across
et al., 2012], the Autism Diagnostic Interview-Revised populations.
(ADI-R) [Rutter, Le Couteur, & Lord, 2003], and their With only a limited capacity to administer existing
new versions for use in toddlers [Kim, Thurm, Shum- observational instruments with discriminative ability,
way, & Lord, 2013]. Such diagnostic tools, which stand- the physician must rely on his or her subjective sense
ardize the collection of observational and reported data of certainty when considering a diagnosis of ASD. Such
have proven to increase diagnostic accuracy in autism an approach can be accurate when evaluating high cer-
research centers and autism specialty clinics in children tainty, uncomplicated cases. However, diagnostically
as young as 24 months [Johnson, Myers, & American complex cases in which there is a possibility of ASD but
Academy of Pediatrics Council on Children With Dis- also a strong clinical suspicion of alternate diagnoses
abilities, 2007]. that may overlap with ASD presentation such as Intel-
However, many subspecialty physicians who have lectual Disability, Communication disorders, Attention
expertise in ASD diagnosis, such as Child and Adoles- Deficit Hyperactivity Disorder (ADHD), Specific learning
cent Psychiatrists, Developmental Behavioral Pediatri- disorder, Anxiety Disorder, Post Traumatic Stress Disor-
cians, and Pediatric Neurologists do not have the der, and Schizophrenia require assessment using

From the Seaver Autism Center for Research and Treatment, Icahn School of Medicine at Mount Sinai, 1 Gustave L. Levy Place, Box 1230,
New York, NY
Grant sponsor: Beatrice and Samuel Seaver Foundation, NIH; Grant number: KL2 TR000069.
Received March 01, 2015; accepted for publication July 25, 2015
Address for correspondence and reprints: David Grodberg, Seaver Autism Center for Research and Treatment, Icahn School of Medicine at Mount
Sinai, 1 Gustave L. Levy Place, Box 1230, New York, NY 10029. E-mail: [email protected]
Published online 25 August 2015 in Wiley Online Library (wileyonlinelibrary.com)
DOI: 10.1002/aur.1539
C 2015 International Society for Autism Research, Wiley Periodicals, Inc.
V

INSAR Autism Research 9: 443–449, 2016 443

diagnostic tools that have low false positive rates [Con- and specificity in detecting an independent DSM-5
stantino et al., 2012]. Indeed, clinicians who work out- diagnosis of ASD in adults also indicated strong psycho-
side autism specialty clinics are more likely to see such metric properties (SE: 0.91, SP: 0.93) [Grodberg et al.,
diagnostically complex patients from more diverse 2014]. A number of free and commercially available
social and demographic populations [Corsello, parent report tools with varying degrees of sensitivity
Akshoomoff, & Stahmer, 2013]. and specificity have been developed [Volkmar et al.,
Recent studies indicate that even when a definitive 2014]. Two of the most common parent report tools
diagnosis of ASD is not possible, or when ASD is confi- include the Modified Checklist for Autism in Toddlers
dently ruled out, reliable documentation of subthres- (M-CHAT-R) and the Childhood Autism Spectrum Test
hold social and communicative impairments along (CAST). Parent report tools are easily integrated into
with repetitive and stereotyped behaviors is critical as clinical practice as they are administered outside the
such comorbidities are associated with poor develop- clinical encounter (e.g., in the waiting room). They can
mental outcomes and quality of life. For instance, chil- be scored and interpreted with minimal training, but
dren with other DSM-5 diagnoses such as ADHD they are vulnerable to parents’ over-or under-reporting
complicated by ASD signs and symptoms have more of their child’s signs and symptoms of ASD [Johnson,
psychopathology, lower adaptive functioning, and are Filliter, & Murphy, 2009].
more likely to be treated with medication [Frazier et al., Two observation tools are commercially available and
2011; Gadow, DeVincent, & Schneider, 2009; Grzadzin- can be used as screening instruments for ASD: The
ski et al., 2011; Mulligan et al., 2009]. Screening Tool for Autism in Toddlers (STAT) [Stone,
The Autism Mental Status Exam (AMSE) was devel- Coonrod, Turner, & Pozdol, 2004] and the Childhood
oped to address this practice gap. The AMSE is a free Autism Rating Scale (CARS) [Schopler, Reichler, DeVel-
standardized diagnostic observation tool that structures lis, & Daly, 1980]. The STAT consists of 12 interactive
the observation and documentation of 8 social, com- items, takes 20 min to administer, and can be used in
municative, and behavioral signs and symptoms of ASD children age 12–36 months. The CARS, one of the most
that typically emerge throughout a neurodevelopmen- widely used assessment scales, was created before the
tal evaluation. The AMSE was developed through a col- current diagnostic classification system and typically is
laborative process that included child and adolescent used now as a general rating of all information avail-
psychiatrists, developmental behavioral pediatricians able from the history and observation [Gotham, 2011].
and pediatric neurologists to ensure feasibility and low The CARS is commercially available, can be used in
clinical burden at the point of care. While the AMSE children over age 2 years as a screening tool, and takes
structures documentation of direct observations, it also about 20 minutes to administer.
provides opportunity to record clinical information Structured data collection using the AMSE can sup-
that is reported by the patient or caregiver, regardless of port observational assessment that in turn, can effi-
whether the behavior is observed during the exam. This ciently and accurately support diagnostic decision
integration of reported data into an observational making in children suspected of ASD. While the AMSE
assessment is necessary, as has been demonstrated by cannot be used independently, given the lack of vali-
studies examining the validity of the ADOS and ADI-R, dated diagnostic observation tools for ASD and the
which indicate that the combination of observed signs complexity of assessing ASD in children, the AMSE
and reported symptoms provides the best diagnostic holds promise as a point of care tool that can guide
accuracy [Kanne, Abbacchi, & Constantino, 2009; Kim clinical judgment when considering a DSM-5 diagnosis
& Lord, 2012; Risi et al., 2006]. The AMSE does not add of ASD and subsequent referral and treatment decisions.
extra work, but instead structures the way naturally More efficient and accurate diagnosis in high-risk pedi-
evolving data is observed and charted. The eight items atric populations is necessary to promote appropriate
include: (1) eye contact, (2) interest in others, (3) point- referrals to evidence based psychiatric, psychological,
ing skills, (4) language, (5) pragmatics, (6) repetitive and educational services as well as research protocols. A
behaviors, (7) preoccupations, and (8) unusual sensitiv- standardized approach to data collection can
ities. Each item is scored on a 0-2 scale with possible strengthen efforts to measure ASD signs and symptoms
total scores ranging from 0 to 14. The AMSE can be eas- within and between different clinical and demographic
ily built into the electronic health record and prelimi- populations.
nary validation of the AMSE in a large, high-risk, The current study aims to establish the diagnostic
sample that ranged from toddlers to adults suspected of accuracy of the AMSE in a population of young chil-
ASD indicated excellent inter-rater reliability and strong dren who were evaluated for possible ASD or social/lan-
classification accuracy when compared to the ADOS guage concerns at an autism research center. This study
[Grodberg, Weinger, Kolevzon, Soorya, & Buxbaum, was driven by the need for a free and low-burden stand-
2012]. Further investigation of the AMSE’s sensitivity ardized diagnostic observation tool that ultimately can

444 Grodberg et al./A Simplified Diagnostic Observational Assessment INSAR

Table 1. Summary of AMSE Items and Scoring Guidelines
Item score 0 1 2
a
Eye contact >3 sec Fleeting None
Interactionsa Spontaneous Passive None
Pointinga Can point/gesture Only follows point None
Languageb Complex sentences Undeveloped sentences, phrases, single words None
Pragmaticsb,c No impairment Reported impairment Observed impairment
Repetitive behaviorsb None Compulsions Stereotypy
Preoccupationsb,c None Reported preoccupations Observed preoccupations
Sensitivitiesb,c None Reported symptoms Observed signs

Complete scoring instructions are available online at: www.autismmentalstatusexam.com

a
Social items must be observed by the examiner.
b
Communicative and behavioral items may be reported or observed.
c
Pragmatics, Preoccupations, and Sensitivities are weighted if observed.

be seamlessly integrated into the subspecialty physi- 29% were latino, and 11% meet criteria for Intellectual
cian’s clinical workflow. To date, there are no known Disability.
diagnostic observation tools that have been validated to Informed consent was obtained from all legal guardi-
the DSM-5 criteria for ASD, that are in the public ans. Participants were at higher-risk for ASD than the
domain, and that can be seamlessly integrated into the general population as they were referred by their
clinical encounter and electronic health record. parents, school psychologists, or pediatricians who
This study follows the reporting guidelines set forth learned about various research protocols at the Center,
by the STARD initiative, which provides guidance for including genetics, clinical trials, and community
the development and reporting of diagnostic test devel- engaged interventions. All had suspected ASD or social/
opment (http://www.stard-statement.org) [Bossuyt communication concerns.
et al., 2003]. Such guidelines are intended to improve
the completeness and transparency of reporting of stud-
ies of diagnostic accuracy. Additionally, the guidelines Materials
provide a framework for readers to use to assess threats Autism Mental Status Exam
to internal validity (i.e., bias) and to external validity
(i.e., generalizability). The AMSE is a free and widely available 8-item observa-
tional tool that prompts the examiner to observe and
document patients’ social, communicative and behav-
Methods ioral functioning in the context of a developmentally
Participants focused clinical examination. The AMSE is intended to
The source population for this study included all guide clinical judgment in the context of diagnostic
patients who received comprehensive autism-focused decision making. Each item is scored on a 0–2 scale
diagnostic evaluations as part of the Institutional with possible total scores ranging from 0 to 14; higher
Review Board (IRB) at the Icahn School of Medicine at scores reflect greater severity. Social items must be
Mount Sinai approved assessment protocol at the Seaver observed during the clinical exam, but communication
Autism Center for Research and Treatment from Sep- and behavioral items can be reported or observed.
tember 2013 through December 2014. The sample pop- Three items—pragmatics of language, encompassing
ulation, which was derived from the source population, preoccupations, and unusual sensitivities—prompt the
included all children 5 years and under who did not examiner to specify whether the item is reported or
have fluent language skills (n 5 45). This inclusion crite- observed. In these three items, the score is weighted if
ria was implemented by enrolling only children who the item is observed. Scoring instructions for those
were administered the ADOS-2 Toddler Module, Module three items also provide flexibility for lower and higher
1, or Module 2 as part of the IRB approved diagnostic functioning individuals. An online training curriculum
assessment protocol. The participants’ language thus provides the scoring manual and video simulation of
ranged from nonverbal to undeveloped sentences. clinical examinations based on individuals of varying
None of the participants had fluent speech. The age ages and levels of functioning. Validated translations
range for this sample population is 18 months to 60 are available in numerous languages. A summary of
months (M 5 41.1 months, SD 5 12.5). Seventy-eight AMSE items and scoring guidelines can be found in
percent of subjects were male, 71% were Caucasian, Table 1.

INSAR Grodberg et al./A Simplified Diagnostic Observational Assessment 445

Autism Diagnostic Observation Schedule-Second Edition Analytic Methods

The ADOS-2 is a semi-structured observational assess- The receiver operating characteristics (ROC) curve is a
ment that is used to assess the presence of autism method that depicts the tradeoff between the true-
symptomatology within two domains: social- positive rate and the false-positive rate of a diagnostic
communication and repetitive, restricted behaviors. test. We implemented an ROC curve analysis to simul-
The ADOS-2 Toddler Module, Module 1, and Module 2, taneously display the line representing each cut point
are intended for individuals who do not have fluent and its associated true positive rate (sensitivity) and
speech. These modules were administered to all partici- false positive rate (1-specificity). The point on the line
pants in this study sample by trained clinicians. that is farthest away from the true diagonal indicates
the cut point with the highest sensitivity and
Autism Diagnostic Interview-Revised
specificity.
The ADI-R is a structured caregiver interview that is The calculated area under the curve (AUC) was used
used for diagnostic purposes. The ADI-R probes for cur- to reflect the accuracy of the test, which indicates how
rent behaviors and a developmental history consistent well the test separates those individuals with ASD from
with autism symptomatology based on questions in the those without ASD. An AUC of 1.0 reflects a perfect test
following domains: early development, communica- and an AUC of 0.5 reflects a test that is no more accu-
tion, reciprocal social interaction, and repetitive, rate than flipping a coin.
restricted patterns of behavior. The ADI-R was adminis- Correlation with the ADOS-2 comparison score [Lord,
tered and scored by a research reliable clinician. 2012] was also investigated. This analysis included only
subjects who were administered ADOS-2 module 1 and
Procedure
module 2 because the toddler module does not have a
Each participant first received a clinical evaluation by a comparison score. As both AMSE scores and ADOS-2
board certified Child and Adolescent Psychiatrist or comparison scores were normally distributed, we used
Developmental Behavioral Pediatrician with extensive the parametric Pearson correlation test.
experience in the diagnosis of ASD. The eight AMSE As we had a fixed population, we did not power the
items were used to structure the physician’s observation study prospectively. Instead, we relied on the calculated
and recording of signs and symptoms of ASD. Partici- confidence interval to signify the precision of our
pants were then administered the ADOS-2 on the same results. Regression methods were not utilized in this
visit. The ADOS-2 was administered by independent analysis.
psychologists at the Center who were blind to the
AMSE score and blind to the physicians’ diagnostic Results
impressions. An ADI-R was administered at a follow up
appointment for all participants who were above the Seventy-three percent of the 45 participants met BECD
cutoff on the ADOS-2 and for cases that were one point for ASD using DSM-5 criteria. Diagnostic accuracy of
below on the ADOS-2. All ADOS-2 and ADI-R adminis- the AMSE was assessed by the nonparametric measure
trations were scored by reliable raters. In order to ascer- of area under a receiver operating characteristic (ROC)
tain the DSM-5 clinical diagnosis in a way that is curve. The ROC curve analysis was used to determine
sufficiently independent from the AMSE score, a best the optimal cutoff for the AMSE compared to the BECD
estimate clinical diagnosis (BECD) protocol was imple- based on DSM-5. Area under the ROC curve (AUC) was
mented in which the supervising psychologist at the 0.99 [95% confidence interval (CI) 0.98–1.00] (Fig. 1).
Center reviewed the full ADOS-2 protocol, the ADI-R The most effective cutoff score was estimated at a
protocol, and the developmental history, which was total score of 6. This cutoff score produced a sensitivity
documented using a standardized intake form listing of 94% and a specificity of 100% (Table 2).
chief complaint, history of present illness, past medical The relationship between AMSE total scores and
history, and developmental milestones. In some cases, ADOS-2 classification was also examined. Eighty per-
the BECD clinician communicated directly with the cent of participants met criteria for ASD on the ADOS-
independent ADOS-2 or ADI-R examiner to gather addi- 2. Area under the ROC curve was 0.95 (95% CI 0.88–
tional information. The BECD clinician was thus unable 1.00) and a score of 6 produced an optimal cutoff with
to gather additional information from the physician a sensitivity of 86% and a specificity of 100%.
administering the AMSE, and did not have access to the AMSE total scores for participants meeting BECD
AMSE score. DSM-5 criteria were then used to guide the DSM-5 ASD criteria ranged from 5 to 13, while total
BECD clinician’s diagnostic formulation of ASD versus scores for participants who did not meet BECD DSM-5
non-ASD. ASD criteria ranged from 1 to 5. On the ADOS-2, AMSE

446 Grodberg et al./A Simplified Diagnostic Observational Assessment INSAR

 Table 3. Axis I Diagnoses and AMSE Scores Assigned to
Non-ASD Patients
Patient Diagnosis AMSE Score

1 Language disorder 4
2 Language disorder 3
3 Language disorder 2
4 Unspecified anxiety disorder 4
5 Unspecified anxiety disorder 5
6 Language disorder 5
7 Stereotypic movement disorder 1
8 No diagnosis 3
9 Language disorder 3
10 Language disorder 4
11 Language disorder 4
12 Unspecified anxiety disorder 3

Among the 12 children who were diagnosed as having non-ASD condi-

tions, AMSE scores ranged from 1 to 5.

A Pearson correlation was used to examine the rela-

tionship between total scores on the AMSE and scores
on the ADOS-2 severity metric. Results indicate that
Figure 1. ROC curve. Receiver operating characteristic (ROC) AMSE and ADOS-2 severity metric are significantly cor-
curve for the AMSE.
related (r 5 0.721).

Table 2. Sensitivity and Specificity of AMSE Cutoff Scores for Discussion

a Best Estimate Clinical Diagnosis of ASD Based on DSM-5
Criteria
The AMSE is the first brief and free diagnostic observa-
AMSE cutoff Sensitivity Specificity
tion tool for ASD designed by physicians to ultimately
1 1.00 0.00 be integrated into the clinical workflow and electronic
2 1.00 0.08 health records of Child and Adolescent Psychiatrists,
3 1.00 0.17 Developmental Behavioral Pediatricians, and Pediatric
4 1.00 0.50
Neurologists. It provides clinicians who have expertise
5 1.00 0.83
6 0.94 1.00 in diagnosing neurodevelopmental delays with an inno-
7 0.79 1.00 vative approach that structures the documentation of
8 0.42 1.00 signs and symptoms of ASD without interfering with or
9 0.30 1.00 adding burden to the clinical exam. The strong correla-
10 0.15 1.00
tion between AMSE total scores and ADOS-2 calibrated
11 0.06 1.00
12 0.03 1.00 severity scores indicate that the AMSE may be a useful
tool in settings not equipped to conduct more compre-
In our study sample, AMSE cutoff scores and their associated sensitiv- hensive evaluations. This study represents the first
ity and specificity indicate that a cutoff score of 6 has a sensitivity of examination of the AMSE’s test performance in a sam-
.94 and a sensitivity of 1.00 in detecting children who ultimately were
ple of young children who were evaluated for possible
diagnosed as having ASD [95% confidence interval (CI) 0.98–1.00].
ASD or social/language concerns at an autism research
center using the DSM-5 criteria. The present study
total scores for participants meeting ADOS-2 classifica- investigates the effectiveness of the AMSE’s ability to
tion for ASD ranged from 3 to 13, while AMSE total differentiate between patients who meet criteria for
scores for participants who did not receive an ADOS ASD using a gold standard research diagnostic protocol
classification of ASD ranged from 1 to 5. Of those par- and those who do not. Results from this exploratory
ticipants who did not meet DSM-5 ASD criteria, all study determined that a total score of 6 is the optimal
received other Axis I diagnoses. Table 3 displays AMSE AMSE cutoff for this sample of children. Findings add
scores and diagnoses for all participants. 15% of chil- to our previous work, which indicated that an AMSE
dren carried a previous diagnosis of ASD given by a total score of 5 best predicted ASD classification in a
physician, 78% had received services through early sample of verbal adults who had suspected ASD or
intervention without a medical diagnosis, and 7% had social/communication concerns. While we recognize
no diagnosis and received no services. that test performance (i.e., sensitivity and specificity) is

INSAR Grodberg et al./A Simplified Diagnostic Observational Assessment 447

regarded as an intrinsic characteristic of the test and support the final BECD diagnosis of ASD. In cases that
not dependent upon context, such an assumption scored at or near the ASD classification threshold on
depends upon the test being the same across different the ADOS or those that were considered complicated,
spectrum of disease. It is known that a broader spec- the BECD clinician reviewed specific clinical material
trum of disease, as found in young children compared (e.g., chief complaint, history of present illness, psychi-
to verbal adults, can decrease sensitivity and specificity atric, medical and developmental history). It could be
of a cut point. It is not surprising, then, that a higher argued that these data may be vulnerable to bias. For
cut point on the AMSE reflects a more optimal sensitiv- instance, a subject’s lower AMSE score may influence
ity and specificity in our study sample. the physician to under-report ASD signs and symptoms
The excellent sensitivity and specificity found in this in the clinical material. To minimize the plausability of
study supports both the clinical utility of the AMSE in such bias, the aforementioned clinical material was
guiding clinical judgment at the point of clinical care documented using a standardized clinical intake form
as well as its ability to support ASD diagnoses in the that structured data collection (e.g., “Age of first
context of autism research protocols. While represent- word”). As mentioned above, the BECD clinician did
ing a stepwise progression from previous test perform- not have access to AMSE scores and was unable to con-
ance data, this study firmly establishes the foundation sult with the physician who administered the AMSE.
for more extensive validation in populations that vary This research indicates that the strong psychometric
in ASD prevalence and spectrum of disease. This properties of the AMSE can support diagnostic decision
research provides justification for replication at other making in a population of children with suspected ASD
tertiary care centers as well as at community based set- or social/communication concerns at an autism
tings where children with social and communicative research center. Additionally, as a low burden diagnos-
concerns may have more complicated clinical presenta- tic observation tool that is in the public domain, the
tions characterized by intellectual disability, communi- AMSE holds promise to support improved documenta-
cation disorders, ADHD, learning disorders, and tion of children’s social and communicative develop-
anxiety. ment across clinical and demographic populations.
While this protocol demonstrates the AMSE’s high
classification accuracy, several limitations must be
addressed. Participants were referred to the Center References
because they were suspected of having ASD or social/ Bossuyt, P.M., Reitsma, J.B., Bruns, D.E., Gatsonis, C.A.,
language concerns and are thus at higher risk than the Glasziou, P.P., Irwig, L.M., . . . Standards for Reporting of
general population. Given the relatively small sample Diagnostic Accuracy. (2003). Towards complete and accu-
size, the generalizability of results remains limited and rate reporting of studies of diagnostic accuracy: The STARD
future studies must examine the use of the AMSE in initiative. BMJ, 326(7379), 41–44.
larger high risk samples that are further stratified by Constantino, J.N., Zhang, Y., Abbacchi, A.M., Calhound, A.,
level of language, cognitive and adaptive functioning as Scofield, F., Grafeman, S.J. (2012). Rapid phenotyping of
autism spectrum disorders: Inclusion of direct observation
well as other neurodevelopmental disorders such as
in feasible paradigms for clinical assessment. Neuropsychia-
communication disorders, ADHD, learning disorders
try, 2(3), 203–212.
and anxiety.
Corsello, C.M., Akshoomoff, N., & Stahmer, A.C. (2013). Diag-
Considering our false positive rate of 8%, it remains nosis of autism spectrum disorders in 2-year-olds: A study
critical that in cases where a clinician does not have of community practice. Journal of Child Psychol Psychiatry,
expertise in ASD diagnosis or in cases that have a com- 54(2), 178–185.
plicated clinical presentation, the AMSE must not sub- Frazier, T.W., Shattuck, P.T., Narendorf, S.C., Cooper, B.P.,
stitute for more comprehensive diagnostic assessment. Wagner, M., & Spitznagel, E.L. (2011). Prevalence and cor-
In such cases referral for further work up must be made. relates of psychotropic medication use in adolescents with
Importantly, the level of sensitivity and specificity an autism spectrum disorder with and without caregiver-
found in this study is dependent on reliable use of the reported attention-deficit/hyperactivity disorder. Journal of
AMSE. Future research aims to determine the AMSE’s Child and Adolescent Psychopharmacology, 21(6), 571–579.
Gadow, K.D., DeVincent, C.J., & Schneider, J. (2009). Compara-
test performance when administered by clinicians with
tive study of children with ADHD only, autism spectrum dis-
varying levels of expertise in ASD diagnosis.
order 1 ADHD, and chronic multiple tic disorder 1 ADHD.
Another potential limitation that must be addressed
Journal of Attention Disorders, 12(5), 474–485.
in any report of diagnostic test accuracy includes Gotham, K., Bishop, S.L., & Lord, C. (2011). Diagnosis of
threats to internal validity. In this case, one must exam- Autism Spectrum Disorders. Autism Spectrum Disorders
ine the independence of the reference standard (BECD) (pp. 30–43). New York: Oxford University Press.
from the index test (AMSE). To address potential bias, Grodberg, D., Weinger, P.M., Halpern, D., Parides, M.,
we used the independent ADOS-2 and ADI-R scores to Kolevzon, A., & Buxbaum, J.D. (2014). The autism mental

448 Grodberg et al./A Simplified Diagnostic Observational Assessment INSAR

 status exam: Sensitivity and specificity using DSM-5 criteria algorithms for toddlers and young preschoolers. Journal of
for autism spectrum disorder in verbally fluent adults. Jour- Autism and Developmental Disorders, 43(7), 1527–1538.
nal of Autism and Developmental Disorders, 44(3), 609–614. Lord, C., Rutter, M., DiLavore, P.C., Risi, S., Gotham, K.,
Grodberg, D., Weinger, P.M., Kolevzon, A., Soorya, L., & Bishop, S.L. (2012). Autism Diagnostic Observation Sched-
Buxbaum, J.D. (2012). Brief report: The Autism Mental Sta- ule. (2nd ed., ADOS-2). Torrance, CA: Western Psychologi-
tus Examination: Development of a brief autism-focused cal Services.
exam. Journal of Autism and Developmental Disorders, Mulligan, A., Anney, R.J., O’Regan, M., Chen, W., Butler, L.,
42(3), 455–459. Fitzgerald, M., . . . Gill, M. (2009). Autism symptoms in
Grzadzinski, R., Di Martino, A., Brady, E., Mairena, M.A., Attention-Deficit/Hyperactivity Disorder: A familial trait
O’Neale, M., Petkova, E., . . . Castellanos, F.X. (2011). Exam- which correlates with conduct, oppositional defiant, lan-
ining autistic traits in children with ADHD: Does the guage and motor disorders. Journal of Autism and Develop-
autism spectrum extend to ADHD? Journal of Autism and mental Disorders, 39(2), 197–209.
Developmental Disorders, 41(9), 1178–1191. Risi, S., Lord, C., Gotham, K., Corsello, C., Chrysler, C.,
Johnson, C.P., Myers, S.M., & American Academy of Pediatrics Szatmari, P., . . . Pickles, A. (2006). Combining information
Council on Children With Disabilities. (2007). Identifica- from multiple sources in the diagnosis of autism spectrum
tion and evaluation of children with autism spectrum disor- disorders. Journal of American Academy of Child Adoles-
ders. Pediatrics, 120(5), 1183–1215. cent Psychiatry, 45(9), 1094–1103.
Johnson, S.A., Filliter, J.H., & Murphy, R.R. (2009). Discrepan- Rutter, M., Le Couteur, A., & Lord, C. (2003). ADI-R. Autism
cies between self- and parent-perceptions of autistic traits Diagnostic interview-revised. Los Angeles: Western Psycho-
and empathy in high functioning children and adolescents logical Services.
on the autism spectrum. Journal of Autism and Develop- Schopler, E., Reichler, R.J., DeVellis, R.F., & Daly, K. (1980).
mental Disorders, 39(12), 1706–1714. Toward objective classification of childhood autism: Child-
Kanne, S.M., Abbacchi, A.M., & Constantino, J.N. (2009). hood Autism Rating Scale (CARS). Journal of Autism and
Multi-informant ratings of psychiatric symptom severity in Developmental Disorders, 10(1), 91–103.
children with autism spectrum disorders: The importance Stone, W.L., Coonrod, E.E., Turner, L.M., & Pozdol, S.L.
of environmental context. Journal of Autism and Develop- (2004). Psychometric properties of the STAT for early
mental Disorders, 39(6), 856–864. autism screening. Journal of Autism and Developmental
Kim, S.H., & Lord, C. (2012). Combining information from Disorders, 34(6), 691–701.
multiple sources for the diagnosis of autism spectrum disor- Volkmar, F., Siegel, M., Woodbury-Smith, M., King, B.,
ders for toddlers and young preschoolers from 12 to 47 McCracken, J., State, M., . . . Adolescent Psychiatry Commit-
months of age. Journal of Child Psychology and Psychiatry, tee on Quality Issues (2014). Practice parameter for the
53(2), 143–151. assessment and treatment of children and adolescents with
Kim, S.H., Thurm, A., Shumway, S., & Lord, C. (2013). Multisite autism spectrum disorder. Journal of American Academy of
study of new autism diagnostic interview-revised (ADI-R) Child Adolescent Psychiatry, 53(2), 237–257.

INSAR Grodberg et al./A Simplified Diagnostic Observational Assessment 449