The Autism Mental Status Exam: Psychometric Validity of a Brief Screening Tool

Background:   The American Academy of Pediatrics (AAP) recommends screening for autism spectrum disorder (ASD) at 18-month well-child visits, but implementation in pediatric settings is slow with fewer than 10% of pediatricians integrating an ASD-specific screen (Barton et al., 2012; Dosreis, Weiner, Johnson, & Newschaffer, 2006).  A multitude of screening tools exist; however, some lack sound psychometric properties, others are cumbersome to integrate into pediatric settings, and the majority are parent-report questionnaires.

The Autism Mental Status Exam (Grodberg, 2011) is an 8-item observational assessment requiring clinician rating of highly predictive ASD symptoms resulting in a global risk score. The AMSE is not intended to add extra work to a clinician’s exam, but rather structures the way that data are observed and documented; thus it carries minimal clinical burden. It is intended for use by a variety of health professionals and proposed for many functions, including rapid assessment of ASD profiles for clinical and research purposes, supplement to ASD diagnostic evaluations, and potentially use as a screening tool. An initial validation study indicated the AMSE had high classification accuracy when compared to the Autism Diagnostic Observation Schedule (Grodberg, Weinger, Kolevzon, Soorya, & Buxbaum, 2012).

Objectives:   The current study contributes to the psychometric investigation of the AMSE as a tool to assess autism symptomatology in children seen in diverse settings.  Specifically, the AMSE’s convergent validity will be verified with comparisons to an empirically validated measure of ASD symptoms.

Methods: The study includes children presenting for an initial visit in three community settings: developmental behavior pediatric clinic, an autism clinic, and a preschool program.  Participants included were those whose health care provider completed the AMSE during the initial assessment and whose parents completed the Social Responsiveness Scale, 2^nd Edition (SRS-2, Constantino & Gruber, 2012).  The sample (n=59) was predominantly male (70%), racially diverse (22% Hispanic, 22% Caucasian, 14% African American, and 8% multi-racial), with a median age of 61 months.

Results: AMSE Total Score was significantly correlated with the SRS-2 Total Score (r(56) = .61, p < .01).  Furthermore, the AMSE’s classification of risk for ASD was significantly related to classification of risk on the SRS-2 (X² (1, N = 59) = 18.88, p < .01).  Construct validity of specific items on the AMSE was supported through significant relationships with related subdomains on the SRS-2.  Specifically, Interest in Others on the AMSE was significantly related to SRS-2 subdomain Social Awareness (r(57) = .34, p = .01), Pragmatics of Language was significantly correlated with Social Communication subdomain on the SRS-2 (r(57) = .35, p = .01), and the SRS-2 subdomain Restricted Interests and Repetitive Behavior was significantly correlated with both Repetitive Behaviors/Stereotypy (r(56) = .49, p < .01) and Unusual or Encompassing Preoccupations (r(55) = .40, p < .01) items on the AMSE.

Conclusions: The AMSE demonstrated strong convergent and construct validity with the SRS-2 within a diverse sample.  Based on continued evidence of sound psychometrics, further research is recommended to explore its validity and clinical utility as a brief ASD-screening tool in pediatric settings.