The Autism Mental Status Exam: Sensitivity and Specificity Using Consensus Diagnosis

Background: The Autism Mental Status Exam (AMSE) is a brief diagnostic observational assessment tool that structures the way we observe and record social, communicative and behavioral functioning in individuals with ASD (Grodberg, 2010).  The AMSE comprises 8 operationalized items, which include: eye contact, interest in others, pointing skills, language, pragmatics, stereotypy, preoccupations, and sensitivities.  The strong clinical utility of the AMSE derives from its low clinical burden.  The AMSE does not add extra work to a clinician’s exam, but rather structures the way data is observed and recorded.  Each item is scored on a 0-2 scale yielding total scores that range from 0-16.  A preliminary study indicates that the AMSE has excellent inter-rater reliability and classification accuracy when compared to the Autism Diagnostic Observation Schedule (ADOS) in a high-risk population. This is the first study to assess the sensitivity and specificity of the AMSE using consensus diagnosis.

Objectives: To investigate the sensitivity and specificity of the AMSE using current DSM-IV-TR criteria and the proposed DSM-5 criteria for autism. This study seeks to determine the most effective cutoffs that predict clinical consensus diagnosis supported by ADOS and ADI-R in a high-risk population.

Methods: 136 children, adolescents and adults between age 19-months and 45-years received comprehensive autism diagnostic evaluations. Evaluations included an initial intake performed by a child and adolescent psychiatrist or a psychologist with autism expertise; the AMSE was administered in the context of this clinical evaluation.  Participants were then administered an ADOS by a different clinician and, in most cases, an ADI-R was performed.  Consensus diagnosis was ascertained by a team of clinicians using data from the clinical evaluation and diagnostic assessments.  Subjects were grouped into two diagnostic categories:  ASD and non-ASD.  In an effort to assess the AMSE’s predictive validity using current DSM-IV-TR criteria and proposed DSM-5 criteria, Analysis #1 grouped all PDD-NOS cases into the ASD category and Analysis #2 grouped PDD-NOS cases into the non-ASD category.

Results: Within this high-risk sample, 68% of subjects received a clinical diagnosis of Autistic Disorder or Asperger’s Disorder, 12% received a diagnosis of PDD-NOS, and 20% received other axis I diagnoses.  ROC curve analysis was used to determine the most effective criterion cut-off scores on the AMSE. In Analysis #1, which reflects DSM-IV-TR criteria, area under the ROC curve (AUC) was 0.91 (95%CI: 0.858-0.971). An AMSE cutoff score of ≥5 predicted ASD diagnosis with sensitivity of 90% and specificity of 89%. In Analysis #2, which more closely reflects proposed DSM-V criteria, the AUC was 0.88 (95%CI: 0.809–0.946). An AMSE score of ≥6 predicted ASD diagnosis with sensitivity of 79% and specificity of 79%.

Conclusions: This study demonstrates that the AMSE provides a standardized method to rapidly assess signs and symptoms of ASD. Findings indicate that the AMSE holds promise as a useful observational assessment tool for use with individuals suspected of ASD. Findings support the validity of the AMSE and add to previous results suggesting that the AMSE accurately predicts ASD diagnosis, even when rigorous standards for diagnosis are applied.