Effectiveness of the STAT and STAT-Q within a Multidisciplinary Assessment for Early Identification of Autism Spectrum Disorder

Background: To date, no consistent assessment protocol is used across health care systems to establish early diagnosis of autism spectrum disorders (ASDs). The American Academy of Pediatrics (AAP) has suggested the use of both Level I screening of all children starting at 18 months using the M-CHAT (Robins et. al., 2001) or similar tools, and comprehensive evaluations for children who screen positive for autism risk. While the AAP recommends ongoing screening starting at 18 months, these guidelines are not being implemented consistently. In Illinois, children who are seen for early diagnosis are assessed by a multidisciplinary team within a Medical Diagnostic Evaluation (MDX) with no consistent protocol. Implementation of a consistent protocol utilizing a standardized level II screener -the Screening Tool for Autism in Toddlers and Young Children (STAT: Stone, Coonrod, Turner & Pozdol, 2004) along with developmental testing and clinical judgment - has the potential to enhance accuracy of ASD diagnosis in young children in a cost-effective manner. In addition, the STAT-Q Parent Questionnaire allows for parent input and increases understanding of the child.

Objectives: This study aims to analyze the specificity and sensitivity of both the STAT and the STAT-Q as part of a consistent protocol for identifying young children with ASD in an effort to develop a standardized screening protocol that is both clinically effective and financially viable.

Methods: 95 children 18 and 36 months of age, both English and Spanish-speaking, were assessed through the MDX utilizing developmental testing, autism specific validated instruments, as well as clinical judgment. Three groups were developed to compare sensitivity and specificity of the STAT and STAT-Q within this multidisciplinary evaluation: Group I- STAT compared to overall diagnosis; Group II-STAT-Q with overall diagnosis; and Group III- STAT-Q with positive screen on the STAT.

Results: Diagnostic Odds Ratios (DOR) was used as a measure of the overall diagnostic power of each tool. Receiver-operating-characteristic (ROC) analysis was performed to assess and compare the sensitivity and specificity of the STAT and STAT-Q when compared to overall diagnosis. Preliminary results show strong sensitivity and specificity of the STAT (97.5% and 73% respectively), with slightly lower sensitivity and specificity of the STAT-Q with clinical diagnosis (approximately 65% and 40% respectively); higher sensitivity when comparing the STAT-Q with the STAT (mothers 78% versus fathers 65%) and lower specificity of the STAT-Q with the STAT (mothers 20% and fathers 36%). A much larger sample (>150) will be available by April 1, 2018 to further describe positive predictive value of using these instruments within a multidisciplinary assessment. We will also examine impact of age, gender, and language on these relations.

Conclusions: This project hopes to demonstrate the effectiveness of the STAT and STAT-Q within multidisciplinary assessment of young children utilizing a standardized tool, including parent concerns, along with clinical judgment. These tools are efficient, sensitive, and cost effective; thus, the project has the potential to contribute to the creation of a standardized protocol for accurate early identification of ASD, allowing children to get into appropriate intervention programs at younger ages.