The Relationship Between the Childhood Autism Rating Scale- Second Edition and Clinical Diagnosis Utilizing the DSM-5 and the DSM-IV

Background: The Childhood Autism Rating Scale, Second Edition (CARS2; 2010) includes two rating scales- CARS2-Standard Version (CARS2-ST; formerly titled CARS) and the newly developed CARS2-High Functioning version (CARS2-HF).  The CARS2-ST maintains the integrity of the CARS from 1988, and as such is considered “the strongest, best documented, and most widely used clinical rating scale for behaviors associated with autism” (Lord & Corsello, 2005). The CARS-ST is for use with children less than age 6 or over age 6 with an estimated IQ score of 79 or lower or with notably impaired communication. The original CARS (CARS2-ST) has been shown to have good agreement with the DSM-III-R, though it was somewhat over-inclusive compared to strict criteria (Van Bourgondien, et al., 1992).  Many subsequent studies reported good to excellent agreement between the original CARS (CARS2-ST) and clinical diagnosis using DSM-IV criteria (Perry et al., 2005; Sponheim & Sparkland, 1996).  In the CARS2 verification study and national field trials, the CARS2-ST had a sensitivity of .88 and a specificity of .86 compared to clinical diagnoses at the time of DSM-IV. For the CARS2-HF, the sensitivity was .81 and the specificity was .87.

Objectives:  To assess the specificity and sensitivity of the CARS-2 when compared to DSM-IV versus DSM-5 criteria for ASD within a community based clinic sample.

Methods: Each participant was seen for a diagnostic evaluation at one of 5 community based TEACCH Centers that are part of the University of North Carolina TEACCH Autism Program (TEACCH). The assessments included administration of the CARS2, the Autism Diagnostic Observation Schedule-2 (ADOS-2), and a developmental history interview. All diagnostic conclusions were made on the basis of direct assessment and observation as well as developmental history. In addition, clinicians rated symptoms of autism on both the DSM-IV-TR and DSM-5 criteria for autism on the day of the diagnostic. Preliminary evidence was obtained for 68 participants using the CARS2-ST and 76 participants using the CARS2-HF. These participants are representative of a variety of socioeconomic and racial backgrounds from both rural and urban centers throughout the state. The TEACCH Centers conduct diagnostic evaluations on all ages from toddlers through adults. Thus, the age range of participants rated on the 1.09 to 18.09 (mean= 4.39 years; SD= 2.85) for participants rated on the CARS2-STand CARS2-HF was 5.03 to 45.05 years (mean= 13.75 years; SD= 8.20).

Results:  In preliminary analyses, the sensitivity of CARS2-ST was .74 and .78 across DSM-IV and DSM-5 respectively. The specificity of CARS2-ST was 1.00 across DSM-IV and DSM-5 criteria. The sensitivity of CARS2-HF was 1.00 across DSM-IV and DSM-5 while the specificity was .88 and .77 respectively. 

Conclusions: Preliminary findings suggest that a high proportion of individuals who meet clinical cut-off for autism on the CARS-2 meet criteria for autism based on DSM-IV and DSM-5 criteria. The relationship between CARS2 ratings and different diagnostic subgroups (e.g. PDD, Anxiety, ADHD, etc.) will be presented.