Validation of an Integrated Autism Screening Model Using DSM-5, CARS-2, and CASD Among Children

<title>Abstract</title> <p>This study examined the validity and diagnostic of an integrated autism screening model using DSM-5, CARS-2 and the CASD among children aged 4 to 6 years in Sabah, Malaysia. ASD identification remains challenging in Sabah because many families, particularly in the rural districts, experience limited access to developmental specialists, intervention centres, and early diagnostic services. Combining DSM-5, CARS-2, and CASD could provide a reliable and consistent approach for identifying autism-related characteristics and severity classifications. A quantitative cross-sectional design was used. 50 children referred for suspected ASD were recruited through purposive sampling from four autism intervention centres in Sabah. Data were collected through multidisciplinary developmental assessments, behavioural observations, and clinical reports. Descriptive statistics, Spearman’s rank correlation, and weighted Cohen’s kappa analyses were conducted using SPSS version 30. Results showed very strong and significant positive correlations among DSM-5, CARS-2, CASD, and Clinical Report Summary scores (p &lt; .001). DSM-5 was strongly correlated with CARS-2 (rₛ = .923), CASD (rₛ = .885), and clinical reports (rₛ = .887). The strongest relationship was found between CASD and clinical reports (rₛ = .996). Weighted kappa values ranged from .593 to .768, indicating moderate to substantial agreement across severity classifications. The strongest agreement occurred between DSM-5 and CARS-2 (κ = .768). These findings demonstrate that the integrated model has good convergent validity and reliability. The study contributes a local empirical evidence for improving early ASD screening, referral decisions, and intervention planning in Sabah, Malaysia.</p>