Acceptability and Pilot Validation of the Diagnostic Autism Spectrum Interview (DASI-2) Compared with Clinical and ADOS-2 Outcomes.

BACKGROUND/OBJECTIVES: There is a growing need for autism spectrum disorder (ASD) assessment tools that are diagnostically aligned, clinically usable, and accessible across diverse service contexts. The Diagnostic Autism Spectrum Interview-Version 2 (DASI-2) is a freely available, semi-structured clinical interview mapped directly to DSM-5 and ICD-11 criteria. This pilot study aimed to adapt DASI-2 into Hungarian and explore the (1) acceptability of DASI-2 administration, (2) agreement with prior clinical ASD diagnoses, and (3) relationship between DASI-2 observational ratings and ADOS-2 classifications.

METHODS

Following a multistep translation procedure, DASI-2 was administered to seven children previously assessed for ASD in a multidisciplinary Hungarian clinical setting. The assessment included a parent interview, direct assessment with the child or young person, and completion of the DASI observational record (OR1-OR4). DASI diagnostic outcomes were compared with prior clinical decisions, and OR scores were analyzed in relation to ADOS-2 classifications.

RESULTS

All participants completed the DASI-2 interview in full. Agreement with prior clinical diagnosis was found in six of seven cases (κ = 0.70, indicating substantial agreement). When exploring the one non-aligned case, the divergence in diagnostic outcome was due to broader contextual information considered by the initial clinical team which influenced clinical opinion. The five participants diagnosed with ASD showed substantially higher DASI observational scores (mean = 15.26) than the two who were not diagnosed (mean = 1.57), mirroring ADOS-2 severity classifications.

CONCLUSIONS

These findings support the acceptability and preliminary validity of DASI-2. Its inclusive structured observational record may provide a practical complement to resource-intensive tools such as the ADOS-2; however, further validation in larger and more diverse samples is needed.