BackgroundIn childhood autism spectrum disorder (ASD), the Childhood Autism Rating Scale–2nd edition (CARS2) instrument is used for diagnosis and assessment of severity and change. Following a feasibility study, we conducted 2 discrete choice experiments (DCEs), each of which used caregivers and clinicians as proxies for autistic children, to assess preferences for CARS2-based attributes. MethodsCaregivers and clinicians from 5 European countries received an online DCE corresponding to either the standard or the high-functioning version of the CARS2. Participants completed 14 choice tasks with 2 hypothetical profiles composed of 13 attributes set at 4 varying levels. To reduce task complexity, the 2 profiles of each choice task had at least overlap in 7 attributes, i.e., attributes were set at the same level, and presented in a stacked layout. Multinomial, mixed and scale-adjusted logit models were used to estimate preference weights. Explorative anchoring to the EQ-5D-Y was undertaken and a rescaled set of DCE coefficients is provided. ResultsModels were fit separately for caregivers and clinicians in each experiment. The final models included 563 caregivers and 666 clinicians for the standard experiment, and 346 caregivers and 310 clinicians for the high-functioning one. Caregivers and clinicians, as expected, prioritized some of the same attributes but not all. For example, in the standard experiment, Verbal communication, Non-verbal communication, and Activity-level were highly important attributes to both groups, whilst Taste, smell, and touch response and use was more important to caregivers than it was to clinicians. In the high-functioning experiment, preferences for caregivers were highest for Thinking/cognitive integration skills, Verbal communication and Fear or anxiety. For clinicians, the most important attributes were Thinking/cognitive integration skills, Fear or anxiety and Social-emotional understanding. ConclusionCaregiver and clinician preferences indicated some disparity in what constitutes the greatest unmet need for autistic children. These findings can support clinicians and caregivers in creating mutual understanding and agreement on therapeutic goals, which can be very direct and near (e.g., Taste, smell, and touch response and use) or future-oriented and developmental (e.g., Relating to people).