Digital mental health assessments can effectively link young people to appropriate health services. Current approaches often lack personalization, failing to recognize the complex and multidimensional needs of young people. 1734 young people aged 12–25 years completed seven standardized measures (49 items) using a digital health assessment tool while receiving mental health care. A multidimensional computerized adaptive test (MCAT) was developed to predict scores on seven standardized scales, spanning clinical symptoms, suicidality, functioning, and alcohol use. Different adaptive tests were simulated under various stopping criteria configurations. Ten-fold cross-validation was performed to determine the accuracy and efficiency of the multidimensional assessment. By administering a personalized subset of items to each individual, the average number of assessment items could be reduced by 69% while maintaining excellent agreement with full-length scores for suicidality (ICC = 0.96), anxiety (ICC = 0.92), and alcohol use (ICC = 0.91), and good agreement for psychological distress (ICC = 0.88), functioning (ICC = 0.86), psychosis (ICC = 0.78), and mania (ICC = 0.75). Estimated average assessment time decreased from 10.5 minutes to under 3.3 minutes (49 items reduced to 15.3 items, per person, with mean absolute agreement ICC = 0.87). This adaptive digital assessment can screen across key domains to identify mental health needs and complexity in youth mental health, leading to rapid decisions about treatment needs and care pathways.
Capon et al. (Thu,) studied this question.