Tutun S, Johnson ME, Ahmed A, et al. An AI-based Decision Support System for Predicting Mental Health Disorders. Inf Syst Front. 2023;25(3):1261-1276. doi:10.1007/s10796-022-10282-5
Recent calls to action have encouraged researchers to leverage artificial intelligence (AI) to develop decision support systems (DSS) that assist mental health professionals and policymakers. In a direct response to this call, this study presents the development and assessment of a mental health assessment tool. Participants completed an online version of the SCL-90-R using a secure platform following a referral from their mental health professional. The SCL-90-R is a 90-item assessment used to screen for ten major mental health disorders. Since the platform opened in 2019, over 6,000 participants have taken the SCL-90-R using the online platform. To optimize the SCL-90-R, data from participant responses was analyzed to reduce the number of items and time to complete the assessment without reducing the number of detectable diagnoses. The final reduced version, named the SCL-28-AI, contained 28 items. This shorter version was tested against the initial 90-item version using a combination of traditional and AI models. The AI models using the shortened version still detected and accurately diagnosed all 10 disorders with only a 9% decrease in performance measures compared to the full 90-item version. The use of AI models in DSS aside, this suggests that the SCL-28-AI can be used by mental health professionals in place of the much longer and more cumbersome SCL-90-R. Future work aims to tailor and implement the DSS with the help of mental health professionals.