Design and Development of an NLP-Based Mental Health Pre-screening Tool for Undergraduate Students in Thailand: A Usability Study

Fear of stigmatization has been a barrier for Thai undergraduate students to actively reach out to health practitioners for support when they experience a mental health problem. Commercially available mental health systems struggle to find a balance between effective clinical diagnostic capability and engaging user experience. In natural ambience, it is proven that users could better articulate their thoughts, which results in a higher efficacy of prescreening results in mental health assessment. Natural language processing (NLP) techniques are promising to enable this by developing a human-like digital assistant for mental health. The usability and user interface design of such mental health assessment tool should support the frictionless interaction while being compatible with the Thai context (e.g., language). To fill these gaps, in this study, we developed an NLP-based digital healthcare assistant as a pre-screening tool that can be used to detect undergraduate students’ anxiety levels and identify their needs for psychological support. We conducted a pilot usability evaluation of the system focusing on the system’s usability effectiveness, efficiency, and reliability. The findings can be concluded that creating a natural and user-friendly experience for users, with flexibility that allowed for individual preferences through a chat-based NLP system, results in an engaging user experience and less friction towards the adoption of the tool. These findings can be used to support the future development of an effective and user-friendly pre-screening tool for this particular user group.