Mental health problems among adolescents and young adults have increased globally, underscoring the urgent need for accessible, preventive, and systemic approaches to promoting student mental health. This pilot study aims to evaluate the feasibility, conversational quality, and preliminary effectiveness of an AI-powered chatbot integrated with wearable-derived physiological data intervention to provide personalized and timely mental health support for Finnish higher education students. The study explores how real-time stress indicators, such as heart rate variability (HRV), can inform the chatbot's adaptive and proactive interactions to promote students' mental health in daily life.
The study setting will be the Faculty of Medicine, University of Turku in Finland. A total of 15-20 participants will be recruited with convenience sampling for this study. The inclusion criteria for participants are: 1) enrolled students at the Faculty of Medicine, 2) voluntary and consenting, 3) Finnish-speaking, 4) having a smart phone (Android/iOS), and 5) having none to mild levels of psychological distress. Participants who self-report a diagnosis of mental health disorder, or having an acute treatment for mental health will be excluded.
The intervention in this study consists of 4-week daily use of a AI-powered chatbot. In addition to access to the chatbot, a wearable device will be given to each participant to collect physiological data related to their stress (HRV), physical activity and sleep. At the end of week 2 and 3 of the study, participants will receive a personalized report summarizing the data collected by the device. Throughout the intervention period, the research assistant will contact participants once a week to ensure their well-being. At the end of the study, an interview will be arranged with each participant.
The study data includes the conversations with the chatbot, physiological data from the wearable device, daily stress questions, background information questionnaire, Depression Anxiety Stress Scale (DASS-21) (at the beginning of the intervention and at weeks 2 and 4), Chatbot Usability Questionnaire (CUQ) (at weeks 2 and 4) and an end-of-study semi-structured interview.