Understanding the Impact of the COVID-19 Pandemic: A Population-Level Natural Language Processing Study of Youth Mental Health

Young people in Canada are struggling with mental health at unprecedented levels. About one in five youth face a mental health disorder, and the COVID-19 pandemic made this crisis even worse. Many youth cannot access the help they need due to long waitlists, their geographical location, or the stigma surrounding seeking support. Kids Help Phone (KHP), a free national mental health service, has recorded over 1.5 million text and chat sessions from young people since 2018 seeking mental health support. While these conversations contain invaluable insight into what youth are experiencing, KHP currently lacks the tools to fully understand patterns in this data, such as who is reaching out, how their needs change over time, and what strategies from responders make the most difference. Our project will use natural language processing (NLP), a type of artificial intelligence that enables computers to understand human language, to analyze KHP's de-identified text and chat records from 2018-2025. This approach will allow us to track important mental health features, such as signs of anxiety, sadness, anger, and overall mood, before, during, and after the COVID-19 pandemic. We will also examine whether these patterns vary across youth from diverse backgrounds, including different age groups, racial identities, and sexual orientations, to support equitable decision-making and care for youth from diverse backgrounds. By uncovering how youth mental health changed during the pandemic and what kinds of support work best, this research will provide KHP with direct guidance to strengthen their counseling practices. Ultimately, my work aims to make sure that young people across Canada receive faster, more effective, and more equitable mental health support when they need it most.