Human-Centered AI In Healthcare: Voice Assistant For Accessible Remote Health Interventions
Open Access
- Author:
- Qiu, Ling
- Graduate Program:
- Informatics
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- February 14, 2025
- Committee Members:
- Dongwon Lee, Professor in Charge/Director of Graduate Studies
S. Shyam Sundar, Outside Unit & Field Member
Xinning Gui, Major Field Member
Jack Carroll, Major Field Member
Saeed Abdullah, Chair & Dissertation Advisor - Keywords:
- health
voice assistant
human computer interaction
health interventions
breast cancer
dementia - Abstract:
- Delivering health interventions in remote and underserved areas has been a persistent challenge. Meanwhile, artificial intelligence (AI) has advanced significantly in recent years, offering transformative opportunities for healthcare. Despite substantial research integrating AI into healthcare, there remains a notable gap in leveraging AI to deliver remote health interventions to these populations. This dissertation aims to address this research gap by focusing on the development of Human-Centered AI (HCAI) systems for delivering remote health interventions. Through a human-centered approach, I designed, developed, and evaluated two AI-driven voice assistants (VAs) tailored for distinct healthcare contexts. The first VA delivers personalized self-management support interventions to women living with metastatic breast cancer (MBC). A 2-week usability study and a 6-month randomized controlled trial demonstrated the system’s feasibility and usability for this population. The second VA provides individual Cognitive Stimulation Therapy (iCST) to persons living with dementia (PLwDs). I explored the system’s feasibility and generated design considerations via exploratory prototyping and semi-structured interviews, laying a foundation for future long-term deployment. This dissertation contributes to leveraging HCAI for healthcare in three key ways. First, it introduces two novel VA systems, guided by the NIH stage model for behavioral intervention development, to support healthcare for women with MBC and PLwDs. Second, it demonstrates feasibility and usability of using VA to deliver remote health interventions in these contexts. Lastly, it provides evidence-based design guidelines for developing VA-based healthcare systems, and offers broader insights into designing HCAI systems that improve accessibility and effectiveness in remote health interventions.