Designing a Verbal Instruction-Based Medication Adherence System for Individuals with Memory Impairment
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Medication management can be challenging for people with memory impairments, especially when dosing instructions must be remembered and applied throughout the day. To address these challenges, this work investigates the design requirements and technical feasibility of using natural speech input to extract the timing information required for medication reminder support from verbal instructions. This approach aims to minimize the need for manual interaction and provide more accessible reminder-based assistance. To understand user expectations and requirements for such a system, we conducted an online survey$(N=44$). Among other insights, survey responses emphasized the importance of effective reminder support, which informed our decision to focus the prototype on the temporal dimension of medication assistance. We developed a proof-ofconcept processing pipeline that transcribes spoken Japanese medication instructions and uses a schema-constrained Large Language Model to generate structured temporal medication events suitable for generating time-based reminders. A preliminary validation showed that text-based inputs were extracted reliably, while speech inputs also achieved high accuracy, with most remaining errors attributable to medication name misrecognition during transcription. These findings demonstrate the technical feasibility of converting spoken medication instructions into reliable reminder schedules, establishing a foundation for future context-aware assistance on wearable and mobile devices.
Fields of Study & Subjects
Medication Adherence and Compliance
Speech and dialogue systems
Mobile Health and mHealth Applications
Publication details
- Journal / source
- IEEE International Conference on Artificial Intelligence and eXtended and Virtual Reality (AIxVR)
- Pages
- 365–372
- Publication date
- 26 January 2026
- Language
- EN