Towards Developing "Oops! Solver" for Elderly Care

Area of Science:

• Gerontology
• Computer Science
• Health Informatics

Background:

• Aging societies face rising long-term care costs and caregiver shortages.
• Information and communication technologies (ICT) offer potential solutions for human resource development in caregiving.
• Existing knowledge resources for caregiving lack computer-interpretability, hindering ICT application.

Purpose of the Study:

• To develop a computer-interpretable knowledge resource for elderly care question-answering systems.
• To address the shortage of caregivers and reduce their burden through technological solutions.
• To create a structured manual of elderly care incidents and corresponding queries.

Main Methods:

• Conducted semi-structured interviews with caregivers to collect "Oops" incidents in daily caregiving.
• Utilized an open dataset to supplement incident collection.
• Developed natural language questions and corresponding computer-interpretable queries based on collected incidents.

Main Results:

• Collected 150 "Oops" incidents from daily caregiving practices.
• Created 33 computer-interpretable queries linked to natural language questions.
• Validated that the developed queries can effectively retrieve relevant knowledge from the structured manual.

Conclusions:

• Structured manuals of elderly care incidents can serve as effective knowledge resources for question-answering systems.
• Computer-interpretable queries derived from real-world incidents enhance the utility of digital knowledge bases.
• This approach offers a scalable solution to support caregivers and improve long-term care services.