Analysis of the Quality of Individualized Care Based on the Structuring of Care Records Using an LLM
Abstract
In response to the growing shortage of long-term care personnel and the lack of educational opportunities, this study developed a system that uses a large language model (LLM) to structure and visualize free-form care records based on frameworks such as the International Classification of Functioning (ICF). This system has been introduced as a feedback support tool for on-site leaders, enabling a multifaceted analysis of the process of transforming the quality of individualized care. Results from a pilot study involving 10 caregivers confirmed a shift in awareness toward multifaceted observation perspectives among 80% of the caregivers, and an improvement in the consistency rate of ICF codes between care plans and care records was observed in 70% of them. This method goes beyond conventional, formal record audits and serves as an effective approach for realizing individualized support that integrates the care recipient’s preferences and life circumstances based on objective data.
Sakura Yui
Master’s Student (M1)
Atsushi Omata
Research Associate
