A Pilot Study on Fall Risk Assessment–Based Support for Fall Prevention Training Reworked Abstract: Falls among older adults are a leading cause of fractures and care dependency, underscoring the importance of early prediction and preventive support. Conventional fall risk assessments require expert-led motor function tests and expensive equipment, imposing significant physical and financial burdens that hinder routine implementation. Moreover, analyses based on fixed, retrospective datasets have shown limited adaptability to local characteristics or temporal changes in target populations. This study aims to develop an application that enables low-burden fall risk prediction and individualized preventive support using a questionnaire designed to be easily answered by older adults. Based on the methodology of Toba et al., we developed an application that performs logistic regression analysis on the collected questionnaire data. For preliminary validation, data were collected from 19 older adults aged 60–90 years (7 men, 11 women, 1 unknown) who visited a hospital for outpatient health examinations. In addition, pseudo-data reflecting previous studies and clinical findings were utilized. The results showed sensitivity and specificity comparable to or exceeding those of conventional methods. The system is designed to continuously incorporate real-world data obtained from medical institutions, allowing iterative model updates. Furthermore, explanatory variable selection was designed to flexibly integrate a wide range of information, considering not only statistical significance but also practical usability. These features enable (1) customization for specific populations and (2) adaptability to temporal changes. Unlike conventional static models, this study focuses on developing a practical diagnostic support system, which represents the key novelty of this approach.

Authors:
Kamiya Yuko, Fukuoka Women's University, Japan
Sakaki Taisuke, Kyushu Sangyo University, Japan
Fujino Tomokazu, Fukuoka Women's University, Japan
Shimokawa Toshihiko, Kyushu Sangyo University, Japan

About the Presenter(s)
Yuko Kamiya, Lecturer at Fukuoka Women’s University, received her Ph.D. in Information Science from Kyushu Sangyo University. Her research focuses on computer and network education, practice-based learning, and student support in IT courses.

See this presentation on the full schedule – Thursday Schedule

Bookmark