Scholarlink

Background <p>Clinical fall risk assessments, such as the Fall Risk Assessment Scale for the Elderly (FRASE), often rely on subjective clinician observation. This protocol described the evaluation of “iBalance”, an AI-enabled device utilizing Frustrated Total Internal Reflection (FTIR) to objectively measure dynamic balance and predict fall risk. This study aims to validate the screening accuracy of the iBalance device and explore its feasibility for clinical integration.</p> Methods <p>412 adults aged 60 years and above and 10 clinical professionals will be recruited from outpatient clinics for this mixed-method study. Older adults’ fall risk will be assessed by the iBalance and clinical assessments (e.g., FRASE scale) separately, and their fall incidents will be recorded during a three-month follow-up. The iBalance captures high-resolution plantar pressure at 30&#xa0;Hz. The system processes data via: (1) a machine learning model trained on extracted 3D Center of Pressure (CoP) and Center of Gravity (CoG) features; (2) a Convolutional Neural Network (CNN) for direct video data analysis; and (3) a statistical method optimized for maximum Area Under the Curve (AUC). To ensure participant safety, all the assessments will be conducted with side handrails and staff guarding to prevent falls. Individual interviews with clinical professionals will be conducted to gather insights on improving the device’s practicality and the application of clinical workflows. The primary outcome is sensitivity, specificity, and AUC of iBalance compared to the clinical assessments. Logistic regression will adjust for key confounders (age, gender, and BMI) for both fall risk factor identification and 3-month prospective fall prediction. Qualitative interviews will undergo thematic analysis.</p> Discussion <p>This protocol outlines a comprehensive evaluation for validating an AI-driven fall risk assessment tool. By addressing both technical screening accuracy and practical clinical utility, the study seeks to establish a standardized, objective framework for early fall intervention in geriatric care.</p> Ethics and trial registration <p>This study has been approved by the IRB of The University of Hong Kong/Hong Kong Hospital Authority Hong Kong West Cluster (IRB reference number: UW 24–540). This protocol has been registered on the US ClinicalTrial.gov website (NCT06767163, registration date: January 16, 2025).</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Evaluation of a Frustrated Total Internal Reflection (FTIR) based balance sensor for objective fall risk assessment in older adults: a study protocol

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年3月2日

查看全文
  • Huanran Liu,
  • Liqing Cai,
  • Xin Ma,
  • Jiaming Chen,
  • Cindy Lo Kuen Lam,
  • Jacqueline Kwan Yuk Yuen,
  • Chin Ping Chan,
  • Hon Ho Sze,
  • Hua Li Wang,
  • Ning Xi,
  • Vivian Weiqun Lou

摘要

Background

Clinical fall risk assessments, such as the Fall Risk Assessment Scale for the Elderly (FRASE), often rely on subjective clinician observation. This protocol described the evaluation of “iBalance”, an AI-enabled device utilizing Frustrated Total Internal Reflection (FTIR) to objectively measure dynamic balance and predict fall risk. This study aims to validate the screening accuracy of the iBalance device and explore its feasibility for clinical integration.

Methods

412 adults aged 60 years and above and 10 clinical professionals will be recruited from outpatient clinics for this mixed-method study. Older adults’ fall risk will be assessed by the iBalance and clinical assessments (e.g., FRASE scale) separately, and their fall incidents will be recorded during a three-month follow-up. The iBalance captures high-resolution plantar pressure at 30 Hz. The system processes data via: (1) a machine learning model trained on extracted 3D Center of Pressure (CoP) and Center of Gravity (CoG) features; (2) a Convolutional Neural Network (CNN) for direct video data analysis; and (3) a statistical method optimized for maximum Area Under the Curve (AUC). To ensure participant safety, all the assessments will be conducted with side handrails and staff guarding to prevent falls. Individual interviews with clinical professionals will be conducted to gather insights on improving the device’s practicality and the application of clinical workflows. The primary outcome is sensitivity, specificity, and AUC of iBalance compared to the clinical assessments. Logistic regression will adjust for key confounders (age, gender, and BMI) for both fall risk factor identification and 3-month prospective fall prediction. Qualitative interviews will undergo thematic analysis.

Discussion

This protocol outlines a comprehensive evaluation for validating an AI-driven fall risk assessment tool. By addressing both technical screening accuracy and practical clinical utility, the study seeks to establish a standardized, objective framework for early fall intervention in geriatric care.

Ethics and trial registration

This study has been approved by the IRB of The University of Hong Kong/Hong Kong Hospital Authority Hong Kong West Cluster (IRB reference number: UW 24–540). This protocol has been registered on the US ClinicalTrial.gov website (NCT06767163, registration date: January 16, 2025).