Turning data into safety: A data-driven approach to mitigating fall risks for ageing adults with type 2 diabetes
Nida Afzal, Amy D Nguyen, Annie Y S Lau Australian Institute of Health Innovation, Macquarie University, Sydney

Growing need for safe independent living With rising prevalence of type 2 diabetes (T2D) among ageing adults, the need for safe, independent living environments grows

Challenges of functional decline Progressive decline associated with T2D increases fall risks and complicates daily activities

Limited research on environmental impact

Despite existing self-management interventions, there is little research on how home environments impact safety and independence

Study objective

This study analyses real-world home data to identify environmental challenges such as poor lighting, clutter, and obstacles in trafficways

Anticipating future declines By considering future declines in vision, cognition, and mobility, we aim to mitigate risks before they arise

Data Collection

~5000 images per patient (26 older people with T2D) using body worn cameras

Activity Classification

Use Machine Learning to classify images into 12 activity classes

Manual Image Coding Health management activities, technological interactions and environmental context coded using a validated tool

Risk Identification Identify areas where future decline might pose challenges by anticipating potential physical, functional and cognitive limitations

Formulating Recommendations

Digital interventions and environmental modifications

Environmental Hazards

Clutter Recurring in several areas (e.g., living rooms, hallways)

Inadequate lighting Contributing to poor visibility

Preliminary Analysis 4 older people with T2D, 22 images

Visual Challenges

Difficulty navigating due to poor lighting

Fine Motor Skill Challenges

Difficulty administering medication

Cognitive Challenges

Confusion or difficulty navigating cluttered environments

Mobility Challenges

Accessibility barriers increasing fall/trip risks

RISK IDENTIFICATION

Image analysis helps foresee risks associated with health decline

DEVELOPING PROACTIVE SOLUTIONS

Smart technology interventions Environmental modifications

BROADER IMPACT

Well-being enhancement

Reducing assisted living reliance - sustainable, cost-effective healthcare system

1. Yin K, Harms et al. Patient work from a context and time use perspective: a mixed-methods study protocol. BMJ Open. 2018 2. Xiong H et al, Identifying daily activities of patient work for type 2 diabetes and co-morbidities: a deep learning and wearable camera approach. Journal of the American Medical Informatics Association. 2022 3. Clemson L et al, Content validity of an assessment tool to identify home fall hazards: The Westmead Home Safety Assessment. British Journal of Occupational Therapy. 1999

To learn more about this study or provide feedback, please feel free to reach out Nida.Afzal@hdr.mq.edu.au Nida Afzal | LinkedIn

2024.09a011e